poverty eradication and the

rural agricultural economy  -

 

Inter household differentials and determinants in

smallholder livelihood outcomes and poverty levels in

Tigray Regional State, Ethiopia

 

 

 

 

 

 

 

 

 

 

Fiona Meehan, MSc(Agr) Rural Development

School of Biology & Environmental Science,

University College Dublin (UCD)

feemee@eircom.net fiona.meehan@ucd.ie          

 

 

 

 

 

The 5th International Conference on the Ethiopian Economy, 2007

 

Contents

 

 

1          INTRODUCTION                                                                                          3

 

2          The rural household and its livelihood                                 

2.1       Household headship                                                                              3

2.2       Household size and age profile                                                  4

2.3       Education                                                                                             4

2.4       Household Labour                                                                                5

2.5       Land access                                                                                         5

 

3          Household livelihood strategies

3.1       Agricultural livelihoods                                                              6

3.1.1    Key characteristics                                                                 6

3.1.2    Cropping                                                                                 7

3.1.3    Livestock                                                                                 7

3.2       Non farm income                                                                                  9

3.3       Filling the food gap

3.3.1    Reliance on food assistance

 

4          The Household Context

4.1       Market system                                                                                      10

4.2       Credit services                                                                                      11

4.3       Extension services                                                                                 13

4.4       Mass associations                                                                                 13

4.5       Cooperatives                                                                                        14

4.6       Land rentals                                                                                          15

4.7       Cultural influences                                                                                 15

 

5          Livelihood outcomes

5.1       Measuring outcomes                                                                             18

5.2       Livelihood outcomes  determinants                                                        19

5.2.1    Household labour                                                                    20

5.2.2    Access to land, and cultivation                                               21

5.2.3    Proximity to markets                                                              22

5.3       Headship related inter household outcome differentials               24

5.4       Poverty and wealth analysis      

                        5.4.1    Characteristics of the poorest and wealthiest households     25

                        5.4.2    Influencing factors for poverty and wealth                            26

 

6          Conclusions                                                                                            27

 

 

Annexes

 

Figures  

 

Fig 1:  Headship, all households

Fig 2:  Headship, married hhs only

Fig 3:  Marital status of female heads of hhs

Fig 4:  Mean labour and dependents per hh, by headship

Fig 5:  Reasons for renting out land

Fig 6:  Oxen ownership, all hhs

Fig 7:  Distance from market measured by no. minutes walk  and % hhs

Fig 8:  Cash credit taken the previous year

Fig 9:  Credit source by % of fhhs who took credit

Fig 10:  Credit source by % other hhs taking credit

Fig 11:  Mean outcome indictors for all survey hhs, in birr

Fig 12:  Relative value of crops and cash income in direct hh income from the previous year

Fig 13:  Median livelihood outcome values by hh head

Fig 14:  median values for outcome indicators, fhhs by marital status

 

 

 

Tables

 

Table 1:    Hh size and age profile, by headship

Table 2:    Livestock ownership, % hhs owning at least one

Table 3 :   Livestock produce as source of cash income, all hhs

Table 4    Median cash earnings in birr from agricultural production  the previous year, by hh headship

Table 5:    Primary and overall sources of non agricultural cash income, by hh head, with median earnings from the previous year 

Table 6:    Membership of at least one of hh in mass associations, for all hhs and by headship

Table 7:    Gender division of labour in crop production

Table 8:   Correlation testing of outcome indicators with hh labour including persons 10 to 64 years old, and hh labour 15 to 64, for all hhs, and split by female headed and other

Table 9:   Correlations between livelihood outcomes and distance from nearest sales market

Table 10 :  t-test comparison of means, livelihood outcomes by hh head

Table 11:  1 way anova testing for marital status of female headed households and outcome indicators

Table 12  Descriptive statistics for households in the lowest quartile and in the highest quartile for value of livestock, value of crops, and cash income less crop sales.

 

 

 

 

1          INTRODUCTION

 

Ethiopia continues to be a predominantly rural country, with 84% of the population living in rural areas, the majority of whom are dependent on small scale agriculture based livelihoods.  High levels of poverty and low productivity continue to present major challenges in transformation and commercialisation of this sector, vital both for Ethiopia’s overall economic development and its current poverty reduction strategy and goals. 

 

This paper is drawn from research on smallholder livelihoods carried out over the past five years in drought prone areas of rural Tigray, the northernmost regional state in Ethiopia.  The overall aim of the research was to identify and analyse key rural livelihood influences and determinants at household level, and to explore livelihood influences and interactions between households and their socio-economic environment.  A 1,800 household survey was conducted in 20 tabias/kebelles drawn from seven of the most drought effected weredas in Tigray.  A case study of one tabia, along with farmer focus group discussions in three tabias, and a wide range of individual interviews at local and regional level provided qualitative information to complement the statistical analysis of the survey data.   In recognition of ongoing policy and strategy issues around poverty in female headed households, this paper focuses particularly on livelihood differentials between female headed and married households.

 

Findings abstracted from the resulting thesis are summarised in this paper, including:

  1. A brief profile of smallholder livelihoods

 

  1. The context within which households pursue their livelihoods is described briefly, focusing on the role of local organisations and culture

 

  1. Livelihood outcomes are measured and analysed for determinants of outcomes, and for inter household differentials in outcomes and poverty levels between female headed and other households

 

  1. Conclusions are drawn and related to policy and strategy recommendations

 

2          The rural household and its livelihood

 

2.1       Household headship

 

Given the increasing recognition that female headed households face particular pressures and challenges in relation to household livelihoods, and continuing debate on whether and to what extent female headed households are poorer than other households, household headship, and in particular female headed households were a key focus in this research.

 

Interviewees were asked who was the head of the household, and based on their responses, five categories of headship were identified.  Households headed by a woman with no male partner normally in residence, designated as female headed households throughout this paper, accounted for 28 percent of all households, and a small number, one percent, were headed by a single man.

 

Text Box: Fig 2: Headship, married hhs only

Fig 1 : Headship, all households

 

Text Box: Fig 3: Marital status of female heads of hhs Of those households headed up by a married couple, (fig 2) 69 percent described their household as male headed, 29 percent said it was jointly headed by husband and wife, who shared the responsibility, and a small proportion, 30 households or 2 percent, identified the wife as head of the household.  Given this range of headship arrangements, when the focus is on female headed households in this report, they are compared to ‘other’ households, meaning all other households, rather than ‘male headed’ households only. 

 

Female heads of household (fig 3) were predominantly widowed, 66 percent, with a further 32 percent either divorced or separated.  Only 2 percent were single, or never married.

 

2.2       Household size and age profile

Text Box: Table 1: Hh size and age profile, by headship Age ♂ Age ♀ No in hh Mn Rng Mn Rnge Mean Range All 49 18-93 40 16-91 5.0 1-12 ♀ hhs - - 48 18-91 3.3 1-10 others 49 18-93 37 16-80 5.7 1-12

Household size ranged from one to twelve, with an average of five (table 1). Household size is significantly larger in married households, with an average 5.7 household members compared to 3.3 in female headed households. 

 

There is little difference in age between female and male heads of household, 48 and 49 respectively.  Female heads of households are on average older, at 48, than the wives in married households, with a mean age of 37 years.  Anova testing showed a significant difference (sig=.000) in age among female headed households, related to marital status.  Separated women reported an average age of 39, and  divorced women 41, compared to 52 for widows.

 

2.3       Education

 

Educational access and qualifications overall were limited. The majority of respondents were illiterate, and none of the men or women of the household had been educated beyond 9th grade. The difference in literacy levels between males and females is considerable, with 53 percent illiteracy reported for the men of the household, and 87 percent among the women of the household.  Male access to church education accounts for 20 percent of this differential.  There was no significant difference in literacy levels between female heads of household and married women in other households.  There was a significant negative correlation between age and literacy, slightly stronger in the case of women (r=.242), than for men (r=-.211),  i.e. the older the men and women, were more likely they were to be illiterate. 

 

2.4       Household Labour

 

Text Box: Fig 4: Mean labour and dependents per hh, by headship In such a non mechanised, labour intensive economy, manual labour is perhaps the single most significant livelihood asset, both for working the farm and hiring out.  Average labour available per household, measured according to ILO standard labour age range from 15 to 64, was 2.53 persons.  There was a significant difference between female headed and other households (fig 4), with an average 1.92 and 2.75 labour units per household respectively. 

 

Among female headed households, those headed by divorced women had the lowest labour assets, 1.7 persons compared to 1.9 and 2.0 for separated and widowed households respectively.  This difference was statistically significant (sig=.006) between widows and divorced women’s households. 

 

2.5       Land

 

In a predominantly agricultural economy, land is clearly a base factor of production and a hugely important livelihood asset.  Almost all households, 97 percent, had access to land for cultivation.  Average landholding size was 2.7 tsimdi, and the maximum landholding reported was 10 tsimdi..  Tsimdi is a commonly used local measure of land, corresponding to approximately 0.25 of a hectare. 

 

2.5.1    Access to land

 

There was a significant difference in size of landholding between female headed and other households, 2.2 tsimdi and 2.8 tsimdi respectively (Annex 1, table 1.1).  The average size of land cultivated the previous year by the households was 3 tsimdi, approximately  .75 of a hectare, and the maximum reported was 16 tsimdi.  Smaller size of landholdings reflects the fact that most married households’ farms consist of shares of land originally allocated to each adult male and female, and part shares for children.  Households headed by women who are single or divorced will not have or will have lost the man’s share of land.  This was borne out in breaking landholding size down by marital status, and the finding that widowed and separated female headed households had mean landholdings of 2.3 and 2.1 tsimdi respectively compared to 1.7 for divorced households. 

 

There was no difference in overall rates of landlessness between female headed and other households, three percent for each category.

 

2.5.2    Land rentals

 

Text Box: Fig 5: Reasons for renting out land Land was rented in and rented out by 23 percent and 21 percent of households respectively.   Lack of oxen for ploughing was the main reason given for renting out land rather than farming it directly, reported by 72 percent of those renting out (fig 5).  Insufficient household labour was the second most common reason for renting out, reported by 22 percent of households.  When reasons for renting out are broken down by household headship, female heads of household are twice as likely to report lack of labour as their reason for renting, 26 percent compared to 13 percent of other households renting out land.

 

When the figures for households renting land in or out are broken down by headship, there is a very sharp differential, with female headed households considerably more likely to be renting out their land (49%) than other households (10%), and far less likely to rent land in, 5 percent compared to 31 percent.  This headship related difference shows up  consistently throughout all tabias, although the scale of the differential varies.   Again, marital status emerged as a highly significant factor, with 61 percent of households headed by a divorced women renting out their land compared to 50 percent of widows and only 29 percent of separated women heads. 

 

3          Household livelihood strategies

 

The main livelihood strategies pursued by households in Tigray are described below, focusing mainly as agriculture as the predominant economic activity in the region. 

 

3.1       Agricultural livelihoods

 

The agricultural sector in Tigray is mainly small scale, mixed farming, although there are some larger commercial farms, located mostly in the western part of the region.  These  provide a source of seasonal employment for farm labour throughout Tigray. 

 

3.1.1    Key characteristics

 

Land cultivation is unmechanised and highly labour intensive.  Agricultural production methods, and tools, have changed little for centuries, at least in smallholder agriculture.  Ploughing is usually carried out with a team of two oxen and a wooden, locally manufactured ploughshare, and weeding and harvesting is carried out by hand, using basic traditional hand tools The harvested grain is threshed using oxen or donkeys, or heavy stones by those without any access to livestock.  Occasionally, grain can be seen spread across the road, to be crushed by passing livestock and vehicles. The grain is then sieved and stored, sometimes in specially constructed grain stores, but more commonly in sacks or other containers within the house.  

 

Continuing heavy reliance on rainfed cultivation has two particularly important implications for rural livelihoods.  Firstly, erratic rainfall and increasing incidence of severe drought, within a context of accelerating climate change, mean that livelihoods are highly vulnerable, not just to periodic fluctuations in output, but to potential total loss of crops and livestock, and breakdown of current livelihood strategies.

 

Secondly, it means the bulk of agricultural work is traditionally tied closely into ‘normal’ rainfall  patterns, so that seasonality is a key characteristic of rural livelihoods in Tigray, with agricultural production cycle linked fluctuations in household food supplies and consumption, and in market supply, demand and prices, particularly for labour and food. 

 

3.1.2    Cropping

 

The most popular crop by far was teff,  grown as primary crop by 38 percent of households, and by 62 percent of households as one of their three main crops.  Hanfets (a mix of barley and wheat) and barley were the next two most common primary crops, 18 and 16 percent of households respectively, and grown by 26 and 34 percent of households as one of their three main types of crop. 

 

Mono cropping is practiced by a relatively small proportion of households, 16 percent overall, 26 percent of female headed and 11 percent of other households.  Diversification is still relatively limited, with 64 percent of households growing two or three different kinds of crops, and less than three percent growing more than five.   Two crops is the mode or most common number grown, but when headship is taken into account, diversification is a little wider (mode of 3) in non female headed households.

 

While cropping is clearly a major livelihood activity, practiced to some extent at least by 97 percent of households, only 21 percent ( 17 % of female headed and 22% of other hhs) reported selling any crops for cash.  Only 18 percent of mono crop farmers reported selling any of their production.

 

3.1.3    Livestock

 

Text Box: Fig 6: Oxen ownership, all hhs Oxen and poultry were the most commonly owned livestock, with 64 and 65 percent of households respectively owning at least one.  Sheep and goat rearing was not very widely practiced, 16 percent only for each. 

 

Oxen are the most important type of livestock for agricultural production in Tigray.  Oxen ownership is commonly used as a measure of household poverty, and sales of oxen in response to drought and other shocks is seen as a crucial indicator of extreme livelihood distress and unfolding food supply emergencies. 

 

Of those owning oxen (fig 6), most (34%) had one only, followed by 28 percent who own two, i.e. a full production team.  This means that even the majority of those who did have oxen still had to borrow another, or engage in labour or cash exchange to get a pair of oxen for ploughing. 

 

Text Box: Table 2: Livestock ownership, % hhs owning at least one Type of livestock All hhs Fhhs Others Oxen 64 30 77 Dairy cows 43 28 49 Bulls 20 12 24 Other cattle 22 14 25 Goats 16 5 20 Sheep 16 11 18 Donk/mules 29 10 37 Poultry 65 48 72 Beehives 13 5 16 Horses .4 0 .6 camels .5 0 .7 Breaking the figures down by headship (table 2) showed a consistently significant differential in all types of livestock ownership between female and other headed households, even for poultry and goats, traditionally seen as more accessible and manageable for women farmers than other animals, and for dairy cows, a popular component of development projects targeting female headed households.  Oxen ownership among female headed households was 30 percent compared to 77 percent of other households.  Divorced households were least likely to own oxen, 20 percent compared to 30 percent of widowed households and 40 percent of separated.  The difference was statistically significant between divorced and widowed women (sig=.040) and divorced and separated women (sig=.002).