1. Introduction
Change in natural
environmental conditions has constantly influenced pastoral livelihoods in the
Afar region of
In this
situation, livestock herders increasingly pursue non-pastoral income strategies
to meet consumption needs and prepare against risky shocks such as drought
(Little et al. 2001). Studies in diversification strategies (e.g. Holtzmann
1996, Kituyi 1990, Little 1992, Zaal and Dietz 1999) show that marked change
and diversification is still discussed as a two-sided coin, which may either
allow herders to better cope with high levels of risk or may exaggerate their
problems. Cultivation is a major avenue of diversification and is seen by some
as a viable risk management strategy (Campbell 1984, Smith 1998), while others
view it as an unsustainable or even destructive option that accentuates the
risks pastoralists face (Hogg 1988). Fratkin
(1991) and Nathan et al (1996) show the potentially negative ecological and
social effects of pastoral sedentarization and diversification. Yet, for Holtzmann
(1996), diversification is seen as a cyclical rather than a linear process,
whereby herders combine different income strategies at different points in
their life-cycle. Equally, income diversification strategies such as farming
among pastoralists do not necessarily lead to a diminished interest in
livestock investments and production (Little et al. 2001). In this paper we
will focus on crop production from a dual perspective: first looking at the
historically coercive way of state intervention and, second, at the current
means of responding to natural calamities. As we will demonstrate, there is
considerable difference within pastoral communities in motivations for
diversification, predominantly along lines of wealth and gender.
Property right changes having to do with
the evolving relationship between pastoralists and agriculture are at the center
of this analysis, which looks into two cases related to agricultural production
systems and Afar pastoralists. One case portrays the conflictive transformation
of the traditional land use arrangements of Afar pastoralists, which came about
due to the coercive intervention of the state in implementing projects
associated with commercial farming, while the other shows the more or less
collective adaptation to farming, as induced by recurrent droughts in the
presence of small-scale and supportive state intervention. Indeed, the two
cases show that pastoralism is under pressure arising from both policy-related
and natural challenges.
Natural
resource degradation and poverty in rural
2. Study
sites and Methods
The Afar region extends from
central to northeastern
Figure 1: Location of Afar Region and Study Districts
Source: Afar Region Administration
(2005)
The dominant source of livelihoods in the study areas is pastoralism, with limited levels of crop cultivation and other activities. Afar pastoralists raise mixed species of primary livestock, including camels and cattle, and keep supplementary herds of goats and sheep, usually for commercial purposes. They manage their livestock under an extensive mobile system, with natural pasturage being the main source of livestock feed.
To investigate both historical and recent
changes in the traditional property rights of Afar pastoralists, we pursued two
different kinds of data sources, namely primary and secondary, and employed various procedures for
data collection. Section five is mainly based on secondary data,
including several unpublished documents accessed from the Middle Awash
Agricultural Development Enterprise (MAADE), the
Section 6 is mainly based on the data collected from 180 pastoral households, dwelling in six purposively selected sites namely: Ambash and Qurqura in Amibara district, Doho and Dudub in Awash-Fentale district, and Harihamo and Daleti in Semu-Robi district. A two-stage procedure was used to select the sample households. First, using lists of household heads in each site (generated for the purposes of this study), with the help of the local elders pastoral households were stratified into three groups: poor, medium income, and better off. Thereafter, ten households were selected from each stratum using systematic random sampling technique. In most cases, household heads (usually male) were interviewed, though in a few cases responses were taken from an adult family member who was not the head. A group of trained enumerators conducted the interviews with individual sample households, guided by a structured questionnaire prepared for this purpose.
The overall data collection process encompassed two phases. The first phase (December 2004 – May 2005) involved several tasks, including implementation of the household survey, collection of secondary data, and collection of detailed qualitative data though group and key-informant interviews. The second phase (October 2006) was organized for a short period in order to strengthen the evidence gathered from the first phase by reviewing secondary sources and conducting expert interviews.
3. Traditional
institutional arrangements
The clan is the lowest and de facto unit of traditional administration in Afar, although there are also smaller social units, such as the dahla or sub-clan. As Getachew notes each clan comprises “a group of people related to each other by decent, living within shared territory and sharing common rituals and political leadership” (2001:54). Each clan has a well-established gerontocracy, whereby decision-making power regarding land and other natural resources resides within the clan council, consisting of the clan leader, elders, the feima[1] and local wise men.
Each clan manages its resources collectively, based on customary principles. Accordingly, herd management follows rotational grazing patterns. When rainfall is normal for successive seasons, clan members are instructed not to use reserved pasture areas. These areas are made accessible to the members only after other areas have been exhaustively used. Although each clan member has an inalienable use right over the resources, intra-clan customary laws (or operational rules) regulate these use rights.
The traditional institutions of the Afar
allow two types of resource users. The first category includes clan members who
use the rangeland permanently. They are primary right-holders (waamo)
who have not only the right to use the resources on the rangeland but also to
exclude others and to transfer to their heirs. The
second type of resource users comprises groups of neighboring pastoralists
whose demands for pastoral resources go beyond their own endowments,
particularly during drought years. These groups are secondary right-holders.
They can be termed “right-holders” because they have frequent access to clan
resources that is generally recognized and accepted by clan members and traditional
leaders. However, certain obligations are operational on secondary
right-holders in order to obtain access to the resources. Ex ante negotiation is required with waamo right-holders, the success of which depends upon the
relationship between the two groups and resource conditions. If they are
allowed access, secondary right-holders are required to honor the customary
rules of the host group. For instance, they should refrain from actions such as
cutting trees, allowing other herders to use the resources and rushing their
livestock into reserved areas.
4. Coercive
ways of property rights change: The state subverting the commons
4.1. Triggers and Processes of
Coercive Change
The intervention of the state in Afar was very limited prior to the 1960s. Farming was limited to the lower Awash flood-fed plains, where some pastoralists in the Asahimarra section of Afar had been practicing mixed crop-livestock farming for generations (Getachew 2001). However, following the 1960s state interventions in these areas have increased, mainly for two reasons. First, the Afar plains - specifically areas in the middle Awash valley - were found to have great potential for wide-scale irrigated farming. The most attractive feature of these areas was their suitability for cotton production, which was critically important for expanding the country’s textile industries: a primary focus of the first and the second five-year national development plans (IGE 1957; 1962). Second, pastoralism was not accepted as a livelihood strategy within the reigning national political mindset of the time. Rather, pastoralism was considered to be a primitive and nonviable way of life − to be avoided rather than preserved (Abdulahi 2004; Getachew 2001). Thus, the intention of the policy makers was to change this mobile mode of life towards sedentary farming. However, the pastoralists neither participated in the decision-making process nor were they convinced about the goal of change.
In 1962, the Awash Valley Authority (AVA)
was established by decree as an agent of institutional change. AVA was
responsible for undertaking several activities, such as the founding and
management of state farms, coordination and financing of pastoral settlements
and other schemes, and monitoring the overall transformation process, for which
some 70,000 ha of dry-season rangeland was targeted (Getachew 2001). AVA had
direct military and financial support from the government to implement the
planned changes, using its military power, for example, to threaten the
pastoralists. The Middle Awash Agricultural Development Enterprise (MAADE)
began operations on the expropriated rangeland with the main objective of
satisfying the demand of domestic textile industries for cotton. Initially, it
had an operating area of 300 ha, which was increased to 13,116 ha in 1985. In
addition to MAADE, several pastoral development schemes were implemented with
directives coming from AVA. These included collective settlement farms and
irrigated pastures[2]. The costs to cultivate
the settlement farms were covered by the state while the pastoralists
contributed nothing except their labor. The output of the settlement farms was
distributed among registered households.
The implementation of the state-driven projects resulted in a mixture of property rights in the area. Firstly, by using its coercive power the state became a de facto owner of part of the land over which the pastoralists had had inalienable rights for generations. Secondly, the introduction of the collective settlement farms brought a new variant of common property, apart from the traditional communal ownership of the rangeland. Indeed, the non-riverine parts of the area remained under the control of the pastoralists and were entirely allocated for livestock grazing, whereas traditional rights were nullified by order from the state in the riverine sites. This implies that the intervention of the state created a “legal dualism”: Claims over the riverine sites were governed and protected by statutory laws, whereas the non-riverine sites remained outside of direct state protection and legitimacy.
Because the state, by the power vested in it, redefined the land use rules without consulting the pastoralists, the process of change was not smooth, with the pastoralists resisting every action of the state. Indeed, throughout the 1980s and 1990s Afar pastoralists put great pressure on the administration of the state farms[3]. The pastoralists expressed their dissatisfaction with and opposition to the implementation of the commercial farm schemes mainly by damaging mature crops in the field, a typical example being the recurrent damage caused by local people on banana plantations, which eventually forced the state farms to abandon banana production. Initially, the state farms allocated compensatory funds to be paid to clan leaders and elders in the form of employment benefits which would, it was hoped, ameliorate the dissatisfied pastoralists. This reward system did not put an end to the grievances, however, as the power of the pastoralists emanate from their great number, which was increasing over time.
In the course of time, the relative power of the two actors has changed
in favor of the pastoralists. At the beginning, AVA had the power of mobilizing
resources to constrain the choices of the pastoralists and was capable of
controlling their actions. However, it couldn’t maintain this power to
continuously influence the choices and actions of its counterparts. This is
partly attributable to the decline of attention paid by the government towards
state farms after 1989. Especially after the economic reform of 1991, the stake
of the state in business ventures dramatically declined. As a result, AVA did
not receive enough financial, political and other supports from the government
to maintain its power. In addition, the shift in the national political
structure towards ethnic-based federalism and the concomitant establishment of
the
These changes had effects on the existing property rights and land use arrangements. With the efforts of the Afar regional government and the decision of the Transitional Government of Ethiopia, MAADE handed over a significant part of its land, including irrigation infrastructure and facilities, to the Afar[4] in 1993. This, in turn, resulted in the existence of two distinct forms of property relations, consequently increasing the number of actors involved. First, the pastoralists subdivided part of the returned farmland and started private farming in collaboration with highlanders, implying the individualization of the traditional communal rangeland. Second, the pastoralists leased-out part of the returned land to local investors, whereby the latter annually transfer cash payments to the pastoralists[5], implying the introduction of a lease contract regime into the area.
In general, this
sub-section shows that the state is the major source of property right changes
in the middle Awash
4.2.
Impacts of coercive change on the livelihoods of pastoralists
Direct intervention of the state has, step by step, changed the traditional property regime of the pastoralists and brought about new forms of land use arrangements that have direct implications for their livelihoods. Four distinct forms of land use arrangements have been realized since the initial interventions of the state, namely: state farms, settlement farms, individual small farms and private large-scale farms. These new variants of property rights have one main feature in common: they are all related to the production of crops. However, each of them is unique in terms of the types of actors interacting with pastoralists and the impacts on rights and capabilities of pastoralists to secure livelihoods that they entail. The existence of state farms implies de facto state ownership as well as the nullification of customary rights which pastoralists had had over land for generations. Indeed, the contemporary rights that pastoralists have over this portion of the former commons have been limited to use rights over crop residues, and only with the consent of officials from the state farms. On the other hand, the expropriation of large tracts of dry season rangeland, without compensation, has resulted in the reduction of the capability of pastoralists to secure livelihoods through the traditional means of livestock production. In this respect, the present vulnerability of Afar pastoralists to recurrent droughts is at least partly associated with such expropriatory measures of the state (Sen 1981; Getachew 2001; Yemane 2003).
The settlement farms
(established for compensatory reasons reflect a kind of interaction between the
state and the pastoralists. In this case, the new resources necessary to produce crops were entirely supplied by
the state. The existing irrigation infrastructure and the road networks were
built by the state through a large outlay. Similarly, farm machinery and
facilities were purchased by the state. The technical personnel and the
management staff had also been installed through the efforts of the state.
While these resources defined capabilities to exercise rights within the
parameters of the new land use system, pastoralists already had well-recognized
rights to the benefit streams from the land. In other words, they had the
rights[6]
as well as the capabilities to generate benefits from the settlement farms.
However, the state was not “benevolent” forever, but rather stopped its support
in the mid-1980s. The termination of state support and the concomitant transfer
of all machinery and facilities to the state farms have debilitated the
capability of the pastoralists to extract benefits from their land, although
their rights to the land have remained intact. Lacking the knowledge and
physical resources needed for farming, the pastoralists have not been able to
continue crop production on the former settlement farms, despite their rights
to do so. As a result, the entire settlement farm has been out of production
and is covered, at present, by an inedible exotic weed (Prosopis juliflora). In fact, this part of the former rangeland is neither cultivated nor
is it efficiently used for livestock production, which has direct implications
for the livelihoods of the pastoralists.
The return of the confiscated land in 1993 was an important action that reduced the influence of the state on the traditional lands of the pastoralists. Actually, the pastoralists were free to decide on what to do with the returned land. Accordingly, the land was partly allocated to clan members and was partly leased out to local investors. In regard to individual parcels of land, the Afar have established partnerships with agriculturalists from the highlands. Individual landowners have the right to choose their partners, define and redefine the land use contracts, and terminate contracts if required. In the lease arrangements, the new partners of the pastoralists are local investors. Under this form of contract, the pastoralists collectively earn 30 percent of the investors’ profits in return for the use of their land, which they distribute among themselves based on predefined criteria. They have formed a standing committee, including an accountant, to monitor all transactions of the investors. The committee has been entrusted to defend the rights of its principals and, hence, to take action when errors or other problems arise.
While the current situation shows the restoration of the rights of the pastoralists over their traditional land, capability limitations are apparent in terms of maximally exploiting the new venture. First, pastoralists have poor knowledge of farming techniques and lack resources (e.g. farm implements) necessary to cultivate crops. As a result, the highlanders are responsible for all farm operations in return of larger shares of the net farm proceeds (up to 70 per cent), whereas the contributions and earnings of the pastoralists are minimal. Actually, the share of the highlanders reflects the costs to be paid by the pastoralists due to their limited capabilities to produce crops on their own. Second, the capacity of the committee to actually carry out their responsibilities concerning the lease arrangements is questionable. The members have no accounting knowledge and some of them do not even know how to read and write. Hence, everything is done based on trust, implying the possibility that the pastoralists could be cheated if the investors desire to do so. Again, this implies the weak position of the pastoralists under such arrangements.
It is also worthwhile to pinpoint the distributional effects of the changes in property rights that have taken place. Traditional property rights allowed multitudes of users to share a resource system in accordance with certain predefined rules. Under the traditional arrangements, all clan members had equal rights to grazing resources and, hence, could extract benefits, provided that they had livestock. However, equality in rights to the communal heritage has not been ensured following the state-induced changes of property rights. During the initial period of the transformation, elites and their allies abandoned the customary rules and facilitated their own entitlement to the benefits from the settlement farms. Others used their physical fitness and connections with project leaders to secure their own benefits, while those households lacking such resources were denied access to them (Getachew 2001). The procedures following the subdivision of the newly returned land has also not been immune to discrimination. Contrary to the traditional land law, about 31 percent of the sample households were left out of consideration during the subdivision. A closer look at the assets of the sample pastoralists chosen for this study shows that those who have not been benefiting from the subdivided land are poorer (average 0.89 TLU of per capita livestock asset) as compared to those who have been benefiting (2.91TLU). This inequity and mistreatment is even more visible with regard to the women. ‘Women-headed’ households were neither considered when the returned land was distributed among clan members nor have they been beneficiaries from the leased-out land because of tradition-based criteria: women are de facto minors in Afar customary laws[7].
5. Non-coercive ways of
property rights changes: Voluntary adoption of farming
5.1. Triggers of voluntary
change
Afar pastoralists in the
study areas have been highly threatened not only by the coercive actions
of the state, but also by recurrent droughts. Two major droughts hit the areas
since the mid-1990s, and short dry spells are common as well. The prevalence of
drought has adversely affected the pastoral economy in two ways. First, it has
reduced the total livestock assets and productive capacities of the area,
thereby increasing mortality and morbidity rates. Sanford and Habtu (2000,
cited in Mesfin 2003:44) have estimated that a 5 to 15% percent reduction in
livestock assets occurred in Afar due to the drought of 1999/00. In fact, this
estimation corresponds to the best-case scenario. Under the worst-case scenario,
livestock loss has been estimated to range from 15 to 45 per cent. Emergency
assessment reports of various development organizations and relief agencies
indicate that the prolonged drought of 2002/03 had even more serious
consequences for the Afar pastoralists (FEWS NET 2002; UN-EUE 2002a; UN-EUE 2002b).
Second, the successive
droughts have re-calibrated the terms of trade against the pastoralists.
Although no systematic records have been found yet, assessment reports of aid
agencies indicate a sharp decline of livestock prices during the droughts. A UN
assessment mission in the area indicated that pastoralists faced more than 50 %
reduction in livestock prices following the drought of 1999/2000 (UN-EUE 2000).
Similarly, livestock prices fell by 50 % to 60 % due to the drought of 2002,
while maize prices simultaneously rose by about 235 percent (Davies and Bennett
2007). The adverse effects of the droughts on the terms of trade were
compounded by other factors, such as export restrictions imposed by Saudi
Arabia in September 2000, following a Rift Valley fever outbreak, and
insecurity around the northern border of the Afar region in the aftermath of
the war between Ethiopia and Eritrea in 1998.
These livestock losses coupled with the
deteriorating terms of trade against pastoralists worsened food insecurity in
the study areas, with the degree of food insecurity reaching its climax in
2002/03 because of the intensified drought. A serious famine hit the area,
during which a large number of pastoralists lacked anything to eat. On 12 July
2002, the Disaster Prevention and Preparedness Commission issued a Special
Alert that publicized the deterioration of food security in several parts
of the country, particularly in the Afar region and the neighboring
The deterioration of food security in pastoral areas in general and Afar in particular necessitated an intensified intervention of external agents (governmental and non-governmental organizations) into pastoral livelihoods. While the most immediate external intervention was provision of food aid to save human lives, a number of programs and projects financed by the government and NGOs, such as FAO, Farm-Africa, CARE-Ethiopia, and Oxfam GB, were designed to improve the livelihoods of pastoralists. One intervention was focused on designing projects and programs to facilitate the expansion of crop cultivation in these areas.
Both traditional authorities and external agents were important facilitators of collective action to begin farming. In this respect, external agents (local government and NGOs) sponsored meetings at the kebele level. While there exist no formal records on the number of local meetings in the study sites, the average number of meetings reported by the sample households ranges between 7.2 (for Dudub site) and 18.6 (for Daleti site) for the year preceding the survey. During the meetings, the external agents explained their visions and commitment toward improving the livelihoods of pastoralists, mainly through programs targeted on farming. The interventions of the external actors were even more direct in three of the study sites, namely Harihamo, Daleti and Doho. In Harihamo and Daleti, the government directly supported collective activities in relation to farming through its food security program. Assistance included provision of farm tools, covering initial costs of farm operations (e.g. costs of tractor for tillage), provision of oxen, and other logistic and advisory support. At the Doho site, support was mainly provided by an FAO livestock recovery project office at Awash-Fentale which provided financial support for initial development of irrigation infrastructure and farm inputs, mainly seeds. Moreover, district level experts on agriculture were responsible for providing advisory support to the “agro-pastoralists”.
Similarly, the role of
traditional authorities was substantial. Specifically, activities such as mobilizing clan members
for meetings; organizing and supervising all activities, such as bush clearing
and land levelling; and imposing sanctions on free riders required the active
participation of the feima members.
Traditional sanctions were to be applied, including asset penalties, like
slaughtering the breeding cows of free riders, and corporal punishment, such as
beating free riders in public to shame them[8].
The pastoralists were required to be involved in all activities to prepare the communal land for cultivation, following which it was allocated to the participants. The preparatory activities were done intermittently for about 4 months in Semu-Robi and for 2 months in Awash-Fentale. For Amibara, the exact duration is not clear, but according to sample respondents it ranged between 30 and 180 days. The overall participation rates across districts in these cooperative activities were 39.1 % (n=70) with 13.3 % (n=8) in Amibara, 23.3% (n=14) in Awash-Fentale and 81.4 % (n=48) in Semu-Robi.
5.2. Analytical model and variables
In this case,
farming is an enterprise that has been induced because of natural shock to the
area. Understanding the movement of pastoralists towards farming entails
comparison between the situation under farming and pastoralism. Let Ui1
and Ui0 be the utilities of individual 
associated with farming and pastoralism, respectively. We
expected that community members would be heterogeneous in terms of the level of
utilities generated from farming. We also expected that community members would
vary in terms of the level of utilities they generate from pastoralism. Thus, Ui1
and Ui0 can be formulated as a function of other
variables such that 
and 
, where a and
bi
are parameter estimates and Xi is a
vector of exogenous variables that cause heterogeneity among community
members. As a utility maximizer, individual decides in favor of farming if 
and otherwise if 
[9].
Accordingly, participation
in collective activities to start farming reveals that 
. If we replace 
by 
and 
by 
for brevity, then the probability that individual will participate in
collective action to start farming can be specified as:
. If a normal distribution function is assumed for , then the model turns out to be a probit model
(Amemiya 1981). Alternatively, if a logistic distribution is assumed, the model
becomes the logit one (ibid). The two
alternative models produce similar outputs, except in rare cases when the data
concentrates around the tails of the distributions (Amemya 1981; Greene 2000).
Here the logit model is used, since it lends itself to easier
interpretation.
Table 1 shows the
description of the independent variables considered for logistic regression
analysis and their hypothesized signs. The dependent variable takes on a value
of 1 if a pastoralist participated in collective action to start farming and 0
otherwise. The
explanatory variables had been tested for their importance by using descriptive
statistics before they were subjected to regression analysis. The results show
that participants are significantly different from non-participants with
respect to all but one variable[10].
Table 1: Description of variables and working hypothesis
|
Variable code (Xj) |
Description |
Mean of Xj or Percent of Xj = 1 |
Hypothesis |
|
AGEHH |
Age of household head in years |
40.1 |
- |
|
EDUCATE |
A dummy variable which takes on 1 if the household head is literate; and 0 otherwise |
25.7 |
+/- |
|
ACTIVLB |
The number of household members within the age range between 10 and 60 years[11] |
+ |
|
|
SUITAGR |
A dummy variable which takes on 1 if the area is either suitable for rain-fed agriculture or can be irrigated given existing water resources and capacity to irrigate; and 0 otherwise. |
66.5 |
+ |
|
PERCPLS |
Per capita livestock holding of household (TLU) |
3.1 |
- |
|
EMPOPP |
A dummy variable which takes on 1 if the household generates income from wage employment; and 0 otherwise. |
10.6 |
- |
|
SUPPORT |
A dummy variable which takes on 1 if external agents provided direct support[12] before and during collective activities; and 0 otherwise. |
49.7 |
+ |
Source: Own survey data
5.3. Regression results and discussion
The outputs of the regression are shown in
Table 2. The signs of the coefficients in the regression are all in agreement
with prior expectations. The chi-square statistic is significant, implying that
the explanatory variables (taken together) are important in explaining the
variability in the dependent variable (cooperation to start farming). The model
was able to correctly predict 86 percent of the cases vis-à-vis participation
in collective activities. Since the standard coefficients in the logistic
regression equation are not directly interpretable, the marginal effects of
explanatory variables were computed by using an additional algorithm in the
LIMDEP statistical software version 7.
Table 2:
Determinants of Cooperation among Pastoralists to Start Farming
|
|
Coefficients |
SE |
Marginal Effects |
|
|
Constant |
-3.6695** |
1.2439 |
-0.6348 |
|
|
AGE |
-0.0143 |
0.01523 |
-0.0024 |
|
|
EDUCATE |
0.5477 |
0.5483
|
0.0947 |
|
|
ACTIVLAB |
0.0561 |
0.0776 |
0.0097 |
|
|
SUITAGR |
3.8085** |
1.1561 |
0.6588 |
|
|
PERCPLS |
-0.1681** |
0.0623 |
-0.0291 |
|
|
EMPOPP |
-2.0585* |
0.8831 |
-0.3561 |
|
|
SUPPORT |
1.5636** |
|||