Livestock Research for Rural Development 25 (2) 2013 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Determinants of poverty in (agro-) pastoral societies of Southern Ethiopia

E Adugna and M Sileshi*

Jimma University, College of Agriculture & Veterinary Medicine, Department of Agricultural Economics & Extension,
P.O. Box 307, Jimma Ethiopia
adugna_e@yahoo.com
* Arba Minch University, College of Social Sciences & Humanities, Department of Civic and Ethical Studies:
P.O.Box 21 Arba Minch, Ethiopia

Abstract

Most previous studies dealing with poverty in the (agro-) pastoral context paid little attention to the application of quantitative methods. Thus, this study employed logistic regression in examining the household level determinants of poverty in (agro-) pastoral societies.

The results showed that over half of the respondents were poor, i.e., they do not meet the subsistence cost of basic needs. On the basis of poverty measure; pastoralism is inferior to agro-pastoralism in terms of escaping poverty. Family size, land size; market distance, extension contact, and annual income significantly explain the variations in the likelihood of being poor. Intervention options in terms of strengthening extension support towards forage and water development, market linkage creation, expansion of veterinary and family planning services would be a key for poverty reduction.

Key words: binary logit, capital, Hamer, livelihoods, livestock


Introduction

Ethiopia has the largest livestock population in Africa and more than 61% of its landmass is inhabited by pastoralists and agro- pastoralists (PFE 2008). Ethiopian pastoralists raise a large portion of the national herd, estimated at 42 percent of the cattle, 7 percent of the goats, 25 percent of the sheep, 20 percent of the equines and all of the camels (PFE et al 2010).  There are two debates on the viability of pastoral livelihoods in this 21st century. The first one argued that pastoralism is still a viable strategy if appropriate development initiatives link rural households to the markets (Little et al 2003; Moritz 2012), while the second disclose the fact that pastoral livelihoods are depressed and unviable (Little et al 2010) due to political marginalization (Eyasu 2008; Pavanello 2009; Elias and Abdi 2010), resource depletion, and drought (PFE et al 2010; UN OCHA 2007); lack or inadequate institutional support like market, education, health services (Gebru et al 2004; Pavanello 2009; PFE 2002). As a result, poverty remains intense in pastoral areas of Ethiopia (Elias and Abdi 2010; PFE 2009; Little et al 2006; Swift 2004); and pastoralists tend to be perpetual famine relief clients (Helland 2006). More than any other pastoral groups in Ethiopia the South Omo pastoralists have been neglected in terms of research and development programmes (Nori et al 2008).

While there has been strong urge from the government to settle pastoralists as a solution to combat poverty (Eyasu 2008), others propose access to free grazing and mobility as key factors of poverty reduction in pastoral areas (Kejela et al 2007, PFE et al 2010; Little et al 2010; Yacob and Catley 2010). Thus, the issue of poverty reduction strategy in pastoralist communities of Ethiopia, and how best to respond to it, remains debatable. Even though different actors have different perspectives, in more recent development debates, pastoral destitution and poverty are often attributed to conflict, climate change and weak governance (Yacob and Catley 2010).

Most studies dealing with poverty in the (agro-) pastoral areas paid little attention to household level correlates of poverty and applied qualitative approaches at community level, they often focused on  understanding of wealth stratification (e. g, Kejela et al 2007; Yacob and Catley 2010) which do not grasp household level severity and poverty gaps. To remedy this we need to analyze poverty based on household level data. Therefore, this study is aimed at identifying the determinants of poverty in (agro-) pastoral society of Southern Ethiopia.


Methodology

Description of the study area

Southern Nations Nationalities and Peoples Region (SNNPR) is one of the largest regions in Ethiopia, accounting for more than 10 % of the country’s land area. The population is estimated at nearly 15,745,000 (CSA 2008); almost a fifth of the country’s population. The region is divided into 13 administrative zones (SNNPR 2005). Among these, South Omo zone (Fig 1) is one of the most remote and the poorest people in Ethiopia.  

Figure 1. Map of the study districts
Sampling techniques

Ethiopian (agro-) pastoralists mainly reside in four regions, Afar, Somali, Oromiya and Southern Nations Nationalities and Peoples Region. This research however focused on the latter due to proximity, and financial limit. Moreover, the specific study area South Omo is selected due to its relative remoteness and ignorance by previous researches conducted in the region. Once the study zone is decided; to select respondents two-stage stratified random sampling was used. Among the six pastoral and agro pastoral areas of south Omo, Hamer and Bena Tsemay districts were selected representing pastoral and agro pastoral livelihoods respectively. Then, two Peasant Associations (PAs) from each district were selected purposely. Finally 197 households were selected randomly.

Methods of data collection

Focus group discussions were conducted at each village in order to understand the ground reality. A total of four focus group discussions consisting of eight to fifteen peoples were conducted. Quantitative data on demographic, social and economic characteristics, livestock production, and wellbeing was collected from the sampled household heads through structured interview.

Methods of data analysis

The major data analysis methods used were descriptive and econometrics. The descriptive statistical tools like mean, standard deviation, X2 test, t tests were used to explore mean variations between the poor and non poor. To identify the determinants of poverty binary logit model was fitted. The data analysis was conducted using SPSS soft ware version 16.

Poverty measures

Poverty is characterized by inadequacy or lack of productive means to fulfill basic needs such as food, water, shelter, education, health and nutrition (World Bank 1986). The concept of pastoral poverty in general goes along with the definition of absolute poverty, a situation of deprivation of basic needs, independent of the general style of living in a given society. Greer and Thorbecke (1986), state that setting the poverty line using the cost of calorie approach is conceptually and computationally simple. Thus, this study was based on the cost of basic needs to determine the poverty line. Accordingly, a bundle of food items usually consumed by the lowest income category is used to determine the local poverty line (Table 1). Once, the poverty line was set, poverty indices were measured using the Foster, Greer, and Thorbecke (FGT) formula. Among the various methods of quantifying poverty, the FGT formula (Foster et al 1984) is the most widely used. The formula has been successful in providing a quantitative description of the spread, the depth and severity of poverty in populations.

Given a vector of suitable measure of well-being, Y, in increasing order, Y1, Y2, Y3,...,Yn, where n represents the number of households under consideration, the FGT poverty index (Pa) can be expressed as:

Where z is poverty line

 q is the number of the poor,

gi is shortfall in chosen indicator of well-being. If, for instance, xi denote the per capita calorie intake of household i, then gi = zi-xi if xi < z ; gi =0 if xi z, and a is the poverty aversion parameter (a 0) 

The parameter a represents the weight attached to a gain by the poorest. The commonly used values of a are 0, 1, and 2. When a equals to 0, then (1) is reduced to the headcount ratio, which measures the incidence of poverty. When we set a equal to 1, we obtain P1 or the poverty deficit. P1 takes in to account how far the poor, on average, are below the poverty line. Setting a equal to 2 gives the severity of poverty or FGT (2) index. This poverty index gives greater emphasis to the poorest of the poor, as it is more sensitive to redistribution among the poor.

 Binary logit model specifications

Many studies have ensured that when the dependent variable of interest is of qualitative nature and dummy, binary logit model is the best (Gujarati 2003; Green 2003). Similarly, since the dependent variable in poverty analysis qualifies to the above notion, binary logit model was used. Following Gujarati, 2003), the functional form of logit model is specified as


As
Gujarati (2003) noted, before running the model, Variance Inflation Factors and contingency coefficients need to be computed to ensure the existence of correlation among explanatory variables. Variance Inflation Factor (VIF) was used to measure the degree of linear relationships among the continuous explanatory variables and contingency coefficient was used to check multicollinearity among discrete variables.


Results and discussion

Poverty status

This study used cost of basic needs approaches to set the poverty line. Accordingly the food and non food poverty lines were determined based on the lowest income quartile groups. The food poverty line (FPL) was indicated in Table 1. Calculating the total expenditure and total kilo calories (Kcal) of all the food (EHNRI, 2000) consumed by the reference population the ratio between the two totals is a price per kcal estimate which when multiplied by the energy threshold provides an estimate of FPL (Ravallion and Bidani 1994). Accordingly, the FPL for this study is 1322.67 Ethiopian Birr (ETB).

Table 1. Food Poverty Line

Food items

Kg consumed

*Mean kcal

Kcal/day/AE

Kcal share

Average price

Food poverty

Expenditure

 

per AE per day

per kg/lt

 

(%)

per Kg/Lt (Birr)

 line/year (Birr)

Share (%)

Maize

0.2307

3751

937

42.58

2.93

247

18.68

Sorghum

0.2618

3805

1079

49.03

3.51

335

25.36

Milk

0.1450

737

116

5.26

8.80

466

35.21

Meat

0.0150

1148

19

0.85

38.60

211

15.94

Butter

0.0021

7364

17

0.77

44.25

34.4

2.60

Oil

0.0001

8964

0.57

0.03

20.00

0.43

0.03

Salt

0.0013

8120

11.2

0.51

3.00

1.39

0.11

Shoforo

0.0179

1103

21.4

0.97

4.2

27.50

2.08

Total

 

 

2200.00

100.00

 

1323

100.00

* Source: EHNRI (2000); shoforo refers to coffee husk.

The Non food poverty Line was determined using a simple linear regression of the share of food (FE) to total expenditure (TE); using data from the reference poor population. That is,

                        Si = α+ βlog (te/fpl)i + errori .

where i runs through the sample households in the reference population. Accordingly, the value of non food poverty is 68 birr. The poverty line is the sum of food and non food poverty values, which is 1392.3 ETB, per adult equivalent per year.  Based on this poverty line, 50.8% and 49.2% households are poor and non poor respectively.

The resulting poverty estimates for the study area (Table 2) shows that the percentage of poor people measured in absolute head count index (α = 0) is about 50.8%. This implies that 50.8% of the population are unable to get the minimum calorie required (2200 kcal per day per adult) adjusted for the requirement of non food items expenditure. Although more than half households are generally considered poorer, there are also significant differences between pastoralist and agro pastoralist societies. The proportion of people with standard of living below poverty line is 58.2% and 44.3% for Hamer (pastoral) and Bena Tsemay (agro pastoral) respectively.

The poverty gap index (α=1), a measure that captures the mean aggregate consumption shortfall relative to the poverty line across the whole population is found to be 0.279 which means that the percentage of total consumption needed to bring the entire population to the poverty line is 27.9%. The figure for Hamer and Bena Tsemay is 0.439 and 0.142 respectively. Similarly, the FGT severity index (the squared poverty gap, α=2) in consumption expenditure shows that 7.8% fall below the threshold line implying severe inequality. This indicates that poverty is more severe pastoral context than agro pastoral.

Table 2. Absolute Poverty Indices

Study site

Head count index (α=0)

Poverty gap (α =1)

Squared poverty gap

(α =2)

Hamer

0.582

0.439

0.193

Bena Tsemay

0.443

0.142

0.020

Overall

0.508

0.279

0.078

The correlation of livelihood capitals with poverty
Human capital

The human capital variables considered in Table 3 includes sex, education and age of household heads, family size and dependency ratios. Among the 100 poor households, 45 were lead by women, which accounts for about 66.2% of the total female headed households; this is a significant difference at less than 1% probability level. This implies that women headed households disproportionately fall under the poverty than men counterparts. Access to education is very limited in pastoral context. Similarly, the survey indicated that 58.2 and 41.1 percent of the poor and non poor were significantly illiterate respectively. The groups with the highest poverty cases have no education while those with the higher education have lower cases of poverty. This implies that even though education does not generate remunerable employment in pastoral context; ability to read and write would have some advantages on poverty reduction. The average age of the poor (41 years) is significantly greater than that of the non poor households (38 years). This may indicate access to livelihood asset and wealth will be limited as one aging. Poor households have more family and fewer dependents than the non poor households; however it is not statistically significant.

Table 3. Human capitals by poverty status

 

 

 

Education category (%)

Illiterate

Primary

Secondary

X2

p value

 

Non poor

41.1

58.0

50.0

5.25

0.072

 

Poor

58.9

42.0

50.0

 

 

 

Total

48.2

44.7

7.1

 

 

 

 Sex of head

Male

 Female

 

 

 

 

 

 Non Poor

Poor

Non poor

Poor

9.873

0.002s

 

 

 57.4

42.6

33.8

66.2

 

 

 

   

Mean

SD

t

P value

 

Age of household head

Poor

40.91

10.71

1.958

0.052

 

 

Non poor

38.01

10.06

   

 

Family size in AE

Poor

4.80

1.69

-1.207

0.229

 

 

Non poor

4.45

2.30

   

 

Dependency ratio

Poor

1.13

1.01

-0.456

0.649

 

 

Non poor

1.20

1.19

   

 

***,* significant at less than 1 &10% probability levels respectively

 

Physical and natural capital

The most important physical asset in pastoral community is livestock ownership and access to grazing land. Majority of the respondents owned livestock (95.4%) although the size of livestock owned varies. Mean livestock holdings are highly skewed: 6.71 TLU/AE for an average household of 4.6 AE. On average, poor households own less TLU (5.75 per AE) than non poor ones (7.73 per AE). The difference between poor and non poor regarding livestock holding is not significant. This may imply that livestock number is not a matter of poverty variation.

The average land size held at household level by the poor and non poor households is 1.65 and 2.22 hectares respectively. There is also statistical difference at less than 1% probability level (Table 4). Concerning soil fertility of their land about 94% respondents reported that their land is of good soil fertility. Grazing land is communally administered by elders and clan leaders who formulate rules about resource use, administer enforcement, and ensure that sanctions and penalties are implemented. With respect to each households perception; about 7.1%, 46.2% and 47.7% respondents indicated that access to traditional grazing land has been increasing, not changed and decreasing respectively. In reality, however; access to grazing areas has decreased over time. This has been accelerated by changes in land-use policies at the national level which have encouraged public and private investments in the study areas. In this regard; the future will bring more challenges as a huge sugar factory is under construction in the study areas.

Table 4. Asset, income & expenditure by poverty status

Natural/physical assets

Poverty status

Mean

SD

t

p-value

Livestock / AE ( TLU)

Non poor

2.22

1.45

 

 

 

Poor

5.75

12.10

-864

0.388

 

Non poor

7.73

19.31

 

 

Income source

Poverty status

Mean

SD

 

 

Income from crop

Poor

754

1853

-3.885

>0.000

Non poor

2172

3069

 

 

Income from livestock

Poor

755

965

-4.275

>0.000

Non poor

2030

2732

 

 

Off farm income

Poor

626

2585

1.697

0.093

Non poor

179

513

 

 

Annual income

Poor

2105

3712

3.977

>0.000

Non poor

4258

3880

 

 

Own consumption value

Poor

3246

8941

5.685

>0.000

Non poor

10856

9809

 

 

Expenditure

Poverty status

 

 

 

 

Food expenditure

Non poor

1082

1215

2.716

0.007

 

Poor

699

599

 

 

Non food expenditure

Non poor

727

901

4.974

>0.000

 

Poor

244

356

 

 

 Annual expenditure

Non poor

12620

10276

6.170

>0.000

 

Poor

4133

9005

 

 

****, *; significant at less than 1& 10% probability levels respectively

 

Financial and economic capital

Financial assets were the scarcest capitals and availability and accessibility to formal financial institutions by pastoral society is very limited. As a result pastoral household’s use of credit even is constrained. From the surveyed households, only 27.41% household’s accessed credit from various sources. The main reason for not using credit was lack of awareness and fear of risk of payment. Similarly, access to market is generally poor in the area. Lack of access to market will increase transaction costs, like transport costs and may hinder poor households from participation. The mean distance to the market place in kilometer for the sample households is 15.6 km with a minimum of 6 km and a maximum of 35 km. The figure for poor and non poor is 17.92 and 11.95 respectively, which implies the poor are living at far distance from market points in remote areas. Moreover, there are no regular transport facilities and the community often travels on foot.  

The annual income of the households is a function of livestock, crops, and employment on off-farm activities. The mean annual income difference between poor and non poor groups was 2152.78 ETB, which is highly significant at less than 1 percent probability level. The mean crop and livestock income, and own consumption value are also significantly different between the two groups at 1 % probability level. However, the off farm income is highest for the poor households. This indicates that participation in off farm activities is a response to cope poverty rather than an investment. Like income, there is a significant association between poverty status and food, and non food expenditure, implying that the poor households have lower values for each predictor (Table 4).

Social and political capital

Social capital and collective actions are the daily realities of pastoral livelihoods. Pastoral communities are socially bound and help each other in various aspects. They support oneanother during herding animals, moving huts, and searching for lost animals and during burials. The majority of the households (92.6%) reported that they have received their relatives support in either forms of goat, milk, grain and labour during occasions of wedding, birth of children or funeral ceremonies. About 14.8 % respondents also reported that they have got dowries or bride-price during the survey period. Obviously, Ethiopian pastoralists were politically marginalized from participating in national decisions that affect their life until recently (PFE et al 2010). Due to political marginalization and lack of education, pastoralists` saying in policy was not considered so far. They, however, have their way of traditional systems of governance in which a clan leader govern affairs of the community.

Access to various services

Proximity to social services is an important determinant of household’s livelihood strategies. Among the institutional supports rendered to pastoralists food aid is crucial. Due to drought and rain fall failure people rely heavily on food aid. In principle relief food is distributed only during crisis times. But food deficit in this area is chronic. Even during the normal periods, many households do face seasonal food deficits. The proportion of households reported receiving relief food prior to survey period were over 28.93%. Among these, majority of the poor (77.2%) depends on food aid.

Access to extension support and service is very limited in pastoral context. Until recently, extension service is biased towards crop production and there has been less attention from the sector towards pastoralism. As a result, there are no or little pastoral specific extension packages so far.  In terms of extension agent consultation, the study result showed that 78.7 percent of the households were contacted (Table 5). Proximity to the different social services such as schools, human and livestock clinics, extension office, flourmills, all weather roads and drinking water has significant effect on poverty condition of pastoral livelihoods, but the poor have less proximity to these sectors

Table 5.  Access to social services by poverty status

Received

Non poor

Poor

X2

P value

 

%

%

 

 

Extension contact

43.2

56.8

10.78

0.001***

 Food aid

22.8

77.2

14.38

0.000***

Distance in kilo meter

Poverty status

Mean

SD

t

P value

Market

Non poor

11.95

8.15

-4.895

>0.000

 

Poor

17.92

8.97

 

 

Extension 

Non poor

2.15

2.63

-1.441

0.151

 

Poor

2.64

2.05

 

 

Grain mill

Non poor

2.29

3.01

-3.231

<0.001

 

Poor

3.80

3.53

 

 

Primary school

Non poor

1.23

1.59

-2.064

0.041

 

Poor

1.89

2.06

 

 

Potable water

Non poor

2.38

5.08

-1.006

0.316

 

Poor

2.97

2.76

 

 

***, **, significant at less than 1and 5% probability levels respectively

 

Determinants of poverty in (agro-) pastoral society

This part presents the determinants of poverty using binary logit model. Obviously, the variables to be included in the model were tested for the existence of multicollinearity and multicollenearity problem was observed between family size and dependency ratio, as well as sex and marital status of heads. Thus, dependency ratio and marital status were removed from the final model. The likelihood ratio has a chi-square distribution and is significant at less than one percent probability. Additionally, goodness of fit in logistic regression analysis is measured by count R2 and the model result show the correctly predicted percent of sample household is 77.1 percent (Table 6) The sensitivity, correctly predicted poor is 80.4 percent and that of specificity correctly predicted non poor is 73.7 percent. This indicates that the model has estimated the poor and non poor correctly.

Among the seven variables entered the model, five variables were found to significantly explain the probability of being poor. The results revealed that the coefficients of independent variables have expected sign and consistent with the logic of economic theory. The interpretation of these significant variables is given below.

Table 6. The maximum likelihood estimates of the logit model

 

Variables

B

S.E.

Wald

Sig.

Exp(B)

SEXHEAD

0.516

0.499

1.068

0.301

1.676

AGEHEAD

-0.031

0.024

1.647

0.199

0.969

FAMSIZ

0.530

0.136

15.258

>0.000

1.699

LANDSIZE

-0.853

0.219

15.191

>0.000

0.426

LIVESTOK

-0.002

0.004

0.468

0.494

0.998

DISMARK

0.079

0.028

7.689

0.005

1.082

EXTCONT

-0.889

0.500

3.154

0.076

0.411

ANNINC

-0.001

0.000

15.622

>0.000

0.999

Constant

-0.609

1.172

0.271

0.603

0.544

Pearson chi square

96.8***

 

 

-2 Log likelihood

169

 

 

 

Sensitivity

73.7

 

 

 

Specificity

80.4

 

 

 

Percent correctly predicted (Count R2)

77.10

 

 

 

sample size

197.0

 

 

 

*, **, *** significant at less than 10, 5 and 1% probability levels

Family size (FAMSIZE): this variable was significant and positively related with the state of poverty indicating that this variable was the cause of poverty. This implies that the odds ratio in favor of the probability of being poor increases with an increase in the family size measured in adult equivalent. The odds ratio of 1.699 implies that, other things being constant, the odds in favor of being poor increases by 69.9% with one additional adult in the family. This result was in congruent with that of Hillina (2005).

Land size (LANDSIZE): was significant and negatively related to the tendency of being poor. Other things being equal, the odds of being poor versus non poor would be 57.4% times greater with one unit increase in land size. This implies that cultivation or expansion of land under cultivation at household level is a means for poverty reduction in pastoral areas. It also supports the wisdom that poverty is more severe in pastoralists than agro pastoralists.

Market distance (DISMARKT): market distance has been found to be positively related with poverty and significant at less than 1 percent probability level. It was expected that households nearer to market centers had better chances to benefit from market incentives than those who are away from market centers. The odds in favor of being poor increases by 8% for one kilometer increase in market distance, citrus-paribus. Marketing in pastoral areas is typically complicated by high transaction costs due to the long distances that the pastoralist must travel and the poor infrastructure

Annual income (ANNINC): the annual income per household which is the sum of crop, livestock and off farm sources is negatively related with the likelihood of being poor at less than 1% probability level, similar finding with that of Hillina (2005). One unit increase in income decreases the probability of being poor by a factor of 0.99. However, households in the study area have very limited room for generation of income.

Extension contact (EXTCONT): The results of this study indicate that households access to extension service measured in terms of whether a household had a significant extension contact or not is found to negatively influence the likelihood of the household being poor. This implies that extension advice, trainings and technology promotion has an impact in poverty reduction efforts in pastoral areas. Access to extension service reduces the likelihood of a household to be poor by 58.9%.

Contrary to expectation, the results show that age, sex, and number of livestock owned are not statistically significant to explain the probability of a household being poor. Mainly, the case of livestock might be explained with the fact that livestock number is no more an option to escape poverty.


Conclusions


Acknowledgement

The authors would like to thank Arba Minch University for financing this research. Our special thanks would be to Dr. Kassa Tadele for facilitation. Our respondents are unforgettable for their valuable information and hospitality.


References

CSA 2008 Central Statistical Authority: population estimates, Ethiopia, Addis Ababa.

EHNRI 2000 Ethiopian Health and Nutrition Research Institute. Food Composition Table for use in Ethiopia, Addis Ababa.

Eyasu E 2008 Pastoralists in southern Ethiopia: Dispossession, access to resources and dialogue with policy makers. DCG Report No. 53, Norway.

Elias E and Abdi F 2010 Putting Pastoralists on the Policy Agenda: Land Alienation in Southern Ethiopia, Key highlights in sustainable agriculture and natural resource management.

Foster J, Greer J and Thorbecke E 1984 A Class of Decomposable Poverty Measures Econometrical, 52(3): 761-766.

Gebru G, Desta S and Coppock, D L 2004 Pastoralism in Ethiopia and the policy environment: Linking research, development actors and decision-makers. Proceedings Report Produced by the Pastoral Risk Management (PARIMA) Project of the Global Livestock Collaborative Research Support Program (GL-CRSP), August 15, 2003, Utah State University, Logan, Utah, USA., International Livestock Research Institute, Addis Ababa, Ethiopia, pp: 1-32. http://pdf.usaid.gov/pdf_docs/PNADE477.pdf

Gujarati D N 2003 Basic Econometrics, fourth edition. McGrew-Hill, Inc. New York.

Green H W 2003 Econometric Analysis. 4th Edn., University Macmillan Publishing Company, New York.

Greer J and Thorbecke E 1986 Food poverty profile applied to Kenyan smallholders. Econ. Dev. Cultural Change, 35: 115-141

Hillina M 2005 Dimensions and determinants of poverty in pastoral areas of Eastern Ethiopia: the case of Shinile zone in Somali national regional state, MSc. Thesis Alemaya University. 

Helland J 2006 Pastoral Land Tenure in Ethiopia Regime Foncier Pastoral in Ethiopia. Michelsen Institute, Bergen, Norway. 

Kejela G, Bezabih E and Waktole T 2007 Livelihood diversification in Borana pastoral communities of Ethiopia: prospects and challenges. Noragric, Department of International Environment and Development Studies, Norwegian University of Life Sciences. 

Little P D, McPeak J, Barrett C B and Kristjanson P 2006 The multiple dimensions of poverty in pastoral areas of East Africa; Proceedings of the Pastoralism and Poverty Reduction in East Africa: A Policy Research Conference, June 27-28, 2006, Nairobi, Kenya.

Little P D, Behnke  R, McPeak J and Getachew G 2010 Future Scenarios for Pastoral Development in Ethiopia, 2010-2025 Report Number 2 Pastoral Economic Growth and Development Policy Assessment, Ethiopia

Moritz M 2012 Review of Risk and Social Change in an African Rural Economy: Livelihoods in Pastoralist Communities by John G McPeak, Peter D Little and Cheryl R Doss. Pastoralism: Research, Policy and Practice 2 (24): 1-4. 

Nori M,  Taylor M and Sensi A 2008 Browsing on Fences; Pastoral Land Rights, Livelihoods and Adaptation to Climate Change. IIED Press, Nottingham, UK. ISBN: 1843697017, Pages: 28.  

PFE (Pastoralist Forum Ethiopia) 2002 Proposed pastoral development policy recommendations; Ministry of Federal Affairs, Pastoralist Forum, Ethiopia, Addis Ababa.

PFE 2008 Millennium Development Goals and Pastoral Development: Opportunities and Challenges in the new Ethiopian Millennium; Proceedings of the Fourth National Conference on pastoral development in Ethiopia, UN ECA Conference Hall, Addis Ababa. http://www.pfeethiopia.org/pub_files/PROCEEDING Pastoralist%20Conference%20final.pdf.  

PFE 2009 Pastoralist perspectives of poverty reduction strategy program: Experiences and lessons from Afar region of Ethiopia: PFE research report. Addis Ababa, Ethiopia. http://www.pfeethiopia.org/pub_files/PASTORAL%20PERSPECTIVES%20OF%20POVERTY%20REDUCTION.REGION%20ETHIOPIA.PFE%20WEBSITE.pdf. 

PFE, IIRR and DF 2010 Pastoralism and land: Land tenure, administration and use in pastoral areas of Ethiopia.  http://www.pfe-ethiopia.org/pub_files/pastoralism%20&%20Land%20full%20Book.pdf. 

Pavanello S 2009 Pastoralists vulnerability in the horn of Africa: Exploring political marginalization, donors policies and cross-border issues. Humanitarian Policy Group (HPG) Overseas Development Institute London, UK. 

Ravallion M and Bidani B 1994 “How Robust is a Poverty Profile?” The World Bank Economic Review 8 (1):75-102.

SNNPR 2005 Livelihood profile regional overview; USAID, Southern Nations Nationalities Peoples Region, Disaster Prevention and Preparedness Commission (DPPC), The Government of Ethiopia.

Swift J 2004 The state of pastoralism monitoring poverty trends in pastoral areas of East Africa and the horn initial feasibility survey: OXFAM Final report. Oxford, UK.

UN OCHA 2007 The future of pastoralism in Ethiopia; UN OCHA Pastoralist Communication Initiative, Addis Ababa, Ethiopia.

World Bank 1986 Poverty and Hunger: Issues and Options for Food Security in Developing Countries. A World Bank Policy Study, Washington, D.C.

Yacob A and Catley A 2010 Mind the gap: Commercialization, Livelihoods and Wealth Disparity in Pastoralist Areas of Ethiopia, Tufts University.


Received 18 November 2012; Accepted 19 December 2012; Published 5 February 2013

Go to top