Livestock Research for Rural Development 24 (9) 2012 Guide for preparation of papers LRRD Newsletter

Citation of this paper

The role of working animals in the livelihoods of rural communities in West Africa

S Mburu, L Zaibet, A Fall and N Ndiwa

International Livestock Research Institute (ILRI)
P.O Box 30709 (00100 Nairobi, Kenya
skmburu@cgiar.org

Abstract

In mix crop livestock systems, farmers derive their livelihoods using natural resources based strategies and off farm income. In these system livestock is both a source of food, fiber cash income and draught power valued in agricultural production and transport. Using the sustainable livelihood framework the contribution of working animals to income, food security and poverty alleviation is discussed in the context of rural West Africa. We use data collected in selected communities to test empirically the impact of ownership of working animals on poverty alleviation. Specifically, a poverty index is developed based on asset ownership and income from different sources including forest and off farm income. Then, this index is regressed on ownership of working animals and other important socio-economic determinants to provide evidence of the role of draught animals in poverty alleviation. Additional discussion also addresses the issues of constraints to the optimal contribution of working animals to the livelihoods and strategies to further improve the role of these working animals are highlighted.

Key words: Incomes, livestock, poverty index


Introduction

In West Africa, livestock plays an important role in the rural livelihoods by providing different functions, such as food, income, and other cultural and social functions.  For the average rural farmer livestock provides a buffer stock and an effective hedge against income fluctuations (Fafchamps and Czukas 1995). Livestock resources in these areas are expected to play an important role in poverty alleviation and food security. In mix crop livestock systems, farmers derive their livelihoods using natural resources based strategies and off farm income. In these systems livestock is a source of food, fiber, and cash income but provides also draught power valued in agricultural production and transport.

In West Africa, the introduction of animal traction came with the development of agricultural production in the 1960s, mainly cash crops for exports such as groundnut in Senegal and cotton in Mali (Fall et al 2003).  This technology has contributed to the intensification of agriculture and increased productivity, which are seen as key conditions to improved farmers’ income and livelihoods. In addition to traditional uses of animal work, transport and draught, animal traction allows for new cropping scenarios and agricultural production opportunities. Further, Fall et al (2003) identifies future areas where the role of working animals will be considerable: “the use of working animals for water and soil conservation techniques, the diversified use of work animals including extended use of cows, single oxen or multiple cattle pairs, and the combined use of equines and cattle where the growing season is long enough to allow timely soil preparation.” Work animals, particularly smaller animals, offer opportunities for conservation farming practices and hillside soil and water conservation, particularly important for smallholder farmers (Sims and O’Neill 2003).

Both Mali and Gambia have seen an increase in cropping areas to the detriment of forest and grazing areas.  There is increasing spread of cash crops, such as cotton in Mali, which covers 67% of cash crop area.  This is believed to have encouraged the introduction of draught animals including Zebu cattle to the region. Information on areas allocated to cotton production in Sikasso and Kayes regions show that cotton area (ha) grew from 117,641 in 1985 to 290,457 in 2000 (ILRI  2010).  Previous research shows also evidence that communities in West Africa assign high value to the functions of draught power and transportation (Zaibet et al 2010); In the Gambia, cattle was ranked first as the species contributing the most to people livelihood because of the multitude of their services, including draught power. Horses and donkeys were ranked second and third because they are the only means of transportation for many families to purchase or sell their products on local markets. Horses and donkeys are also preferred as draught power for field work because they work faster than cattle.  This paper provides further empirical evidence on the role of working animals in the livelihoods of people and poverty incidence.  

Study Objectives

We use data collected in selected communities in the Gambia and Mali in 2010 to test empirically the impact of ownership of working animals (cattle, horses and donkeys) on poverty incidence.  For this end the paper follows three steps: first, using the sustainable livelihoods framework, we construct a poverty index based on asset ownership and income from different sources including crops, livestock, and off farm income. Second, the asset based index is regressed on ownership of working animals and other important socio-economic determinants to provide evidence on the role of draught animals in poverty alleviation.  Third, the paper uses the results to discuss the issues and constraints to the optimal contribution of working animals to the livelihoods and strategies to further improve the role of these working animals are highlighted.


Methodology

The sustainable livelihoods framework (Freeman et al 2004) is suitable methodology to understand the role of working animals in local communities in West Africa. According to the sustainable livelihoods framework, livelihood strategies defined as the activities that households engage in to make a living are a function of the assets to which the household has access, given the broader socio-political and agro-ecological context.  Household assets are defined according to five categories: human, social, physical, natural and financial (Adato et al 2007). Different livelihood strategies require different endowments of the different assets. Household asset endowments, along with their preferences, determine the strategies in which they can engage.

In this paper, guided by the sustainable livelihoods framework we develop a poverty index based on core assets in possession of households and we used it to divide the population into three to four categories of poverty groups. These groups are then characterized based on their assets and livelihood strategies. Finally, we investigate the role of working animals as key determinant of poverty by specifying and estimating a Tobit regression model. 

Sampling and data collection

Data used in the analysis were collected using two types of surveys, community or participatory rural appraisal (PRA) and household surveys. Table 1 below gives a summary of the sampling plan for the two types of surveys. The data used to construct the poverty index were mainly from the household survey with data from the PRA being used in the classification of the household data into the poverty categories. The PRA provided the community perception of poverty criteria as well as the proportion of poverty categories. Household survey data covered all assets belonging to households, in addition to household characteristics, such as age of household head, education, and household size. Data on income sources were aggregated to generate farm and household income levels.

Table1. Data sampling strategy in Mali and Gambia

Sampling method

Hierarchies

Sample size per level

Sampling Method (within level)

Sampling Frame

PRA

Sites

3 Pre-defined (Nianija, Kiang West and Niamina East in the Gambia and Manankoro, Sagabary and Madina Diassa in Mali). 

Villages

3

1 in each village size category

List of villages selected for the Household survey

Participants

30 – 35 gender balanced, aged 18 – 65 years, includes: livestock owners (with variation in herd/flock sizes), herders and non-livestock owners

Household survey

Sites

3 Pre-defined (as above)

Villages

9 – 10

Random

List of villages

Households

5 – 15*

Random

List of independent households in village

* depended on village size

Poverty index in the Gambia and Mali

Poverty, or socio-economic status, has been traditionally measured based on income or consumption expenditure (Mukherjee and Benson 2003; Coulombe and Mckay 1996). Income measures do not capture non monetary values such as income in kind (home consumption) or other non market values which is quite important in the case of the rural communities in West Africa (Zaibet et al 2010).  More recently there has been development of asset based poverty index (ABPI) (Filmer and Pritchett 2001; Sahn and Stifel 2003). The ABPI is more in line with the Sustainable livelihood framework.  Rather than income or expenditure, the ABPI requires data on household asset ownership and access, which is relatively easier to collect than income via household surveys.

The construction of the ABPI is based on an aggregation of a range of variables to derive a measure of poverty.  Principal component analysis (PCA) has been suggested as appropriate to this purpose (Vyas and Kumaranayake 2006; Krefis et al 2010).  PCA is used to reduce the number of original variables into smaller number of uncorrelated variables called principal components. Typically, the first principal component, i.e. the one with the highest variance (and eigenvalue), has the meaning of the measure of size, which could be interpreted as socio-economic status (Kolenikov and Angeles 2009; Vyas and Kumaranayake 2006).  The ABPI is therefore defined as follows:

The weights, ai, correspond to the eigenvectors of the correlation matrix given by the PCA. The higher the weight, the better is the corresponding variable as discriminating factor in PCA. We used principal component procedure in SAS software to analyze the data and obtain these indices.

We used data from household surveys as described above and selected variables with three categories; variables of structure (physical assets) such as land and livestock ownership; variables of socio-economic characteristics such as education, gender, household size; and variables of management, e.g. disease prevention and treatment, as a proxy for service/market access. These variables were then transformed in categories and used in the PCA. The first principal component for each case (The Gambia and Mali) was used to compute the poverty index for all the households (Figure 1).  Other studies use arbitrary cut-off points to classify poverty categories (Vyas and Kumaranayake 2006). The cut-off points assume the poverty index is uniformly distributed; which is not always the case.  Here the classification was based on the wealth ranking during the PRAs.  Such distribution is based on the perception of communities of poverty distribution.

The households were then grouped (based on PRA workshops in the two countries) into four poverty levels in the Gambia; very poor, poor, moderate poor and rich; and three in the case of Mali; very poor, poor, and rich.  Results are presented in Figure 1.

Figure 1.  Poverty index scores and distribution in the Gambia and Mali

According to Figure 1, the poverty index scores are almost normally distributed in the case of Mali (mean=0.31 and std deviation = 1.03) whereas the distribution in the case of Gambia showed more skewed curve (mean = 0.16 and std deviation = 0.8).  The three poverty categories in Mali have mean scores of -1.04, 0.25 and 1.68 respectively for the very poor, poor and rich. In the case of the Gambia these mean scores are -0.9, 0.11, 1.1 and 1.7 respectively for the very poor, the poor, the rich and very rich.  The form of the histograms in the two countries (normality and skewedness) shows that community perceptions are in accordance with the computed poverty scores. 

Livelihood strategies in the Gambia and Mali

Livelihoods are very diverse, although the most important are crop income and livestock income in both countries. Crop income in Mali is however dominant (81%). These activities require ownership and access to assets, i.e. land and farm implements, livestock and labour (human assets). Livestock is an important asset for integrated crop-livestock production by providing the required power to farming activities. In the Gambia there seem to be more diversification of the sources of income; business-services, remittances, and trading are contributing more than 5% each (Table 2).  

Table 2.  Livelihood strategies in the Gambia and Mali

Livelihood strategies

Gambia

Mali

Source of income

Mean (Dalassi)

%

Mean (F CFA)

%

Business-Services

1,929

0.07

1,106

0.00

Trade in non-livestock agricultural products

 

0.00

7,181

0.01

Field crops and fodder

9,869

0.36

1,030,227

0.81

Formal salaried employment

2,563

0.09

3,037

0.00

Fruit production

115

0.00

48,072

0.04

Gardening / vegetable production

907

0.03

28,305

0.02

Hired labourer

49

0.00

 

0.00

Hunting, forest and fishing

919

0.03

7,953

0.01

Livestock herding

253

0.01

3,638

0.00

Livestock income

3,963

0.15

79,086

0.06

Remittances

2,161

0.08

 

0.00

Rent out land / sharecropping

36

0.00

755

0.00

Sale of poultry & poultry products

64

0.00

10,202

0.01

Trade non-agriculture products.

1,449

0.05

5,722

0.00

Trading in fish

25

0.00

0.00

0.00

Trading in agricultural products

1,194

0.04

24,507

0.02

Trading in livestock products

313

0.01

4,829

0.00

Trading of livestock

1,236

0.05

14,805

0.01

Working on other farms

120

0.00

3,500

0.00

Looking at the sources of income by poverty categories (Table 3), we see also that the main source of livelihoods across the different households was field crop production both for sale and consumption. The trend was similar across the different income groups. The rich households also earned from formal salaried labour as well as selling their own cattle and trading in livestock. The poor households however earned from trade in non-agricultural products such as firewood, water; working on other farms, fishing and remittances.  

Table 3. Livelihood strategies by social category in the Gambia and Mali

 

Gambia(Dalassi)

Mali (CFA)

Mean Income

Very poor

Poor

Moderate rich

Rich

Mean Total

%

Very poor

Poor

Rich

Mean Total

%

Crop income

8,671

32,234

35,680

54,424

29,980

0.58

655,742

880,987

1,665,903

1,027,591

0.78

Livestock income

72

1,438

18,018

25,280

5,704

0.11

34,717

54,334

164,053

77,653

0.06

Off-farm income

10,184

11,324

15,393

17,307

12,196

0.24

64,607

91,424

210,533

115,424

0.09

Gardening &

fruits

837

1,249

973

893

1,081

0.02

66,548

52,926

131,167

76,129

0.06

Forest income

1,559

3,600

1,800

1,000

2,511

0.05

19,400

12,357

36,717

20,248

0.02

Total

21323

49845

71864

98904

51,472

1.00

841014

1,000,000

2,000,000

1317045

1.00

The role of working animals to livelihoods

Working animals play a fundamental role in livelihoods improvement as they provide farm power and contribute to food security and poverty reduction, income generation and gender equity. Smallholders who use animals for soil tillage can cultivate larger areas more efficiently and quickly than with human labour, thereby greatly increasing their yields. Working animals create synergy in nutrient cycles, farming and marketing systems: animals allow farmers and traders to transport harvests, market products, fodder for livestock and manure. They increase people’s transport capacity and range, providing families and entrepreneurs with access to supplies, services and livelihoods. Working animals are multipurpose, producing profitable livestock products, including manure, milk and sometimes meat. Human, animal and mechanical power are not mutually exclusive and each has advantages depending on the environment, scale and socio-economic context. People aspire to prestigious, modern machines but tractors may be unaffordable and inappropriate on small farms or in difficult terrain (FAO 2011). 

Results shows that the rich categories are in general better endowed with assets than the poor, mainly in livestock and land assets. Working animals in particular are important since access to and productivity of land depends largely on possession of farm implements and, in particular ownership of working animals. Working animals as shown in Table 4 are defined by four categories: adult castrated males for draught, adult intact males for breeding and draught, donkeys and horses. The poor and very poor don’t have working animals in the Gambia, but these categories may typically have one donkey and an adult castrated bull in Mali. The moderately rich in the Gambia have on average one adult intact bull and a horse. The rich households however would have 1 to 5 bulls and 1 to 2 donkeys or horses. About 46.6 percent of households in The Gambia owned working animals while in Mali about 72.8 percent owned working animals.

Table 4. Mean number of working animals by household categories in the Gambia and Mali

 

Gambia

Mali

Working Animals

very poor

poor

Moderately rich

Rich

 

very poor

poor

rich

Adult Castrated Males for

draught

0.04

0.22

0.11

1.40

 

0.97

1.57

5.04

Adult intact males for

breeding & draught

.00

0.13

0.81

2.80

 

0.13

0.21

1.04

Donkeys

.00

.00

0.14

0.00

 

0.49

0.97

1.75

Horses

0.27

0.56

0.86

1.15

 

0.06

0.05

0.00

Table 5 displays the levels of income by source with respect to ownership of working animals. Looking at the p-values, which are used to show whether these differences are significant, we can notice that ownership of working animals have a significant effect on crop income in both countries, and also have significant effect on forest income in the Gambia and off-farm income in Mali.    

 

Table 5. Sources and levels of income by ownership of working animals

 

Gambia (Dalassi)

Mali (CFA)

 

Own*

Don’t own

P-value

Own*

Don’t own

P-value

Crop income

37,819

22,355

.028

1,205,796

550,890

.001

Livestock income

9,381

2,489

.019

99,954.22

17,998

.000

Off-farm income

13,469

11,082

.252

136,613

58,743

.001

Forest income

3,625

1,644

.037

22,161.33

15,131

.220

% of HHs

46.60

53.40

 

72.80

27.20

 

*Owns at least one type of draught animal ( intact or castrated males, horses, donkeys)

 

Figure 2. Levels of income by ownership of working animals in The Gambia  

Households in The Gambia with working animals had higher crop, livestock, off-farm and forest incomes compared to households without working animals.

 

Figure 3. Levels of income by ownership of working animals in Mali

Similarly in Mali, households with working animals had high crop, livestock, off-farm and forest incomes compared to households without working animals. 

Further, we used a Tobit model to determine the determinants of poverty in the Gambia and Mali with a focus on the role of working animals. The dependent variable defined in the model is the poverty index as described above and the explanatory variables were chosen among exogenous variables which are not included in the PCA. Table 6 presents the results of the Tobit estimation; where the two sets of variables (working animals and socio-economic variables) are displayed. Socio-economic characteristics which have significant effect on poverty incidence are found to be gender and household size in the Gambia and only household size in Mali. In particular the age of household head and membership to group associations were not significant. The results confirm also the role of working animals; in Mali both cattle draught animals and donkeys were significant whereas in the Gambia horses were the most significant among other categories. The fact that cattle draught animals have significant estimates in the Mali model but not in the case of Gambia could be explained by the existence of zebu breeds in the cotton zone of Mali and/or the (large) format of the NDama cattle in these regions. Knowing that in the Gambia almost 100% of cattle is NDama, this explains why people prefer horses to cattle as working animals and explains the significant role of these animals in poverty incidence.  In more general terms, the gender dimension is very important for poverty as shown by these results but also for the use and ownership of working animals.

 

Table 6. Tobit results on the determinants of poverty in the Gambia and Mali

 

The Gambia

Mali

Poverty Index

Coefficient

Std. Err.

t-statistic

 

Coefficient

Std. Err.

t-statistic

Gender of head (1=male)

0.40

0.19

2.05**

 

 

 

 

Household size

0.06

0.01

6.26***

 

0.03

0.01

4.68***

Age of head

0.00

0.00

0.34

 

0.00

0.00

-0.88

Group membership (1=yes)

0.01

0.01

1.34

 

0.12

0.13

0.98

Draught animals (numbers)

0.05

0.05

1.05

 

0.09

0.02

5.26***

Donkeys (numbers)

0.11

0.15

0.75

 

0.12

0.05

2.54**

Horses (numbers)

0.09

0.05

1.72*

 

-0.04

0.11

-0.33

Constant

-0.98

0.25

-3.93

 

-0.44

0.23

-1.94

/sigma

0.72

0.03

 

 

0.88

0.04

 

LR chi2(6)

75.49

 

 

 

94.42

 

 

Prob > chi2

0.00

 

 

 

0.00

 

 

Number of observations

236

 

 

 

294

 

 


Conclusions


References

Coulombe H and A. Mckay 1996 Modeling determinants of poverty in Mauritania. World Development. 6(24):1015-1031.

Fafchamps M, Udry C, and Czukas K 1998 Drought and Saving in West Africa: Are Livestock a Buffer Stock? SSRN Electronic Journal, 55, 273-305. doi: 10.2139/ssrn.71374.

FAO 1997 The national agricultural research system of The Gambia: analysis and strategy for the long term. Final report. Field Document TCP/GAM/6611/FAO, 193pp.

Filmer D and Pritchett L 2001 Estimating wealth effects without expenditure data—or tears: An application of educational enrollment in states of India. Demography, 38(1), 115–132.

Freeman A F and Allison E 2004 Livelihoods and rural poverty reduction in Kenya. Development policy review. 22(2): 147-171.

Itty P, Ankers P, Zinsstag J, Trawally S and Pfister K 1997 Productivity and Profitability of Sheep Production in The Gambia: Implications for Livestock Development in West Africa. Quarterly Journal of International Agriculture 36 153-172

Kolenikov S and Angeles G 2009 "Socioeconomic Status Measurement With Discrete Proxy Variables: Is Principal Component Analysis A Reliable Answer?," Review of Income and Wealth, Blackwell Publishing, vol. 55(1), pages 128-165, 03

Krefis A C , Schwarz N G, Nkrumah  B,  Acquah  S, Loag W, Sarpong N, Adu-Sarkodie Y,  Ranft U and May J 2010 Principal component analysis of socioeconomic factors and their association with malaria in children from the Ashanti region, Ghana. Malaria Journal. 9:201.

Mukherjee S and Benson T 2003 The determinants of poverty in Malawi, 1998. World Development 2(31): 339-358.

Sahn D E and Stifel D C 2000  Poverty comparisons over time and across countries in Africa. World Development, 28(12), 2123–2155.

Sahn D E and Stifel D C 2003  Urban–rural inequality in living standards in Africa. Journal of African Economies, 12, 564–597.

Vyas S and Kumaranayake L 2006 Constructing socio-economic status indices: how to use principal components analysis. The London school of Hygiene and tropical medicine, Oxford University Press. UK. www.fao.org/ag/againfo/themes/animal-welfare/blog/en/)

Zaibet L, Traore S, Ayantunde A, Marshall K, Johnson N and  Siegmund-Schultze M 2010 Livelihood strategies in endemic livestock production systems in sub-humid zone of West Africa: trends, trade-offs and implications.  Journal of Environment, Development and Sustainability. 13(1)96-105.


Received 27 July 2012; Accepted 31 July 2012; Published 3 September 2012

Go to top