Neocolonial agenda: Agrarian transformations in Ethiopia and Sri Lanka

ABSTRACT The aim of the article is to show how neocolonial development has led to increasing ‘de-agrarianisation’ despite agricultural expansion, and this is at the expense of peasant production and pastoralism. Based on the authors’ own ethnographic research and literature studies, the article presents two cases of agricultural expansion: the Mahaweli Development Programme in Sri Lanka, aimed at redeveloping small-scale agriculture, and the reorganisation of pastoralist areas into large-scale cash cropping areas in Ethiopia. The authors find that recent agricultural ‘developers’ have failed to acknowledge the role and value of the traditional agrarian economy, as well as the pressure put on small-scale farmers and pastoralists through increasing capitalism in agriculture, land grabbing, and expropriation for agri-business. In conclusion, despite the historical and cultural differences between Sri Lanka and Ethiopia, both countries exemplify how the expropriation of state land under colonialism and after has facilitated the development of capitalist agriculture, involving irrigation, new settlements and migration, cash cropping, land alienation, and external public and private control. In Sri Lanka, increasing capitalism and technical reforms have led to social inequity and de-agrarianisation among small-scale farmers. In Ethiopia, agricultural development has been a political and economic process of alienation and exploitation for pastoralists.


Introduction
As the world became industrialised, questions emerged about how agriculture would manage and be affected by the process. The status of peasant farmers after serfdom, freehold, sharecropping, and landless workers has been highly debated (Kautsky 1900). Indeed, as early as the first decades of the 20th century, all aspects of land tenure, landownership, forms of organisation, and the status of farmers as landowners or 'proletarianised' land workers were raised. For instance, when discussing forms of organisation and land tenure for the peasantry, Chayanov (1991Chayanov ( [1919) rejected the state farms introduced by the socialist Russian state in favour of cooperatives and non-authoritarian forms of organisation.
Also, the fate of agriculture under colonial rule has been widely discussed, particularly in the context of the 'agrarian reforms' that affected rural populations. Some geographers, including Sautter (1978), have described the situation in colonised countries as a struggle between local traditional peasantry and modern, state-supported agriculture in plantations. World War II heralded a transition to modern agriculture in previous colonies in Africa and Asia, implying the use of modern agricultural techniques, state control, and a cash economy. Today, all agrarian societies in tropical Asia and Africa have been involved in a cash economy, with money contributing to regulating social relationships (Shanin 1971;Sautter 1978;Lund 1979;Ellis 1988).
The aim of this article is to uncover how contemporary agricultural systems and practices are embedded in the previous colonial powers of the elite, the state, and the government. Such power structures continue to set the agenda for the expansion of agricultural land, the introduction of new technologies, and external interventions by aid agencies, international corporations, and neoliberal policies, hence we use the term 'neocolonial development'. Neocolonial development has led to an increasing 'de-agrarianisation' (i.e. withdrawal from agriculture) at the expense of peasant production and pastoralism. To illustrate this, we present two cases of agricultural expansion: (1) the Mahaweli Development Programme (MDP) in Sri Lanka, which is aimed at redeveloping small-scale agriculture, and (2) the reorganisation of pastoralist areas into largescale cash cropping areas in Ethiopia.
A fundamental aspect of our discussion is the status of land. In Sri Lanka and Ethiopia, state-owned land refers to the collective property of rural communities in agricultural or agro-pastoral/pastoral areas, where traditional private property is not compatible with the mobility requirements for cattle and people. Alternatively, state-owned land can refer to the private property of individuals or corporate entities (private investors, local or foreign owners). Despite the historical and cultural differences between Sri Lanka and Ethiopia, both countries exemplify how the expropriation of state land under colonialism and after has facilitated the development of capitalist agriculture, involving irrigation, new settlements (as well as migration), cash cropping, land alienation, and external public and private control. In Sri Lanka, landless farmers from the populous southwestern part of the country and shifting cultivators from ancestral villages had to move into organised village hamlets with markets, irrigation canals, and agricultural fields to scarcely populated areas in the area covered by the MDP, where they became small-scale capitalist producers. In Ethiopia, under the 19th century Emperor Menelik, conquered land and surrounding kingdoms were subjected to 'indigenous colonial' rule, with the result that pastoralists and agro-pastoralists became alienated from their lands.
The following sections up to the Discussion section present the conceptual frame and describe the Sri Lankan and Ethiopian cases. Both cases show how political powers/governments contributed to escalating the tensions inherent in agricultural development by ignoring farmers' culture, knowledge, and resistance. The cases build on the authors' research experience in the two countries, as well as on relevant literature. 1 The purpose of the empirical presentations is to illustrate how different processes of agrarian transformation are surprisingly similar with respect to neocolonial development. Accordingly, we do not compare two countries but we analyse the situation in different areas of the countries. The MDP in Sri Lanka covers nearly half of the country's total land area, whereas the Ethiopian example shows how pastoralism is affected in a few Ethiopian regions. Furthermore, whereas the Sri Lanka example shows how modernisation and land reform lead to polarisation as well as de-agrarianisation, the Ethiopian example documents how de-agrarianisation is result of a political and economic process of exploitation. In the Discussion section of this article, we discuss the dispossessions that occur in modern agriculture due to state, technical, and administrative actions, as well as policy and power structures under economic reform. We conclude that despite different historical legacies it is possible to identify similar processes of change in agricultural development in the two countries.

Conceptual approach
Three stages of agrarian development may be identified. All three stages are intertwined, and together they have led to the complicated situation faced by agrarian communities today: the post-colonial period (1960s to the present day), the transition to capitalist and technical production (1970s to the present day), and the neoliberal period (1980s to the present day) in which de-agrarianisation is increasingly taking place.
In the post-colonial period, debate arose over the difference between 'peasant economies' and postcolonial agricultural production (Hydén 1980;Ellis 1988). Hydén (1980) described a network of interactions, communication, and support among Tanzanian and Kenyan peasant groups in East Africa. These groups, identified in terms of their relationship to structure, were united by kinship, community, religion, and other affinities. Such an 'economy of affection' is at the heart of African governance, which involves a mix of formal and informal institutions. This may contribute to explaining the prevalence of clientelism in African politics and peasant resistance to formal institutions in Africa. Likewise, similar agrarian changes have taken place in Sri Lanka, starting with the refurbishment of ancient irrigation canals, the establishment of permanent settlements, and the cultivation of paddy and vegetable fields in the 'Dry Zone' in the country's Central Province and Eastern Province in the 1930s.
From a planning perspective, Sautter (1978) described the distinction between peasant production and capitalist production, mainly with respect to their goals of social reproduction. Peasant production aims at the mere survival of families, communities, and even villages. Ellis (1988) observed that, despite relatively recent economic developments, farmers co-opted their traditional logic of production by maximising one or more of their household objectives. By contrast, the lens of capitalist production sees farmers as mainly producers or consumers, with the aim of producing, acquiring, accumulating, consuming, and, if necessary, using the maximum amount of goods produced.
To avoid risk, peasants have had different options for exiting their work in agriculture (Ellis 1988). Today, such exit options are increasingly practised, as population growth fragments small farms, farmer families become pauperised, and younger generations search for alternative methods of livelihood generation. The outcome of these trends has been termed 'de-agrarianisation' (Bryceson 1996;Jakobsen 2017), which describes the local development of non-agricultural income solutions for rural families, as well as emigration to urban areas and the take-up of seasonal employment on capitalist plantations.
In parallel with the shift from peasant production to small-scale family farming, plantations have been established in both Sri Lanka and Ethiopia. Sri Lanka has demonstrated a dual agricultural strategy, with plantations employing bonded Indian labour in the mountainous region in the central part of the country (normally referred to as the hill country) and cash crop production on small family farms in the scarcely populated state land in the Eastern Province and Central Province. From its inception, Sri Lanka's plantation industry has undergone drastic change, from the colonial period  through to nationalisation, to the present period of private management by agency houses (Poholiyadde 2018). In Ethiopia, European investors were initially attracted to the promise of establishing plantations near the highlands along the Kassam River, close to the Awash River. In the 1960s, the first large-scale irrigation development was undertaken along the Awash River. As in Sri Lanka, national authorities were established to oversee the development, and with the support of international agencies, both infrastructure and transportation routes were developed. Gradually, local pastoralists lost their rights to communal land and were forced to become sedentary.
However, in both countries, the transition to technical and capitalist agricultural production proved complicated. Recent discourse on 'rendering technical' provides some explanation for this. In common with Parange, Li (2007) and Ferguson (1990) emphasise that 'development agencies and governments attempt to frame potential problems as technical to make them appear both intelligible and fixable' (Paranage 2019, 2). Accordingly, rendering something technical means framing problems and their solutions in a way that pursues technical fixes without addressing root causes, through a 'will to improve', and capturing a host of social development aspirations (Li 2007). Paranage's article on the MDP in Sri Lanka explores how the 'official' narrative ignored why the programme faced problems by underestimating what was perceived as problematic, failing to pursue technical solutions to these problems, and overlooking 'expert' advice on how these issues could be fixed (Paranage 2019). Similarly, in Ethiopia, large-scale capitalist production on state farms and plantations was structured according to the whims of administrators, politicians. and, recently, agri-businesses, while the needs, knowledge, and traditions of pastoralists were generally and systematically ignored.
Finally, recent agricultural development carries the shadow of post-colonial and technical features, traditions, and structures. At the beginning of the 1980s, the liberalisation of commerce and the free transfer of financial resources were introduced in the Global South, leading to deepfelt social, economic, political, and cultural change in rural communities. Policies that allowed investors and corporations to invest freely were implemented, and agriculture became increasingly market-oriented and dominated by state intervention and transnational corporations (Amin 2006;Conneli & Dados 2014). The integration of production into global trade led to increasing environmental and social insecurity, neglect in food production, land grabbing, and emigration (Bryceson 1996;Jakobsen 2017). Such developments occurred worldwide, irrespective of the type of agriculture and specific cultural features.
Resettlement, irrigation, and cash cropping as a rural development strategy: the Mahaweli Development Programme The policy and implementation history of the MDP began on an optimistic note, with a vision of the promise of rapid modernisation: For Sri Lanka, the successful completion of the accelerated Mahaweli Programme will mark the closing of a sad chapter of dependence, and the opening of a new onean era of development, hope and fulfilment. All this will ultimately mean the general uplift of the quality of life of the people in Sri Lanka. It will mean a challengea challenge that will raise the spirit of man and create a new confidence in democracy and freedom. (Ministry of Mahaweli Development 1979, 10) Prior to the first colonisation schemes in the northeast of the country in the 1930s, Sri Lanka's Dry Zone had long been isolated. In 1963, the government (with the assistance of the Food and Agriculture Organization of the United Nations (FAO) and the United Nations Development Programme (UNDP)) initiated a preliminary mapping of a vast development project (the MDP) in the Mahaweli River Basin, which revealed that the area had sufficient water, land, and space to cater for a vast irrigation and resettlement programme (Supplementary Figure 1). Since then, the MDP, which covers approximately half of Sri Lanka's footprint, has played an essential role in the country's development, contributing to alleviating both demographic and economic problems (Lund 1993b). The first phase of the construction, during the period 1978-1984, was termed the Accelerated Mahaweli Development Project (AMDP). The project was completed within 6 years, instead of 30 years as initially planned. Lund conducted field studies in hamlets developed under the AMDP repeatedly during the period in which it was operational, and in 2004.
The MDP aimed at achieving national sufficiency in rice, stimulating the production of non-traditional crops, creating income-generating activities in agriculture, and establishing sufficiently large power to enable intensive cultivation through transbasin irrigation. The settlement package was characterised by central planning, implementation, management, and monitoring by the Mahaweli Development Board (Karunatilake 1988). Lund (1983;1993a;1993b) conducted research in the System H area, which was the target of the AMDP. In particular, the AMDP promised the irrigation of 1659 km 2 of land at a yearly capacity of c.7 million m 3 , a total capacity of 500 MW, the resettlement of 75,000 families, and the provision of roads, schools, hospitals, markets, and medical centres (for an example of the development of a town see Fig. 1). Lund (2003;2013) has shown that the landholding patterns of the MDP deviated significantly from those of the pre-settlement period with regard to social structure, settlement pattern, ownership, and workloads (traditional settlement patterns and social structure in the Dry Zone have been documented extensively by Pieris 1956, Knox 1966, Leach 1973, and de Silva 1981. While the previous compounds were small family units (20-25 houses) and homogenous in nature (by class and caste), the new settlements were much larger (100-125 or more houses), heterogeneous, and formed according to a hierarchical pattern of hamlets, villages, and townships (Figs. 2 and 3). Land was given to three types of settlers 2 : (1) resettlers, or local people who had been resettled under the new scheme, some of whom had lost land due to the construction and had received compensation in the form of cash and land; (2) new settlers, or landless people from various parts of the country (mainly from southern, heavily populated areas); and (3) evacuees, from land that had been affected by dam construction. Each family was allocated 2.5 acres (1.02 ha) of irrigated land, a well, and 2.5 acres (1.02 ha) of home garden land. They were also provided with tools and raw materials to set up a one-room 'temporary' mud hut on a self-help basis. Although settlers were defined as landholders, they were only given the right to use the land, and not to own it.
The geographical and ethnic backgrounds of the MDP settlers varied widely (Tilakasiri 1979). However, most settlers came from surrounding villages or areas that had been evacuated for the purposes of irrigation and cultivation. The settlers also varied in their status.
While most were poor, a few were relatively affluent and influential, and owned vast fields and herds of cattle and buffalo. In terms of ethnic composition, most were Singhalese Buddhists, while only a few were Tamils and Muslims (David 1986;Scudder 2005). According to Scudder (2005), of the 75,000 households settled under the AMDP, only 1.9% were Hindu and 2.9% were Muslim.
Following settlement, the department or organisation that initially assisted the farmers ceased to be actively involved, and individual settlers had to fend for themselves or work their way through a hierarchy of officers to receive assistance. At the top level, the Resident Project Manager held ultimate responsibility for the project, while a wide range of line departments (relating to, for example, agriculture, water and irrigation, land administration, community development, marketing, and credit development) managed particular areas. There was very little cross-sectoral collaboration to facilitate the situation for the settlers (Puttaswamaiah 1990).
Various studies have shown how the change from small, homogeneous family compounds to large, densely populated, and heterogeneous settlements in the Mahaweli brought about massive socio-economic and social change. Family structure changed from an extended to a nucleated pattern, land was given to the head of the household (usually the husband, with gender relations transformed to the detriment of women) and success depended on previous knowledge, including the ability to adapt and achieve economic success (Lund 1983;de Soyza 1995;Mathur & Cernea 1995). Over the years, the Mahaweli settlements turned into differential communities comprised of some successful farmers (i.e. those who were educated, of high caste and class, indigenous to the Dry Zone, and with good political connections) and many less successful farmers (i.e. settlers who were uneducated, of low caste and class, and foreign to the Dry Zone).

Challenges in the Mahaweli
According to Agarwal (1988, 3), a common feature of rural development strategies across Asia has been 'a concerted effort to increase productivity and agricultural surpluses through the introduction of new technologies and practices, embodied especially in the green revolution package of practices, and in significant irrigation and resettlement schemes'. Since the 1960s, such schemes have always been initiated and implemented by the state, often as prestigious projects carried out by reform-friendly governments and politicians. In 1977, Sri Lanka experienced a shift in government from the Bandaranaike regime, which preached redistribution and socialist, decentralised, and smallscale industrial and agricultural development, to the Jayawardena regime that preached economic reform, cash cropping, and strong, centralised systems of control. President Jayawardena's government was intrigued by the development models of newly industrialised countries, such as Singapore and Malaysia. In Malaysia, economic reform was linked to the modernisation of the country's agricultural and industrial sectors. However, shifting political priorities dominated and often ignored realities on the ground.
For several decades, the MDP was marketed as bringing Singhalese civilisation back to the Dry Zone. Furthermore, various presidents tried to create a populist ideology to legitimise Singhalese nationalist policies. To incentivise the movement of people to Mahaweli settlements, the authorities used symbols, rituals, legends, and stories about the glorious past (e.g. the 'Dawn of the Mahaweli Era') to connect the settlements with ancient 'homelands'. Once the dams, canals, and new settlements had been constructed, temples were built and traditional rituals (e.g. inauguration ceremonies involving oil lamps and prayer, certain thanksgiving rituals) were reintroduced (Lund 1993a).

Structural changes
When the MDP was formed, a Mahaweli Ministry was set up, and large international donors (e.g. World Bank, Asian Development Bank (ADB), UNDP, FAO, Overseas Development Institute (ODI), German Agency for Technical Cooperation (GTZ), United States Agency for International Development (USAID), Swedish International Development Cooperation Agency (SIDA), Canadian International Development Agency (CIDA)) provided vast amounts of support in terms of money, equipment, and expertise. The preoccupation with growth led to a neglect of both distributional and ecological considerations and uneven agricultural capitalist development (Agarwal 1988). This in turn exacerbated class and regional inequalities, and it impinged on the economic institutions and social fabric of the societies (Lund 1993a).
Thus, resettlement in the Dry Zone of Sri Lanka was of considerable geographical, economic, political, and cultural significance. It contributed substantially to changes in Sri Lanka's population map and increased the production of rice and other cash crops. However, politically and symbolically the project implied the resettlement of Singhalese people in (or near to) previous Tamil areas. On this basis, some have argued that the MDP was a direct cause of the intensified ethnic clashes between the Singhalese and Tamils in 1983 that were seen as the start of the civil war 1983-2008 (Balasingham 2004).
Housing is one of the significant challenges of construction and its aftermath. The official planning documents and research material relating to the MDP reveal that some subsidies have been available for housing, but they reveal very little about the creation of homes, and a sense of identity and belonging (Lund 2013). According to Leach (1973, 27), the household group is more important than the family: 'The compound group as a piece of the ground, is a continuing entity transmitted unchanged from generation to generation.' Thus, what constitutes a home is intrinsically connected to the physical structure, such as the house or courtyard. Home also relates to a family pattern, access to land/resources, and the organisation of local society. In the colonisation scheme, the available house was, at first, only a temporary hut; however, the hut was later expanded into a permanent home with a surrounding garden, latrine, and fence. The MDP encouraged settlers to construct their own core house up to 'basic standards', and then to develop a homestead plot. However, throughout her repeated visits to the area, Lund noted that several settlers did not manage to build a new home. They lived in modifications of the initial core house and remained poorin some cases, becoming even more impoverished (Lund 2013). Some had to mortgage their land, and others (especially those who were young) had had to move. The initial farmhouse stood as it always had, with an ageing population unable to sell its land due to usufruct rights and an inability to secure a collateral-based loan from a bank. The indigenous population suffered the most (cf. the situation in the village of Henanigala South (Eastern Province), documented by Lund in 2003 and2013). These situations occurred frequently, and only a few farmers managed to consolidate land and build wealth, as shown by Figs. 4 and 5. According to the authorities, the housing effort was constrained by the relatively high cost of house construction. Despite the 'substantial support' (Rs 5000 (46 USD)) offered to each family, 'the settlers invariably find it difficult to supplement this assistance with their own resources' (Ministry of Mahaweli & Ministry of Lands and Land Development 1986, 52). Furthermore, 'the normal State intervention mechanisms which are designated to assist the people of this category, cannot be made fully operational, due to the same reasons of low resources endowment' (e.g. no access to credit, no collateral, legal grant documents not issued on time) (Ministry of Mahaweli & Ministry of Lands and Land Development 1986, 53). New settlers were also constrained in their reconstruction efforts, due to a lack of cheap labour (e.g. transport/lorry owners, carpenters, masons) and building materials: 'Since the period of induction is one of trauma and discomfort, the use of family labour is also restricted for this kind of activity' (Ministry of Mahaweli & Ministry of Lands and Land Development 1986, 53).

Towards a differentiated and unequal society
The Mahaweli is an area in which different population groups (e.g. farmers, politicians, officials) pursue often-divergent interests. The degree of freedom for different groups is determined by several systemic constraints, related to the size and multifunctionality of the system and the speed of construction (Muller & Hettige 1995). The Mahaweli Authority's response to simultaneous action was strongly centralised management functions. In hindsight, one may question the wisdom of this response, as centralisation led to societal and agricultural problems. Hence, the fundamental problem of the MDP related to its programmed structure and political economy paradigm. In particular, the planners and experts ignored farmers' knowledge. Such ignorance in the Dry Zone has been documented by social science scholars since as early as the 1930s, when the first irrigation schemes were initiated (Lund 1983;1993a;1993b;Paranage 2019). Furthermore, the highly donor-dependent economy struggled to diversify the local economy and sustain livelihoods as the programme led to highly differentiated returns for the farmers. For example, the MDP led to water distribution problems and paddy cultivation in the Dry Zone (Tilakasiri 1979; Wichanachchi et al. 2014). In addition, the allocated land was not of equal quality. Several farmers were allocated land unsuitable for cultivation, drinking water was contaminated, and incentives such as pesticides and fertilizers were maldistributed. Examples of illegal appropriation of land and corruption also took place (Lund 1983;2003).
The tenurial relations that emerged from illegal land transactions are a key to understanding the new differentiation processes in the Mahaweli settlements (Lund 1983;1993a;1993b;Shanmugaratnam 1984;Muller & Hettige 1995;Sørensen 1996;Scudder 2005). Many farmers became marginalised and 'proletarianised' because they were unable to cultivate their own land (Siriwardena 1981;Krimmel 1986;Lund 2003). Today, the differentiated society is still apparent in terms of household composition and access to agricultural resources.
The MDP also highlights the issue of forced versus voluntary migration (Sørensen 1996). In the Mahaweli, forced migrants previously lived inside the area but were affected by the planned interventions or flooding. By contrast, voluntary migrants were primarily from outside the project area. Forced migration in the Mahaweli could be linked to limitations to land user rights too, as many settlers who could not sell their house or land moved, to acquire bankable assets. Hence, population density and structure in the Mahaweli changed over the years, leading to new developments and challenges. In addition, the impact of population growth and density was not considered at the project planning stage. Today, most families who live in the Mahaweli are either secondgeneration or third-generation settlers. They face problems due to the deteriorating agricultural economy, an inability to move, old age, and dependency on external funding, such as money from children working in the army, in cities, or abroad (Azmi 2008;Lund 2013).
Over the years, numerous policies and population related factors have contributed to de-agrarianise the Mahaweli project area. There has been an observable shift from the situation of pioneer settlers working the land with very simple means and creating a home during the early years of settlement, to the present situation, in which the children and grandchildren of those pioneer settlers are leaving the Mahaweli project area to take up seasonal work. As land in the area is scarce due to loss or fragmentation of land, second-generation and third-generation settlers aim at making a life outside farming. in industrial areas or abroad. However, most individuals work in the Sri Lankan diaspora without the accompaniment of their dependents. Grandparents and young children are frequently left without the work input and social contribution of the most active generation, with the result that they often give up farming. Such labour migration may lead to the splitting apart of households and families (Azmi 2008).
Finally, tensions in the settlement process, which began during colonial times, occurred throughout the planning and implementation of the MDP, and have continue to the present. Such tensions relate to the competition for resources, bureaucratic control, and top-down management. In addition to the MDP's technocratic bias in implementation and lack of coordination, the programme was also ridden with attitudinal problems and ethnic and class biases. One official document states: A certain degree of complacence on the part of the majority of the settler families, in being tolerant towards living in squalid conditions, despite the opportunities given to them, and a somewhat fatalistic attitude […][has] frustrated most of our enthusiastic attempts. (Ministry of Mahaweli & Ministry of Lands and Land Development 1986, 150) To summarise, the MDP represented a significant effort to reconstruct the rural economy of Sri Lanka as a whole, with every settler given an equal share of land and a home (a one-room mud hut). Accordingly, the programme may be seen and analysed as the country's first structural adjustment programme. However, over the years the Mahaweli has developed into a highly differentiated society, as a few settlers have succeeded in creating a good income and home, while the majority have failed and been forced to move, leading to rapid de-agrarianisation. This trend is mainly affecting second-generation and third-generation settlers who have not been able to find additional land in the planned settlements. Finally, labour migration has split households and families (Azmi 2008). These effects have significantly impacted the ways in which settlers perceive and identify themselves, invest in their farms and pioneer houses in the settlements, and manage their vulnerable and impoverished life situations (Lund 2003;2013).

The birth and extension of 'modern' plantation agriculture in Ethiopia
In Ethiopia, the use of 'modern' agricultural techniques is geographically constrained to the Rift Valley (Supplementary Figure 2). The Rift Valley is a lowland area with extensive plains and a lower population density relative to the highlands, where traditional peasant agriculture remains dominant. Individuals who live in the Rift Valley are mainly pastoralists and agro-pastoralists belonging to ethnic groups that were incorporated (by force) into the Ethiopian empire/state at the end of the 19th century.

Introduction of modern agriculture in Ethiopia
At the beginning of the 20th century, modern agriculture was introduced in the Harari Region 3 (Gascon 1995), with coffee plantations initiated by Ras Tafari (who later became Emperor Hailé Sélassié) in the Awash Valley, which was mainly populated by the Afar (or Danakil) people. One plantation in particular seems to represent one of the first 'European' plantations in Ethiopia: an irrigation-based plantation created in 1904 by Armand Savouré, a well-known arms trader, entrepreneur, and businessman in Addis Ababa. The plantation was located along the Kassam River, a tributary of the Awash River, in a place called Awara Malka. Savouré acquired the land as a concession (4000 ha for 50 years), 'given' to him by Emperor Menelik. At that time, the Ethiopian state had only very recently acquired theoretical control over the Afar Region (northeastern Ethiopia), which was inhabited by pastoralists. Punishing expeditions were used to exert state control, and examples of the tense relationship between highlanders/Amhara and pastoralists have been recorded since the 1920s. For example, a 1928 entry in Menelik's journal illustrates how the government representative in Awara Malka, a small settlement on the Kassam River, a tributary of the Awash River, considered the Afar: The chief representative of Abyssinian authority […] came to call on us. He was an Amhara named Ibrahim […] Awara Malka lay in the borderlands, and it was one of the few places which were firmly held by the government. Ibrahim […] hated the Danakil and killed them on the slightest provocation. (Nordic Africa Institute n.d.) In a short brochure addressing potential European investors, Savouré (1912) provided instructions for creating a plantation, drawing on his own experience in Awara Malka. The text resembles a manual for investors, covering all of the necessary components for plantation success: ideal land lease conditions, requirements for soil quality, the availability of water and irrigation, and proximity to markets. It also indicates the crops that were most likely to flourish (cotton, sugar cane, fruits) and emphasises the importance of easy access to cheap labour. Additionally, Savouré's brochure notes that Awara Malka is located close to the highlands and the planned railway line; the railway line was very convenient for European personnel.
Despite the noted strengths, the Awara Malka plantation ultimately failed. After some years, the land was only partially cultivatedin part due to increasing tension and violence. On 14 February 1912, Michel Bourgeat, a visitor, described the plantation as poorly managed and vulnerable to attack by spoliating Afars (Baudouin 2018). The plantation was located on lowlands that had traditionally been used by the Afar for cattle grazing during the dry season. In retaliation for the plantation's encroachment onto their territory, the Afar burnt down some of the buildings. Also, malaria had caused difficulty for the Awara Malka plantation, in common with other plantations in the Rift Valley.
It is noteworthy that Savouré's brochure did not contain a single word about the Afar; it totally ignored their existence in the Rift Valley. Reality did not fit with the glorious perspectives sketched in the text. Savouré almost went bankrupt due to his inexperience as a plantation owner. Nonetheless, the plantation still exists today. It was taken over by Italian investors during the Italian occupation (1936)(1937)(1938)(1939)(1940)(1941), and thereafter by the restored imperial power. In 1974, the plantation was nationalised under the Derg regime. It is now under private ownership.
Savouré's plantation was not the first to have been created by foreigners in Ethiopia. However, Savouré was the first to promote and experiment with an agricultural model based on irrigationa model that eventually became generalised in Ethiopia. Savouré's brochure was therefore a landmark publication, providing a model for the introduction of plantation and irrigated agriculture in Ethiopia. Such agricultural production was aimed at the wider market. Thus, it required transport facilities, irrigation systems, a seasonal and partly permanent workforce with minimal pay, and a land lease contract from the authorities in exchange for low rent. Today, modern agricultural projects apply a similar sociotechnical and economic model based on official support and profit maximisation, which is blind to social responsibility and indigenous rights.
In Ethiopia, irrigation is mainly practised through small-scale installations; however, there are also some large installations and dams (see Supplementary Figure  3) (Müeller-Mahn et al. 2010). The present discussion deals only with large-scale dams and irrigation systems in the Awash Valley (Koka Dam, Kessem River, Tendaho Dam) and the Lower Omo Valley in southwestern Ethiopia (Gibe III Dam).

Land tenure under Haile Selassié's imperial power
Prior to the creation of new plantations, it was necessary to reassess the land tenure rules, and thereby reaffirm the all-encompassing state lands. Hailu (1975, 26) wrote the following about Article 130 (d) of the revised constitution of 1955: All property held and possessed in the name of any person, natural or judicial, including all land in escheat and all abandoned properties, whether real or personal, as well as all products of the sub-soil, all forests and all grazing land, water courses, lakes and territorial waters, are state domain.
It is worth noting that the right of pastoralists to receive compensation for lost grazing land was acknowledged.
In the 1960s, the 'plantation model' began to expand. The first large-scale irrigation development was undertaken in 1961 at Tendaho, in the lower Awash Valley (Said 1994, 2), following the creation of a large dam (the Koka Dam). The dam, and its associated reservoir, enabled the development of irrigation infrastructure and the establishment of new plantations along the Awash River. A national agency, the Awash Valley Authority, was created in 1962 to organise and monitor the process, and with the support of international agencies, to 'develop' the area.
In the literature, several scholars have critiqued the general neglect of pastoralists and attempts to 'sedentarise' them during the development of irrigated agriculture since the early 1970s. In December 1972, a conference on the pastoralism in tropical Africa (Monod 1975) was held in Niamey, during which participants advocated the adaptation of the pastoral way of life to the constraints of semi-arid and arid zones. A few years later, Flood (2002Flood ( [1975) voiced the need for the protection of Afar pastoralists (and other pastoralists, such as the Karrayu) against societal damage induced by a policy aimed at their 'sedentarisation', and the negative consequences of irrigation that was not well controlled for the environment, including soil salinisation (i.e. in the irrigated areas) and changes to vegetation.
Since the 1970s, several studies that have analysed pastoral policies have criticised the government's policies towards pastoralists and addressed the consequences (both social and environmental) of the installation of plantations on land formerly used by pastoralists (e.g. Hailu 1975;Said 1994;Flood 2002Flood [1975). However, these studies have not had any discernible effect in Ethiopia and other African countries such as Tanzania, Kenya.

Land tenure under the Derg
During the Ethiopian Revolution of 1974, land was nationalised, and state farms were established on communal land and former privately owned plantations. As also described by Helland (2015), the 1975 Land Reform nationalised the lands of the Sultan and all other commercial concession holders in the Afar Valley. However, the land was not returned to the Afar pastoralists. Said points out the contradiction between state declarations and realities in the field as follows: Although the proclamation of 1975 stated that the pastoral people have possessory right over the land they customarily used for grazing, the state's control of pastoral land increased through the establishment of additional farms in the valley. (Said 1994, 11) Farms were placed under various forms of government management, as either government-owned commercial enterprises or state farms. In March 1990, at the end of the collapsing Derg regime, a period of a mixed economic policy started and marked the end of the collectivisation of agriculture: land belonging to all collective farms was distributed to members (i.e. of the collective farms) (Zerihun et al. 2002).
Land tenure after 1991: the EPRDF After 1991, former collective state farms were privatised. This did not change the problems faced by pastoralists. Between 1994 and 1997, the new central government of Ethiopia seemed willing to give pastoralists some form of land tenure security. However, the government was reluctant to change the main structures and policies of the 1975 Land Reform. The 1994 Constitution asserted Ethiopian pastoralists' right to free land for grazing and cultivation, and the right to not be displaced from their land. These rights were to be implemented in law (Helland 2015, 7), with regional states drafting their own regulations. However, in 2005, a complete turn in policy opened up for privatisation. The federal government issued a revised proclamation (The Rural Land Administration and Land Use Proclamation No. 456/2005), allowing the 'government being the owner of rural land to change communal rural land holdings to private holdings' (paragraph 5.3) (Federal Negarit Gazeta of the Federal Democratic Republic of Ethiopia 2005). This move invited in private investors, thereby significantly weakening pastoral communal land tenure rights. Abdulahi (2007) noted that the new proclamation mirrored many provisions set out in previously established Amhara and Tigray regional policies, suggesting that 'the land tenure laws of these two states [were] the foundations for the 2005 Proclamation' (Abdulahi 2007, 123). The proclamation therefore appeared to represent the deepening ideological commitments of the Tigrayan political elite towards a universal prescription or tenure model for the whole of Ethiopia (Crewett et al. 2008, 19). Thus, 2005 represented a point of acceleration and generalisation of the land grabbing process, duly documented by Rahmato (2011), which had begun a few years earlier when cultivated farmland was handed over to international investors for the commercial production of flowers. This initiated limitless land expropriation by both national and international investors: 'In the past four years (2007)(2008)(2009)(2010) it [the government] has signed over some 2.9 million acres (1.17 million ha) of land, mostly to foreign investors […] expect to lease a total of 7.4 million acres (31,000 km 2 )' (Abbink 2011, 517). Additionally, in certain contexts, the development of hydroelectricity became a priority and, against which no pastoralist or agropastoralist could compete to protect their rights.
Land leases for flower plantations spread rapidly in the late 1990s, as the conditions for them and future concessions were very attractive (Abbink 2011, 516). An Oakland Institute report on land deals lists a selection of federal incentives (Mousseau & Sosnoff 2011, 15), including the exemption of import and export duties, other taxes, and custom duties, as well as easy access to credit and no income taxes under certain conditions. Additionally, there were other incentives, such as extremely low rent and sometimes even free rent, no requirement for environmental impact assessments, and low fees for water (Makki 2012).

Foreign and domestic investors
Geographically, there appears to have been substantial investment from the Gulf States, as well as from India and, more recently, from China. While much media attention has focused on large-scale foreign acquisitions, the contribution of small-scale domestic investors to land deals is of critical importance. The majority of investors are domestic, and they represent a privileged group in Ethiopian society. They are reflective of the significant corruption under the EPRDF, 4 when members of the Tigray ethnic group and supporters of the regime monopolised land and resources. As Mousseau & Sosnoff (2011, 23) explain, 'In those regions most of the businesses are owned by the Tigrayans (and other Highlanders to a lesser degree), and almost all of the domestic agricultural investment lands are held by the Tigrayans.' Researchers have expressed different opinions about the future way of life for the Afar. According to Helland (2015), the source of ecological problems/degradation is the weakening and destruction of communal land tenure regimes. Population increase and a worsening resource base have led individuals to diversify their activities in non-farm and non-pastoral activities, while relying on relief and facing destitution. However, in some places such as the lower Awash Valley, the local Afar, under the Sultanate of Awsa, have managed to develop their own solutions, showing some capacity to adapt to change or maintain control of their land. Specifically, some plantations in the lower Awash Valley are owned by Afar people, and some Afar households have even transitioned to agro-pastoralism. Helland (2015, 2) observed a capacity for transformation and survival of pastoralism among the Afar: In sum, there is increasing evidence of a growing imbalance between the human population and the resource base in Afar. This imbalance is exacerbated by the fragmentation and loss of key rangelands to irrigation, invasive bushy species and to conflicts along the borders, with the result that levels of vulnerability are on the increase. Not all households can be supported by pastoralism, even under normal circumstances […] Despite these constraints, it can be expected that pastoralism will continue to play a significant role in Afar as livestock production can generate income and wealth from Afar's poor resource base.
By contrast, Markakis (2021, 95) expresses a far more pessimistic account, heralding the end of pastoralism: What is obvious already is the rapid decline and ultimate disappearance of mobile pastoralism, for the very simple reason that mobility is no longer possible. 'From our interviews with regional officials, the overriding consensus was that pastoralism has reached a dead end when it comes to sustaining the economic livelihood of the communities,' is a recent appraisal.
Environmental issues of water scarcity, soil salinisation, rangeland degradation, and the proliferation of invasive species such as Prosopis Juliflora (a thorny evergreen srub/small tree)) are widely documented (Said 1994, 4-7;Abebe et al. 2015;Mekkonen et al. 2019;Kebede 2021;Shiferaw et al. 2021). Currently, large swathes of rangeland in the Awash Valley are being degraded, leading an increasing number of pastoralists to turn to agropastoralism or to become destitute and forced to move to urban areas. Poverty is spreading, and cotton plantations are being replaced by sugarcane and fruit production industries, with a workforce that is mainly recruited from the highlands.

A new wave in the lower Omo Valley
A total of 50 years of failure and 50 years of analysis have not managed to modify the ideology and policies at work in Ethiopia. Ethiopian authorities have only mentioned pastoralists' rights and wishes formally in official documents, and the objective of the policies, when applied, is still to 'sedentarise' pastoralists. The most astonishing proof of this was offered by Prime Minister Meles Zenawi in a speech (Meles 2011, 1-2) that made no secret of the government's plans for pastoralists, which fell under Jaulin's definition of a planned ethnocide (Jaulin 1970). Specifically, Meles outlined a plan to build a series of larger dams (i.e. Gibe III), a massive irrigation system, and sugar plantations (Supplementary Figure 4) that would encroached upon large areas belonging to Omo National Park and Mago National Park, both od which are in the Lower Omo Valley. Meles's plans were aimed at tackling the shortage of water, which he claimed would 'help cattle raising to become productive and modern'. Ironically, Meles spoke as if his plans had been made in collaboration with pastoralists: 'Our government realized this problem and made plans with the pastoralist community' (Meles 2011, 1).
Meles used 'modernity' as a keyword with a clear message: 'We want our people to have a modern life and we won't allow our people to be a case study of ancient living for scientists and researchers' (Meles 2011, 2). His speech ended with a global wish: I promise you that, even though this area is known as backward in terms of civilization, it will become an example of rapid development. To attain this modern life, the solution is the Gibe 3 dam and a big irrigation system in this wide and fertile area of South Omo. (Meles 2011, 2) No mention was made about the fact that the majority of the land used by pastoralists would be given to domestic and foreign investors, or that residents in the affected lands would be subjected to forced eviction.
Meles' speech provoked harsh criticism from pastoralist representatives (e.g. Belachew 2011) and NGOs (e.g. the Oakland Institute, Human Rights Watch), as well as academic researchers (mainly anthropologists). Despite Meles' rosy vision, the reality on the ground was a neglect and eviction of the original population, pastoralists, and others. This was reinforced by ethnic, cultural, and historical prejudice, coupled with greed and corruption. Highlanders (Amharas, Tigreans) perceived the Afar as dangerous and backwards people, and they (the highlanders) had a widespread wish to eradicate pastoralism through the establishment of 'proper' agriculture and forced sedentarisation. Such efforts were deemed necessary to achieve modernity. Modernity was thereby imposed on the local population, who had no way of participating in their own evolution. When local people did not follow orders, the bureaucracy used the army and police to force compliance. Regassa Debelo et al. (2018, 143) synthesises the situation as follows: For all three successive regimes in Ethiopia, expropriating land from people in the peripheries has always been positioned within the discourse of transforming the 'unruly' and 'violent' people and the wilderness environment they inhabit into modernised, governable and transformed subjects and areas […] Because this dominant narrative does not recognise the people in the peripheries as the rightful owners and inhabitants of the land, an enclosure that restricts them from their land is justified as the state's sovereign right to exercise power over its inhabitants and territories. Abbink (2011, 528) used the expression 'irony of history' to describe the situation: In Ethiopia, the old revolutionary ideal of 'land to the tiller' has been abandoned, as the tiller is urged to work on the 'land of the foreigner'. In fact, the LSLA [large-scale land acquisition] process in Ethiopia is an example of the curious global alliance of international capitalist enterprise (including China, reinventing state capitalism) and undemocratic governance in the countries with the land resources.
Fortunately, due to corruption, incompetence, lack of capital, and other reasons, few investors and land grabbers have succeeded in their enterprises (The Economist 2020). In 2018 less than 3% of the 90,000 ha leased to investors in three of South Omo Zone's 5 districts was being farmed (The Economist 2020).

Discussion
In this article we have shown that despite the significant cultural, social, and political differences between Sri Lanka and Ethiopia, both countries have witnessed similar processes of change in agricultural development. There have been six main changes and they are respectively discussed in the following six sections.

The political appropriation of land and resources
Both cases of agricultural change presented in this article were fundamentally politicised. Farmers had to adapt to shifting political and economic regimes and privatised forms of production, which subjected traditional peasant production to significant stress. In Sri Lanka, private production and the expropriation of land occurred at the expense of less resourceful farmers in traditionally peripheral lands. In Ethiopia, the plantation economy shifted from a structure of state-owned plantations to one of large industrialised production units owned by international corporate entities.

Land alienation and land grabbing
Both cases discussed in this article illustrate neocolonial reform and the expropriation of state land, with the support of leading politicians and heads of state. In Sri Lanka, such land was previous forest land inhabited by traditional smallholders practising semi-subsistence paddy production combined with the shifting cultivation of other food crops. In Ethiopia, state land was taken for large-scale state farms and plantations, which eventually displaced pastoralists from their traditional grazing lands. In the long term, global threats to agricultural production stimulated a rush of investments, agro-businesses, and land speculation at an unprecedented scale.

A lack of environmental sustainability
Both Sri Lanka and Ethiopia suffered from land degradation, poor maintenance, numerous failures among experts and bureaucrats, a lack of professionalism among contractors, and conflict with local populations, putting the long-term sustainability of the physical environment at risk. In Sri Lanka, a lack of productivity due to monocropping, drought, and irrigation failures represented major problems. In Ethiopia, land alienation, environmental degradation, and local conflicts were particularly significant.

Technocratic and capitalist development
In both cases, agricultural development was dominated by 'rendering technical', involving foreign experts, private actors, and businesses. In Sri Lanka, resettlement and reconstruction efforts suffered from an unwillingness to provide land deeds, biased selection of beneficiaries, and poor coordination among external implementing and participating actors and agencies, both domestic and foreign. Furthermore, modern incentives (seeds, tools, machinery) in agriculture led to capitalised production. In Ethiopia, traditional pastoralists and agro-pastoralists lost access to grazing land due to plantation development.

De-agrarianisation and mobility of the rural workforce
Neocolonial agriculture increased the differences between actors, groups, and 'beneficiaries' involved in agriculture, with some moving socio-economically upwards and others becoming alienated from agriculture. In Sri Lanka, some farmers gained under the new settlement regime, while others did not. Secondgeneration and third-generation settlers increasingly moved away from the settlements to find work outside, including abroad. However, lack of ownership rights prevented them from selling their houses and acquiring collateral. Increasingly, aging populations are remaining behind in the MDP area. In Ethiopia, 'proletarianisation' dominated agricultural development. Dominant ethnic groups and foreign businesses and corporations took control of agricultural production, while traditional pastoralists were forced into non-pastoral livelihoods. Predominantly, only paid workers from the highlands were able to find work in agri-businesses.

Cultural erosion
Both cases illustrate a loss of cultural heritage and traditional rights. In Sri Lanka, capitalist production grew at the expense of traditional semi-subsistence cultivation, and farming became increasingly performed by farmers from areas outside the Dry Zone. The traditional village (cf. Fig. 3) disappeared and was replaced by large, planned settlements. Similarly, local drought-resistant crops and ways of production were replaced by modern means of production, controlled by the state and local bureaucrats. Likewise, in Ethiopia, pastoralists' traditional means of survival began to disappear, due to land grabbing, ethnocide, and 'proletarianisation'.

Conclusions
In this article the focus is on how the present-day agricultural development policies in Sri Lanka and Ethiopia have their historical roots in the colonial power structures and Eurocentric development thinking, and how these agricultural development processes have had negative effects for the local farmers and pastoralist groups in the studied places in the respective countries. Recent agricultural 'developers' have failed to acknowledge the role and value of the traditional agrarian economy, as well as the pressure put on small-scale farmers and pastoralists (from the colonial period to the present day) through increasing capitalism, technical reforms, and political interference. While the Sri Lanka example shows how increasing capitalism and technical reforms have led to social inequity and de-agrarianisation among small-scale farmers, the Ethiopian example shows how agricultural development has been a political and economic process of alienation and exploitation for pastoralists. Irrespective of forms of rule, ideology, or local relations, the same forces are at work in colonial, neocolonial, and development policies.