Large-scale land investment in Ethiopia:

How much land is being allocated, and features and

outcomes of investments to date

Report for the Bill and Melinda Gates Foundation and the International
Institute for Environment and Development (IIED)

James Keeley, Wondwossen Michago Seide, Abdurehman Eid and Admasu Lokeley

June 2013

1
Contents

Summary of main findings 4

1. Introduction 7
2. Methodology 7
3. Background to large-scale land investment in Ethiopia 9
4. The process for agricultural land investment 14
5. How much land has been allocated to investors? 20
6. Features of land allocations 24
6.1 Size of land allocations 24
6.2 Where is land being allocated? 24
6.3 Profile of investors 29
6.4 Crop focus for large-scale land investment 31
6.5 Implementation of land leases 32
7. Outcomes 36
7.1 Revenue from land investment 36
7.2 Employment 38
7.3 Food security 40
7.4 Resettlement and loss of land 41
7.5 Land use and environmental impacts 43
8. Conclusion 47
References 49

2
Exchange Rate

18.7 ETB = 1 USD

Acronyms

AISD Agricultural Investment Support Directorate

EIA Environmental impact assessment
EPRDF Ethiopian Peoples Revolutionary Democratic Front
GTP Growth and Transformation Plan
ILC International Land Coalition
MOA Ministry of Agriculture
SNNPR Southern Nations, Nationalities and Peoples Region
Woreda district

Figures

Box 1: Advertising poster produced by Agricultural Investment Support Directorate, Ministry of

Agriculture, Ethiopia 12
Box 2: Challenges in establishing accurate figures for land allocations 20
Box 3: Environmental impacts of Saudi Star rice farm 43
Box 4: Environmental protection measures taken by Saudi Star 44

Chart 1: Land available in the federal land bank (hectares) 15

Chart 2: Large-scale land allocation in Ethiopia by region 18
Chart 3: Allocation of land in Ethiopia, all sources 19
Chart 4: Regional and federal land deals by different size category 22
Chart 5: Allocation of land by region from the federal land bank 23
Chart 6: Land allocated by regional governments 25
Chart 7: Allocation of land for sugar production 26
Chart 8: Federal land bank investors by nationality or diaspora status 27
Chart 9: Regional government land allocations, Ethiopian and federal investors 28
Chart 10: Federal land bank allocations by crop type 29
Chart 11: Projected investment for selected land leases 35
Chart 12: Projected jobs per 1000 hectares 36

Table 1: Allocation of land by region from the federal land bank 24

Table 2: Origin of investors with agreements with the federal government 28

Map 1: Ethiopia, with regions marked 10

Map 2: Ethiopian Sugar Corporation map of major sugar production areas. 26

3
Summary of main findings

1. At least 1 million hectares of land have been leased by the Ethiopian government:
around 380,000 hectares from the federal land bank managed by the Ministry of
Agriculture, 335,000 hectares by regional governments, and 335,000 hectares for
state-run sugar plantations. While foreign investors (particularly Indian investors) are
important, most land investment is by domestic investors, and also by the Ethiopian
state for industrial crops.

2. The Ethiopian government has strategically promoted land investment as part of the
Government’s five-year Growth and Transformation Plan, which envisages that
Ethiopia will be food secure and a middle-income country by 2025. Land investment,
possibly in contrast to some other countries, is not primarily driven by overseas
interests.

3. Land is constitutionally a regional responsibility, but the federal government has taken
control of administration of large areas of land to facilitate investment, and actively
courted international and domestic investors. The federal land bank may be a unique
system for making land available to investors. It is facilitated by the fact that land is
owned by the state and the people in Ethiopia, rather than individuals or communities.

4. Land has been given out rapidly by regional governments with limited scrutiny of
investors, and until recently no environmental impact assessment. In many regions
record keeping about land investment has been poor. This is gradually changing as
regional land bureaux seek to verify land allocations, check on investments and where
necessary cancel leases.

5. Monitoring and evaluation remains a major challenge given the size of many regions,
the remoteness of investment locations and lack of staff and vehicles.

6. Only a small amount of land allocated has been developed. This is because of lack of
roads, bridges, power and other infrastructure in investment areas, high costs of land
development, poor technical and financial capacity of investors, the security situation in
some regions, and deliberate abuse of land investment licenses or land lease
agreements. Land rental prices are low (as little as US$ 2 per hectare per annum in
some regions), and in some cases land has been taken for speculative purposes or to
take advantage of tax and finance privileges.

7. Some of the land leases are very large: 8 are over 25,000 hectares and one is 100,000
hectares. Some government officials suggest that these sizes are unmanageable and
that 10,000 hectares would be a more realistic upper size limit.

8. Most land has been allocated in developing regional states (Benishangul-Gumuz and
Gambella), or lowland parts of SNNPR. These areas have not previously been
intensively cultivated, and are either part of shifting cultivation, or agro-pastoralist
systems (Benishangul-Gumuz and Gambella), or are part of pastoralist rangeland
systems (SNNPR).

9. Land allocations in some lowland areas has the potential to significantly undermine
pastoralist systems as access to important water resources is lost, eroding the viability

4
of rangelands where herders need to move across large areas to take advantage of
spatial and temporal variability in the availability of resources.

10. The degree to which land is vacant is overstated by some officials. There should be
clear acknowledgement when investment lands are part of shifting cultivation or
pastoralist systems and proper compensation. There should be provision of alternative
services where it is no longer possible to access land. In many cases where relations
with communities are difficult because of loss of land, they could be improved at the
margins at least, if access to water points or off-season grazing were given attention by
investors. When investors are assessed priority should be given to those with clear
plans for delivery of services such as clinics, schools and training programmes for local
communities. Adherence to these commitments should be part of monitoring
processes.

11. There is a need for more studies of the economic returns to large-scale land
investment, compared to other land use systems. It is possible that pastoralism may
generate better economic returns than large-scale commercial farms in some dryland
areas, although it may be harder for the state to take a share of pastoralist revenue
flows compared to income from large-farms.

12. Land investment should contribute to improved food security through generation of
foreign exchange, improved incomes as a result of on- and off-farm employment
created by investment projects, and food production that is marketed within Ethiopia.
However, if people directly lose their land without compensation or adequate
resettlement (including access to productive resources) then they will be likely to be
much worse off and more food insecure. Where there is a loss of access to resources
that are important parts of livelihood systems and coping mechanisms (forests,
rangelands, and water resources), then there are clear risks that there will be pockets
of greater food insecurity at the local level.

13. On paper, land investment in Ethiopia promises to create significant amounts of

employment. However, lack of implementation means that the number of jobs created
has not lived up to expectations, although this may change if investments are fully
operationalized. For a limited number of investments large amounts of employment
have been created. In developing regions these are often taken by workers coming
from outside the regions, rather than ethnic groups from the region. This has
contributed to conflict in some instances, and needs to be better addressed through
quotas and training programmes. However, the cultural difference between a
pastoralist lifestyle and wage labour employment on a plantation should not be
underestimated, the change in identity required would be resisted by many
pastoralists. As with other aspects of land investment creation of clear baseline
information, in this case on employment and livelihoods in land investment areas would
allow for monitoring of performance over time.

14. Sharing of information about regulations, land allocations and investors could be
improved, particularly between federal and regional governments, and between
government and the public. Senior regional officials, and even technical experts, in
some cases were unclear about the identity of major investors and the size or location
of land allocated by the federal government. Within regions a lack of clarity about land-
leasing regulations and roles of regional and federal government was evident.

5
15. The government has commendably made federal land agreements available on the
Ministry of Agriculture website. More information however could be made available, for
example, maps showing the exact areas for land leases, business plans related to
investments, environmental impact assessments, and monitoring reports on the
progress of different investments.

16. Data collection processes by government need support. For example, documentation
of which crops are being grown, expenditure on infrastructure and other aspects of
investment, and jobs created.

17. There are reasons to be concerned about the environmental impacts of land
investment, in terms of loss of forest resources, erosion of shallow soils, overuse of
agrochemicals and changes in water use and salinisation. Clear environmental
baselines need to be established to allow for meaningful assessment over time.

6
1. INTRODUCTION

Ethiopia is an important case in the international debate on large-scale land acquisitions. It

challenges assumptions about foreign dominance of land investment, or that large-scale land
investment is primarily for food crops for export, it is a case where land investment is central to
government agricultural strategy, but also where allocations have been subject to controversy
in terms of impacts on rights and livelihoods at the local level.

This study presents an inventory of land investments in Ethiopia1, and provides a narrative
description of key features of the large-scale land investment process in Ethiopia. The current
situation in relation to land investment is discussed including the rationale and process for land
allocation, the amount of land allocated and features of allocations to date including the profile
of investors, terms of leases, status of implementation of land investment agreements,
employment creation, and social and environmental impacts.2

The study is part of a larger four-country and international study on monitoring of large-scale
land investment conducted by the International Institute for Environment and Development
(IIED) for the Bill and Melinda Gates Foundation.

2. METHODOLOGY

Land investment in Africa is a sensitive and controversial subject, and many assertions are
made about land grabs or agricultural investment on the basis of limited evidence or poor
research methodologies. Given these pitfalls, this section sets out how this research was
carried out and some of its limitations.

Time and resources for this research were quite limited and the research is best viewed as a
rapid appraisal rather than an in-depth and extended piece of research based on substantial
fieldwork3. The research brief for this country study was to put together both an inventory of
land deals and to comment on features and outcomes of processes to date. Given the
importance of the inventory the team invested effort in building relations with the Ministry of
Agriculture in order to access data on land acquisitions and to secure introductions to regional
governments to access data from them.

A Memorandum of Understanding was signed with the federal Agricultural Investment Support
Directorate (AISD) in the Ministry of Agriculture (MOA). The AISD helped with provision of
some basic data on land agreements at the federal level (which was useful) and data for
regional land allocations (principally from investment agencies and not reflecting actual
allocations, and therefore of fairly limited use). We tried to access environmental impact and
monitoring data from the AISD, but were not able to obtain this.

1
See attached inventory document.
2
This Ethiopia country study was managed by James Keeley, IIED Research Associate, with regional
fieldwork carried out by Abdurehman Eid, Admasu Lokeley and Wondwossen Michago Seide. We
would like to thank Lorenzo Cotula, Fiona Flintan, Diana Grusczynski, Andrew Hilton, Carlos Oya and
one anonymous reviewer for comments on an earlier version of this report.
3
Research was carried out over a three month period July- September 2012, with around a week for
each region, and two to three weeks for meetings in Addis Ababa.

7
The AISD provided a letter of introduction to regional governments where limited fieldwork was
undertaken (these were Afar, Benishangul-Gumuz, Gambella, Oromia, Somali and SNNPR
regions). These introductions were very helpful and allowed us to interview a range of officials
and to access regional level data on land allocations. We are reasonably confident that the
data we have from regional governments are the best that are available on how much land has
been allocated, and to whom. The exception to this is Tigray where we relied on data from the
federal government, which is probably not as up-to-date as data for other regions.

The inventory data collection process set out to collect a range of data for each land investment
over 1000 hectares since January 2005, including information on investment identification,
location, timeline, investor details, land details, agricultural activity, job creation, fiscal regime
and infrastructure, and social and environmental impacts. In practice it was not possible to
collect much of this information as official data often only indicate how much land was
allocated, who the investor is, and possibly how much capital should be invested, how many
jobs are expected to be created, and the nationality of investors. Information on land use for
land leases is often only in terms of broad categories such as crop production or agriculture.
Data on actual employment, and categories of job, or the amount of land currently being
farmed, and social and environmental impacts, including numbers of households affected, and
processes of compensation were generally impossible to compile. Much of this data is either
not collected or not shared by government authorities. The inventory therefore is not complete,
and for some columns in the inventory spreadsheet information is available for some regions,
but not for others.

The second part of the research, features and outcomes, was addressed through over 50 key
informant interviews and a limited number of case studies of farm investments. Interviews were
carried out with: technical experts in the AISD, MOA, and in investment agencies, land and
agricultural bureaux at regional level, in Benishangul-Gumuz we also spoke to officials in the
President’s office; we spoke to woreda officials in SNNPR and Gambella. We spoke to a range
of company representatives4 primarily in Addis Ababa, but also in Gambella, Somali, SNNPR
and Afar, to donor representatives (World Bank, DFID, the Dutch government and the Finnish
government), to researchers, staff at the Ethiopian Wildlife Conservation Authority, the
Ethiopian Sugar Corporation and NGO officials with expertise in land issues in Ethiopia. The
majority of informants, however, were from government or from companies investing in land,
who for the most part take an optimistic view of the benefits of large-scale land investment; we
have tried to take this into account in our analysis. Given the range of issues covered it has not
always been possible to triangulate information from all informants to separate statements
which are fact-based from those which are merely opinions. We have tried to indicate in the
report when evidence in support of a particular statement is limited.

We carried out a limited number of case studies of farm-level investments: one in Gambella,
two in SNNPR, two in Somali region and two in Afar. These were selected on the basis of
accessibility, and willingness to engage with researchers. Case studies in Afar, Somali and
SNNPR involved discussions with community members, as well as company representatives.
We recognise that this is a very limited number of case studies and cannot be considered to be
a representative sample of agricultural land investments in Ethiopia. We are mindful of this and
use findings from this part of the fieldwork carefully to illustrate issues, but with the caveat that
more research at the farm and community level is needed to corroborate findings.

4
These included Karuturi, Saudi Star, S and P Energy, Whitefield and Adama companies.

8
Various written sources were used for this work. These included: federal land agreements
where they are available; various government regulations and proclamations on investment
and land; and some regulations and monitoring documents produced in Amharic. We also
reviewed a range of secondary literature including research reports and journal articles looking
at land investment in Ethiopia and more broadly. We examined relevant articles in Ethiopian
newspapers (mostly the English language press). As noted we were not able to access
government monitoring reports (with the exception of one region), we were also unable to
access company documents, such as business plans, and progress reports.

Another methodological issue is that it is difficult to comment with great rigour on some impacts
of land investment such as social and environmental issues, because investments are only
relatively recent, and because in many cases there is not a good baseline from which to assess
changes, for example, data on environment and land use is very limited for many areas.
Likewise when looking at employment, there is a lack of easily accessible information on wages
and economic returns to different livelihood options, in farm investment areas. Again it might be
possible to estimate how people were using various areas of land that are now given over to
land investment by putting together micro-level studies where they exist, but this would be
somewhat piecemeal, and for this research impossible in the time available.

3. BACKGROUND TO LARGE-SCALE LAND INVESTMENT IN ETHIOPIA

Ethiopia is a highly diverse country geographically, ethnically, linguistically and in terms of

livelihood systems. There are over 80 different ethnic groups in the country, of these the Oromo
are the largest accounting for around 40% of the population. The Tigrayans, Amhara and
Oromo together account for around two-thirds of Ethiopia’s 93 million population.

Ethiopian government is based on a system of ethnic federalism5. The country is divided into 9
regional states and 2 urban regions managed by the federal government. This system was
designed to ensure more autonomy for different ethnic groups than was experienced under
previous regimes, which were highly centralised and dominated by those coming from Amhara
region. Most regions are organised around a dominant ethnic group6. The federal system
means that there can be some variation in law and policy across different regions, including for
land. Of the 9 regions7, 4 larger regions are viewed as having satisfactory government capacity
(Amhara, Oromia, Tigray and SNNPR), and are categorised as ‘established’ regions. Other
regions including, Afar, Somali, Gambella and Benishangul-Gumuz are accorded special status

5
Ethiopia has been ruled by the Ethiopian People’s Revolutionary Democratic Front (EPRDF) since
1991. The EPRDF is a grouping of ethnically-based parties dominated by the Tigrayan People’s
Liberation Front (TPLF) which took power from the previous military government, known as the
Dergue, after a protracted civil war. The Dergue, a Marxist dictatorship, came to power in 1974,
overthrowing the imperial government of Emperor Haile Selassie.
6
This is the case for Afar, Amhara, Harar, Oromia, Somali and Tigray. Benishangul-Gumuz, Gambella
and SNNPR have a range of ethnic groups. In the case of SNNPR they are particularly numerous.
Zones in SNNPR can be organised around ethnicity (eg Sidama, Gurage), or woredas in the case of
South Omo Zone, SNNPR (eg Dassenetch, Nyangatom etc). The smallest ethnicities may only form a
dominant group at the level of the kebelle.
7
The federal system is organised into regions, below these are zones (which have greater or lesser
degrees of functionality in different regions), below this woreda (district), and below this kebelles
(former Peasant Associations).

9
as Developing Regional States, requiring special interventions in governance and economic
policy from the federal government.

Map 1: Ethiopia, with regions marked (source: UN OCHA)

The country roughly divides into two distinctive geographical areas: the highlands (Tigray and
Amhara and parts of Oromia and SNNPR), and a range of different lowland areas8.
Smallholders, who constitute around 80% of Ethiopia’s population, are mostly concentrated in
the highlands; these areas form the core of traditional Ethiopia, what was historically known as
Abyssinia. The rural lowlands are very diverse and are characterised by largely forested areas
that are home to shifting cultivators (Benishangul-Gumuz and Gambella), and semi-arid and
arid areas, or rangelands, that have been historically dominated by pastoralists, with some
cultivation in river valleys (Afar, Somali and parts of SNNPR).

Four main points can be made about the lowland/highland divide, which are relevant to the
subject of large-scale land investment. The first is that the lowlands only became integrated
into Ethiopian state relatively recently, during the reign of Emperor Menelik at the end of the
19th Century, and in some ways the process of full incorporation and state building is ongoing in
these areas (see Markakis, 2011; Makki, 2012; Hammond, 2011). There is an uneasy tension
between this process of state-building, which is linked to an idea of development in the national
interest, entailing some level of central control over remote territories, and the principle of
decentralisation and autonomy implied by ethnic federalism9. Land investment politics should
be understood in this context.

Secondly, most of the lowland areas are border areas, which are either themselves unstable or
influenced by political instabilities in neighbouring states (eg Eritrea, South Sudan, Sudan and
Somalia). Land investment in these areas forms part of these complex dynamics as well

8
Markakis (2011) adds to these the ‘highland periphery’ where the main highland plateau shades off
into lowland areas, this would include parts of Amhara, Oromia and SNNPR. Most lowland areas are
differentiated from highland areas by sudden dramatic changes in elevation.
9
See Lavers (2012) for an extended discussion of this point.

10
(Mosley, 2011). A third point is that pastoralist systems dominate large areas of the lowlands10,
and while the Ethiopian constitution respects the rights of pastoralists to pursue their
livelihoods, sedentarisation is encouraged, and statements are sometimes made by
government officials that pastoralism is seen as somehow backward and not desirable in a
modern Ethiopian state11. Land investment for crop agriculture is seen as a more productive
use of natural resources in these areas.

Finally, there is a long history of people moving within Ethiopia, particularly from the high-
population density highlands (where land is a very scarce resource) to lowland areas
(Pankhurst and Piguet, 2004). This could be voluntary or with little real choice for the people
involved. The Dergue government in particular resettled large numbers of people during the
1980s. The Voluntary Resettlement Programme of the current government sought to resettle
2.2 million people from the eastern and central highlands to western lowlands between 2003
and 2006 (Hammond, 2008). Regions such as Gambella and Benishangul-Gumuz now contain
large numbers of Oromo and Amhara that have moved from the highlands or highland
periphery in recent decades. Villagisation programmes are also being implemented in these
areas.

Large-scale land investment is an important part of the Ethiopian government’s strategy for
development of the country. Agriculture is at the heart of the country’s economy, contributing
50% of GDP, 85% of employment and 85% of exports. Nevertheless Ethiopia is chronically
food insecure, with significant food deficits each year. In the highlands plots are small,
dependant on erratic rainfall and low in productivity. Lowland livelihoods in rural areas are
heavily-dependant on pastoralism, agro-pastoralism or shifting cultivation.

Rapid agricultural transformation is planned involving innovation in the smallholder sector and
the development of the large-scale commercial sector. The Federal Government’s Growth and
Transformation Plan (GTP)12 envisages that Ethiopia becomes a food-secure, middle-income
country by 2025, and increases output of major crops from 19 to 27 million tonnes during the
period of the plan (FDRE, 2010a)13.

The GTP document notes that:

This strategy will support strongly the intensified production of marketable farm
products for domestic and export markets, by smallholders and private agricultural
investors. Fundamentals of the strategy include a shift to production of high value
crops, a special focus on potential high-productivity areas, intensified
commercialisation, and support for development of large-scale commercial agriculture
where it is feasible. The commercialisation of smallholder farming will continue to be
the major source of agricultural growth. Concerted support will be given to increase
private investment in large commercial farms [FDRE, 2010a: 22-3].

10
Pastoralist areas cover about 0.7% of Ethiopia and were around 12% of the population in 2005 (9.8
million people). Of this population 56% are pastoralists, 32% agro-pastoralists, and 22% urban
dwellers (EEA, 2005 and Mulat, 1998) cited in SOS Sahel, undated.
11
An often cited example of this is the statement by former State Minister of Agriculture, Aberra
Deressa, that: ‘We are not really appreciating pastoralists remaining as they are. We have to improve
their livelihood by creating job opportunities. Pastoralism, as it is, is not sustainable. We want to
change the environment.’ (Butler, 2010). See also Eyasu and Feyera, 2011.
12
The GTP is a five-year plan running from 2010/11 to 2014/15.
13
Agricultural output is expected to double from 20 million to 40 million tonnes by the end of the plan
period (MOA, 2012a).

11
Government policy documents suggest that Ethiopia has considerable potential in the
agricultural sector that is currently unfulfilled, not just in terms of addressing food security and
poverty reduction objectives, but as a core driver of national economic growth and job creation.
The Agricultural Transformation Agency, a special unit reporting to the Council of Ministers,
and supported by the Gates Foundation and others, has the mandate to identify bottlenecks in
the smallholder sector and support implementation of change by relevant stakeholders in areas
such as seed, fertiliser use, agro-processing and agricultural extension, and priority crop value
chains14.

The second component of the agricultural development strategy is to develop land that is
perceived as underutilised. Ethiopia covers an area of 111.5 million hectares, of this area 74.3
million hectares (according to official figures) are suitable for agriculture, but only 14.6 million
hectares are being used by smallholders (Mahlet, 2012). Irrigation potential is reckoned to be
4.3 million hectares, at present only 1 million hectares are irrigated (MOARD, 2009a).

It is against this background that the government has promoted large-scale land investment.
Large-scale commercial agriculture is particularly to be promoted in the lowland areas of the
country, with horticulture, labour-intensive agriculture and outgrower schemes promoted in
more densely populated agricultural areas (namely, the highlands)15. Large-scale commercial
agriculture is perceived by the government has having a number of clear benefits, including
promoting food security, creating jobs and transferring technology (see box 1).

Why invest in agriculture?

1. Boosts food security

2. Highly profitable business
3. Creates job opportunity
4. Promotes technology transfer
5. Export promotion
6. Enables capital accumulation
7. Works under and with nature
8. Environmentally-friendly business
9. Enhances land value

Box 1: Advertising poster produced by Agricultural Investment Support Directorate, Ministry of

Agriculture, Ethiopia

The GTP comments on the large-scale commercial farm sector as follows:

Large-scale farming will be undertaken by private investors in lowland areas where

abundant extensive land exists will be expanded and given due attention in the next
five years. The necessary arrangements will be made to increase the private investors’

14
There has been a shift in policy to concentrating relatively more resources on higher-potential parts
of the smallholder sector than was the case in the past: the flagship. For example, the World Bank (and
other donor) funded Agricultural Growth Programme is implemented in high-potential woredas.
15
According to the GTP: ‘The other element of the agricultural development strategy is the promotion
of private large scale commercial farms in areas that are not occupied or utilized by people’ (FDRE,
2010a: 23)

12
participation by identifying areas that are not inhabited but are suitable for agriculture.
Exploratory studies will be conducted to determine which forms of agricultural
production enterprises are most suitable for each area identified. These areas and the
data concerning them will be registered and organised in a land bank… The necessary
support will be given to encourage the participation of Ethiopian investors. Efforts will
be made to attract foreign investment in a manner that will be beneficial for Ethiopia’s
agricultural sector development [FDRE, 2010a: 54]

The GTP also specifies that Ethiopia will expand production of industrial crops (such as cotton,
sugar, rubber and palm oil). Production will increase from 0.7 million tonnes to 1.2 million
tonnes (FDRE, 2010b). The GTP sets out that:

While supporting private investment in large scale farms, government’s focus is to

ensure that the products produced from these farms are primarily for export or raw
materials for domestic industries. For these reasons, emphasis will be put on cotton,
date palm, tea, rubber tree and similar types of crops [FDRE, 2010a: 55]

The government has targets to dramatically increase sugar production. The country currently
produces only 60% of consumption (importing 150,000 tonnes in 2010) and plans to increase
production eightfold to 2.3 million tonnes, with a surplus of 1.25 million tonnes for export,
making Ethiopia one of the top-ten exporters in the world (Davison, 2011a).

There are also clear biofuel production targets with 182 million tonnes of bioethanol expected
to be produced from the 10 sugar factories and plantations currently being constructed under
the sugar intensification plan. There is also a plan that all diesel will contain 20% biodiesel by
2015, requiring increased production of suitable feedstocks (Davison, 2011a).

Ethiopia is seen by many as having a particularly favourable investment regime in certain

sectors, particularly for foreign investors (MAI, 2012; Dessalegn, 2011)16. According to the
Ethiopian constitution, land is owned by the state and the people rather than private individuals
or communities. This facilitates allocation to investors. Policies on income tax, capital
requirements for investment, repatriation of profits, import duties and land lease rents are all
seen by many as favourable for private investment17. These are discussed below.

However, policies have also been subject to criticism. Land allocation is seen by many critics
as too rapid, without adequate capacity for assessment and monitoring of investments and
biased towards foreign investors (Dessalegn, 2011). The largest parcels of lands and the core
of the governments land bank is in the lowlands (Benishangul-Gumuz, Gambella, SNNPR, Afar
and Somali regions and parts of Oromia). Land chosen in these areas is presented as ‘empty’,
despite claims to the contrary that it forms part of pastoralist land use systems or shifting
cultivation systems, often by minority ethnic groups. In some areas the government is
implementing policies of villagisation (Benishangul-Gumuz, Gambella and Somali regions),
relocating households to enable better service provision. Critics argue that villagisation is partly
designed to clear land for land investment (HRW, 2012a; OI, 2011). However, interviews for
this research have not confirmed this, neither have field visits by donor assistance groups.

16
Although Ethiopia ranks number 127 in the World Bank Ease of Doing Business ranking of 185
economies, scoring particularly badly on starting a business and trade across borders. See:
http://www.doingbusiness.org/data/exploreeconomies/ethiopia/.
17
This point is made in a presentation by Morrell Agro-Industries, a large commercial land investor in
the country (MAI, 2012).

13
4. THE PROCESS FOR AGRICULTURAL LAND INVESTMENT

According to the 1994 Constitution, all land in Ethiopia is owned by the state and the people.
This means that the federal and regional governments have a key role to play in managing the
land investment process. It also means that land can only be leased, rather that bought and
sold18. According to Dessalegn, Ethiopia lacks a ‘robust system of land tenure’, farmers only
have rights to rent land, and land use is subject to several conditions and possibility of
expropriation for private investment (Dessalegn, 2011). The 2005 Federal Proclamation on
Land Administration and Use declares that: ‘the government as sole owner of rural lands may
change communal holdings to private holdings as may be necessary’19. Processes of land
certification have been carried out in Amhara, Tigray, Oromia and SNNPR regions, and are
currently underway in Benishangul-Gumuz and Gambella. While the constitution asserts that:
‘Ethiopian pastoralist have a right to free land for grazing and cultivation as well as a right not to
be displaced from their own land’ 20, certification of land use for pastoralist or other uses is
limited, largely because no effective system has been developed to certify group rights21.

The Ethiopian constitution gives regions power to administer land within the region, consistent
with the constitution and federal laws. This is reinforced in the Federal Rural Land
Administration and Use Proclamation (Proclamation 89/1997) which allows regions to make
laws to manage and administer land within their region. Administration includes determining
systems for expropriation and compensation, land rental, communal rights, and land use
planning. This vesting of powers at the regional level means there can be great diversity
between regions in terms of rules, practices and incentives for different types of investment.
For example, as discussed below land lease fees vary between regions, and are paid to the
regional government in some cases (for example, Benishangul-Gumuz)22 and to the woreda
government in others (in SNNPR and Gambella). The length of land leases also varies for
different regions23.

In 2009, the federal government decided to more actively encourage large-scale land
investment and to improve various aspects of technical management of land investment. In an
upward delegation of a regional mandate, the Council of Ministers issued a proclamation
(Proclamation 29/2001 EC) that plots over 5000 hectares would be administered by federal
authorities and included in a land bank.

18
Article 40.3 of the Ethiopian constitution states: ‘The right to ownership of rural and urban land, as
well as of all natural resources, is exclusively vested in the State and in the peoples of Ethiopia. Land is
a common property of the nations, nationalities and peoples of Ethiopia and shall not be subject
to sale or to other means of transfer.’ (FRDE, 1994).
19
Article 5.3. (FDRE, 2005).
20
Article 40.5 (FDRE, 1994)
21
Land certification in Benishangul-Gumuz will certify land for shifting cultivators with existing
cultivators given 10 hectares within which to practise and new households 3-5 hectares. Prior to this
shifting cultivators might have operated over much larger areas.
22
In Benishangul-Gumuz, lease fees go to the region and are returned to the woredas with 20 per cent
extra paid to the region where an investment is located.
23
In Benishangul-Gumuz this is 25 years, and Amhara up to 40 years, for example (MFAF, 2012).

14
Five key regions for land investment were chosen and asked to identify parcels of land of 5000
hectares and above that are suitable for large-scale commercial agriculture. A total of 3.31
million hectares was identified in 2009 (MOARD, 2009b) in Afar, SNNPR, Gambella,
Benishangul-Gumuz and Oromia (see chart 1). This land bank was based on satellite imaging
of the regions and was imprecise about communities or important natural resources on the
ground. In Benishangul-Gumuz, the area in the land bank technically now stands at 1.4 million
hectares, but in practice only around 300,000 hectares have been subject to further verification
in 10 specially identified woredas. Maps of the land bank have not been made publicly
available. This research team asked for current figures for different regions in the land bank
and was told they were unavailable as they were currently under revision.

Land available in federal land bank (in 2009, hectares)

Afar, 409,678, 12%

SNNPR , 180,625,
Oromia, 1,200,000,
5% Afar
37%
SNNPR
Gambella
Gambella, 829,199, Benishangul-Gumuz
25% Oromia

Benishangul-
Gumuz, 691,984,
21%

Chart 1: Land available in the federal land bank (hectares), MOARD, 2009b

The land bank is managed by the Agricultural Investment Support Directorate (AISD), a
directorate in the Ministry of Agriculture. The role of the AISD is to create a more coherent land
investment system than previously existed, put together information and attract investors,
provide technical support (such as use of GIS, surveying and mapping, capacities that regional
governments generally do not have), streamline the investment process, assess the capacity of
investors, and carry out auditing and monitoring of investments.

The land investment process entails first submitting an application to the Ethiopian Investment
Agency detailing the proposed project, the capital to be invested, employment creation,
marketing plans, and utility and raw material requirements. The investor then receives a foreign
investment licence. Following this, a land use agreement is then developed with AISD. The
process when working with the AISD is that the investor puts together a business plan, AISD
investigates the capacity of the investor, including technical competency and financial capacity,
then land is identified from the land bank, a feasibility study is carried out, and then a land
lease agreement contract is signed. The land use agreement specifies terms and conditions,

15
such as the need to carry out an environmental impact assessment within 3 months of signing
the land use agreement, the land rent, and any requirements to develop land within a certain
period, as well as arrangements for termination of the agreement.

Land use agreements made with AISD are available on the MOA internet site, in English in the
case of international investors, and Amharic for diaspora and local investors. These
agreements specify the woredas and possibly kebelles where investments are located, but not
precise grid references. There are no publicly available maps showing where investments are

16
located in different regions. We were shown one such map in one region, but told that we could
not reproduce it for this report.

The change in responsibility for allocating large concessions of land meant that in one case a
land deal previously agreed by a regional government was revised by the federal government.
As noted, the Indian agribusiness firm Karuturi was allocated 300,000 hectares by the
Gambella Regional Government in 2008, but this was subsequently reduced to 100,000
hectares by the federal Government. It seems that this is the only case where the federal
government specifically overruled a deal made by a regional administration.

17
The exact boundary between regional and federal responsibilities seems to be a subject of
some confusion. Some officials interviewed in Benishangul-Gumuz asserted that foreign
investors always went through the federal government, as well as any domestic investor
seeking more than 5000 hectares of land. Others said that foreign investors could go through
the region for units of land below 5000 hectares. Still other government informants said that
there was a federal land bank of 300,000 hectares that had been demarcated, and a regional
land bank of 20,000 hectares for 2012/13 allocation; and for land from the federal land bank -
irrespective of size or nationality - the point of contact was the federal government, and for the
latter the regional government. Others claimed that the federal/regional cut-off size was 3000
hectares not 5000 hectares. These views were all from regional government officials and
illustrate that there needs to be much clearer elaboration and sharing of information on
regulations, the nature of different land banks and roles and responsibilities of different levels of
government.

Another argument made by regional government is that the transfer of land to the federal
government for administration by MOA was necessary because of lack of capacity at the local
level. With improved capacity, as evidenced by the creation of new land bureaux and better
regional land use planning frameworks and land administration regulations at the regional level,
the justification for federal engagement will be less clear and more authority will need to return
to the regional level. The difficulty in securing land for the federal land bank in Benishangul-
Gumuz and Gambella is referred to in the recent performance assessment report for
agricultural investors published by the MOA. The study notes:
Pertaining to administrative bodies of regional states, especially in Gambela and
Beneshangul (sic) regional states; though aware of the significance of transferring land
to the federal government executive bodies were deficient in diligently working towards
that end. [MOA, 2012b]
This relates to the wider argument noted by Lavers (2012) and others that the transfer of
powers from the regions to the federal level in relation to land sits uneasily with the principles of
regional autonomy on which ethnic federalism is based. Lavers argues that this is felt most
keenly in regions such as Oromia where there is more capacity.

In practice, it seems that the federal government has built up capacity and experience in
dealing with large-scale – and, particularly, international - investors that regional governments
can’t match and a division of labour has some justification. It is evident however that there
needs to be better sharing of information between the centre and the regions. In Benishangul-
Gumuz almost all senior government officials were unclear about the second largest land lease

18
in Ethiopia situated in their region24. Most were unable to name the investor, the exact size of
the lease, and the precise location, or status of the investment. While there seemed to be some
joint monitoring between federal and regional government, federal officials on occasion seemed
to go directly to woreda government, bypassing the region. This lack of awareness in the region
is problematic in terms of good development planning or strong regional ownership of
agricultural development activities.

There is also variation in regional land allocation practices in different regions. Until recently in
Afar investors dealt directly with clans and there was no role for regional government –
agreements were made between clan elders and investors, and these agreements had legal
status in court. A share of the output would go to the clans as payment. Now there is regional
land administration policy (since 2011), and a bureau of land administration has been set up.
Also in Afar rents are collected through land taxes, and no land lease rates have yet been
formulated. In some regions investors identify land and inform the Regional Investment
Commission, in other cases, such as Oromia, the Regional Land and Environmental Protection
Bureau would play more of a role. In Somali it is the Investment Bureau. In Afar the Investment
Directorate is under the Trade Bureau.

In Benishangul-Gumuz, until recently it was the Investment Bureau that allocated land to
investors, now the bureau grants a license and investors apply for land from the land bank,
through the new Bureau for Environmental Protection and Land Use and Administration, and a
decision is taken by a specially convened Investment Board chaired by the regional president
(with representatives from water, energy, agriculture, land and trade bureaux).

Land lease agreements specify timelines for development of the land investment. Land
development should start within 6 months of signing a lease agreement or on receipt of
government approval documents. One third of the plot should be developed in one year, and
the whole plot within 3 years. Government can annul a land lease if land is not developed
within these specified periods. There are cases where areas of land have been taken back, or
where agreement to lease more land has not been approved. Generally, expansion of a plot is
only allowed when the plot is brought fully into cultivation.

Monitoring of land investments has been weak, but officials claim that the situation is now
gradually improving. Monitoring visits are supposed to be carried out on a quarterly basis (for
all land over 200 hectares for perennial crops, and 500 hectares for annual crops). For lands
allocated from the federal land bank monitoring visits can be monthly, and for some investors a
weekly report is completed. Spot check farm visits are also carried out.

At the regional level monitoring seems to be constrained by available resources. Therefore in

Benishangul-Gumuz, where the areas are large and vehicles are limited, it is difficult to visit
investors regularly. In Oromia government offices, informants suggested that investors near to
the regional government offices in Addis Ababa were more likely to be visited than parts of the
region that are more remote.

Monitoring in many regions focuses primarily on whether the investor has started to farm or not,
and whether the land is being used for the correct purpose. Assessments of employment,
environmental and social impacts, and productivity are less likely to happen. Additional case
based monitoring may happen when there has been an incident or some other reason to check

24
This is S and P Energy, an Indian firm allocated 50,000 hectares to grow biofuels.

19
on an investment. In one region officials only seemed to know where investors were on paper,
but not whether they had really started to farm or not. At woreda level government officials
complain that they lack documents and resources to monitor, and sometimes they are not clear
who the investors are in their district.

In Benishangul-Gumuz there is now an annual investor forum when investors in the region are
called to the regional capital to discuss progress and investment issues with government. The
government in this region has also imposed a moratorium while they investigate the status of
current land leases, and verify the size of plots.

5. HOW MUCH LAND HAS BEEN ALLOCATED TO INVESTORS?

Compilation of data for this review suggests that around 1.06 million hectares of land have
been allocated in Ethiopia for commercial agriculture (see charts 2 and 3)25. This land is
allocated in three main ways: by the federal government to commercial investors from the
federal land bank, by regional governments, and by the federal governments for state-run
sugar plantations. Each of these categories accounts for roughly one-third of total allocations.

Large-scale land allocation in Ethiopia (by region, hectares)

54,000, 5%
51,544, 5%
121,370, 11% Afar
Amhara
Benishangul-Gumuz
348,009, 34%
160,630, 15% Gambella
Oromia
Somali
SNNPR
26,000, 2%
Tigray
22,300, 2% 272,112, 26%

Chart 2: Large-scale land allocation in Ethiopia by region

25
As explained above, data for the federal MOA allocations comes from the MOA. Data for sugar
lands comes from Ethiopian Sugar Corporation. Regional data comes from regional governments, with
the exception of Tigray where data was provided by the federal MOA.

20
Land alllocations from different sources by region (hectares)

250,000

200,000

150,000 Regional land allocations

Sugar land allocations
100,000 Federal land bank

50,000

0
uz

PR
lla
ra

ia

ay
ar

i
al
um

m
ha
Af

be

gr
N
ro

So

Ti
Am

SN
-G

am

O
ul

G
ng
ha
s
ni
Be

Chart 3: Allocation of land in Ethiopia, all sources (minimum unit size, 1000 hectares)

While these land allocation figures look very precise, we recognise that they need to viewed
with some caution. There are several reasons for this (see Box 2). Datasets often fail to
distinguish between areas of land where there has been an MOU, or initial agreement, or even
a request for a certain amount of land, and land contracts where land has been allocated to an
investor. For example, initial data provided for this project by the MOA suggested that 422,000
hectares of land had been provided in Benishangul-Gumuz region (including only deals over
1000 hectares), with two allocations of 100,000 hectares, one to Awadh Tareef Mohammed
Alkutbi and the other to Al Sharif Mohammed. Likewise in Oromia a total of over 600,000
hectares appeared to have been allocated with individual deals of 140,000 hectares (Bon
Green Farm) and 150,000 hectares (Watany Agricultural Food Production) respectively.
Further examination revealed that these were not land agreements, but only amounts
requested by the investor when applying for an investment license. In fact, no amounts of land
this size were transferred in the regions, and there is only one deal over 100,000 hectares in
the whole country. Given this, the regional land allocation totals for Benishangul-Gumuz and
Oromia are much lower than these figures indicate.

Even when data relates to an actual agreement where land has been accessed, the amount of
land that an investor takes may be more or less than that officially recorded. Sometimes
agreements are cancelled but still included in datasets of commercial land concessions. For
example, Hunan Dafengyuan is sometimes cited as an example of a large-scale Chinese
investor, farming 25,000 hectares of sugarcane in Gambella. In fact the agreement with has
been cancelled, although details of the contract are still available on the MOA website 26.

Sometimes datasets include figures given by companies of the total area they plan to farm, for
example, a figure of 300,000 hectares for Karuturi in Gambella is often quoted, a figure which
seems to be based on an early agreement with the regional government which has been
superseded by a smaller (but still very large) agreement for 100,000 hectares with the federal
government. There can also be a delay in recording new deals in official datasets. Sometimes

26
See also OI, 2011, which has a useful annex on these methodological problems

21
there can be double-counting and recording errors, for example, failure to record a land
concession, or inaccurate recording.

Challenges in establishing accurate figures for commercial land allocations

 Figures may be only an MOU, no lease exists

 Figure may be an expression of requested land, for example from an investment agency
 Lease may have been cancelled
 Double-counting in a dataset (for example, a federal land agreement in a list of land
agreements for a region)
 Recording errors, for example, lease not recorded or inaccurately recorded
 Only a small part of an agreed area is actually farmed
 Delays in updating databases (often only done annually)

Box 2: Challenges in establishing accurate figures for land allocations

Figures for federal land allocations (via AISD, MOA) to the private sector are precise in terms of
land allocated (385,000 hectares), and overall status of the investment (operational or pre-
implementation). Although the category operational covers anything from all the lease being
farmed to only a very small percentage (for example, S and P Energy, has had a lease of
50,000 hectares to farm pongomia (a biofuel) in Benishangul-Gumuz for several years, but on
the ground only 2,500 hectares is actually being farmed27.

Regional land allocations are much more imprecise, as indicated above, and because there
has been a lack of good documentation of allocations and current status of land use at this
level. For example, in Somali region the investment bureau grants a licence and then the
investor visits a district and makes a local agreement, but there is no regional agreement on
what is actually allocated. Taking these kind of problems into account, we suggest around
335,000 hectares have been allocated by regional governments. This is substantially lower
than a figure of 800,000 to 2 million hectares that are sometimes quoted. This is partly because
this research uses a 1000 hectare threshold to emphasise large-scale land investment, and
partly because records of regional deals seem most prone to error (particularly confusing
expressions of intent with actual allocations).

In addition to regionally allocated land, large amounts of land are allocated by the state to state
corporations for industrial crops, particularly sugar and rubber, these are large areas of land,
but often not included in official inventories of land investment. At least 335,000 hectares have
recently been made available for sugar development (including 70,000 hectares from the
federal land bank), in addition to earlier allocations of 50,000 hectares in Afar. Reportedly,
200,000 hectares is also being made available for rubber in SNNPR and Gambella (Muluken,
2012), although we have not been able to confirm this.

Our figure of 1 million hectares allocated in Ethiopia for commercial farming is much lower than
that suggested by some other studies. This is because the study only includes deals since
January 2005, it does not include deals below 1000 hectares, which added together can
amount to large sums, and research has tried to be precise about only including land

27
Interview with S and P company manager, July 2012.

22
agreement figures, rather than amounts of land requested from government, or figures in
MOUs.

The figure for land allocation given by the government in recent press statements is 2.2 million
hectares with only 372,088 hectares of this developed to date28. According to the government
11,773 domestic and foreign agricultural investors have been licensed, with cumulative
registered capital of 132 billion birr (US$ 7.3 billion). 5284 of these investors have received
land, of these 126 are foreign investors. Over one quarter (27%) of the 2.2 million hectares is
allocated to foreigners (73% to domestic investors). According to this statement, 567,651
hectares have been distributed among foreign investors (Eskedar, 2012: 14).

Some researchers contest the publicly declared figures. Dessalegn argues that overall land
allocations amount to 3.5 million hectares, and that by the end of the GTP 7 million hectares
will have been allocated (equal to 38% of the land farmed by smallholders) (Dessalegn, 2011).
The Oakland Institute put the figure at just over 3.6 million hectares of land leased by January
2011 (OI, 2011).

A study for the World Bank identified 406 land investments over 500 hectares between 2004-8
and found that 1.2 million hectares were allocated (for Benishangul-Gumuz, Gambella,
Amhara, Oromia and SNNPR). 23 of these projects (or 5.7% of the total) were foreign
investments. These equalled 51% of the allocated land area (World Bank, 2010). A study by
the International Institute for Environment and Development (IIED) in 2009 found that 602,760
hectares were allocated between 2004 and 200929, with investment commitments of US $
78,563,02330 (Cotula et al., 2009).

More recently the International Land Coalition has compiled an International Land Matrix portal
(www.landportal.info). For Ethiopia, they suggest that 2,412,562 hectares have been allocated
in 56 deals over 200 hectares since 200031. ILC note the difficulty in checking and verifying the
status of reported deals on the website, and the Ethiopian data appears to include several
errors, for example, the Karuturi land agreement in Gambella should be 100,000 not 300,000
hectares, and there is no government record of 600,000 hectares being allocated to Varun, or
100,000 hectares to 2H 25 International Business. Also the 25,000 hectares allocated to Hunan
Dafengyuan has been cancelled. The database also provides no detail of which region of the
country the land is allocated in or whether it is a federal or regional allocation, making it hard to
cross-check entries.

28
Statements by Tarekegn Tsigie, MOA spokesman (Eskedar, 2012: 14)
29
These were deals over 1000 hectares, a total of 157 projects.
30
However, this data was largely from the Ethiopian Investment Agency and it is unclear how much
was actually transferred or put to use on the ground.
31
Available at: http://landportal.info/landmatrix/get-the-detail/by-target-country/ethiopia

23
6. FEATURES OF LAND ALLOCATIONS

6.1 Size of land allocations

At present, of 131 land agreements by regional and federal governments (excluding state sugar
plantation land allocations) there are 17 deals that are 10,000 hectares or over, and only 2 that
are 50,000 hectares or over (see chart 4)32.

Number of leases of different sizes

140 131
Number of land leases

120
100
80
60
40
16
20 6 9
1 1
0
100 50 to 100 25 to 50 10 to 25 5 to 10 1 to 5
Size of lease (1000's has)

Chart 4: Regional and federal land deals by different size category

6.2 Where is land being allocated?

Three clear trends are evident in relation to the location of large-scale land concessions. The
first is that just under 80 per cent of all land leased out or transferred in Ethiopia is in lowland
regions, principally Benishangul-Gumuz, Gambella and SNNPR. For federal government
allocations to private investors, all land allocated is in these lowland regions33.

The second trend is that the large amounts of land are allocated by regional governments in
Amhara, Oromia and Tigray, regional states in the highlands34, which do not have special
developing status, and where government capacity is much greater. Land in these states is not
allocated by the federal government35. Average plot sizes are low.

A third feature is that if lands allocated for sugar production are added to total land allocations,
SNNPR becomes the region with the largest amount of land allocated to commercial
agriculture, by some way (a total of 33 per cent of all land allocated in Ethiopia). However, most
of this land has only been demarcated, and only a small amount is under the plough.

32
These figures are based on land agreements where we have a figure for land allocated, and not
MOUs.
33
The one exception to this is the federal government land agreement with CHC, where some of the
land is in Benishangul-Gumuz and some in Amhara
34
Parts of Oromia are lowland
35
The exceptions to this are sugar concessions and the case mentioned above in footnote 15

24
An analysis of leases from the federal land bank managed by the Agricultural Investment
Support Directorate shows that by September 2012, 385,000 hectares of land had been
allocated under 31 different land lease agreements36 (see chart 5 and table 1). Of these 31
projects, 24 are operational and 7 in the pre-implementation stage. 97% of this land was
allocated in just three regions: SNNPR, Gambella and Benishangul-Gumuz, with just over half
the land allocated in Gambella (53%), and 26% allocated in Benshangul-Gumuz and 18% in
SNNPR. Land allocations were also on average much larger in Gambella, with 8 leases with an
average size of 25,751 hectares. For Benishangul-Gumuz, there are 13 leases with an average
size of 7610 hectares, for SNNPR, 10 projects with an average size of 6836 hectares.

No land allocated has been allocated in Afar by the federal government, despite the existence
of 400,000 hectares in the Federal land bank being advertised for investment by AISD. This is
possibly due to the high risks of land conflict with pastoralist clans in these areas.

Land leased to commercial investors from the federal land bank

(hectares)

Amhara , 12,500,
3% Benishangul-
Gumuz, 98,937,
SNNPR, 68,360, 26%
18%
Benishangul-Gumuz
Gambella
SNNPR
Amhara

Gambella, 206,012,
53%

Chart 5: Allocation of land by region from the federal land bank

36
This includes one land agreement which has been cancelled

25
Region Number of Percentage of Average size Total land
leases land allocated allocated
Benishangul-Gumuz 13 26 7,610 98,937
Gambella 8 53 25,751 206,012
SNNPR 10 18 6,836 68,360
Amhara 1 3 12,500 12,500
TOTAL 3137 100 12,445 385,803

Table 1: Allocation of land by region from the federal land bank

Land investment in lowland regions (Benishangul-Gumuz, Gambella and SNNPR) is

associated with significant social and economic change. These areas have low population
densities, particularly compared to highland regions (Amhara, Tigray and central parts of
Oromia). They also have many ethnic minority groupings, which are quite distinct in terms of
language, livelihood systems and culture from highland Ethiopia. Gambella and Benishangul-
Gumuz are Developing Regional States, partly administered from Addis Ababa by the Ministry
of Federal Affairs. Data show that 46.1% of all land allocated in Ethiopia over 1000 hectares in
size is in Developing Regional States38. Land in these areas is often perceived as unused or
free land, indeed the federal government AISD claim that it only allocates land that is
uncultivated and which is not part of existing livelihood systems 39.In contrast with Oromia
where many Oromo farm the land, or Amhara in Amhara region, in Benishangul-Gumuz and
Gambella, no major investors are ethnic groups from the region.

Areas such as Gambella have historically been marginalised parts of the political economy of
Ethiopia, with the arrival of land investment they are now receiving much more attention, and
are viewed by many as strategically important. Infrastructure is gradually being constructed,
and the number of flights going to Gambella City and Assosa (capital of Benishangul-Gumuz)
each week has increased (although not solely due to land investment). Gambella is the location
of the two most highly publicised land investments in Ethiopia, Karuturi and Saudi Star (owned
by Sheikh Mohammed Al-Amoudi, a major investor in many sectors of the Ethiopian economy)
and has become part of a globalised discourse around ‘land grabbing’ and human rights (see
HRW, 2012a; OI, 2011; Pearce, 2012).

Regional government land allocations present a slightly different picture from federal
allocations (see chart 6), with more land allocated by developed regions, namely Amhara,
Oromia and Tigray (19% of the total), and relatively less by developing regions such as
Gambella and Benishangul-Gumuz (32%). The largest amount of land allocated by a regional
government is by SNNPR (31% of the total).

37
Note that as one lease is in both Amhara and Benishangul-Gumuz the total number of leases is only
31, see footnote 15.
38
The Developing Regional States are Afar, Benishangul-Gumuz, Gambella and Somali.
39
Discussion with technical expert in AISD

26
Land allocated by regional governments (hectares)

Tigray, 6,544, 2% Afar, 34,000, 10%

Amhara, 33,870, Afar
SNNPR, 104,649, 10% Amhara
31%
Benishangul-Gumuz
Benishangul- Gambella
Gumuz, 41,693, Oromia
12% Somali
Somali, 26,000, 8% SNNPR
Tigray
Oromia, 22,300, 7% Gambella, 66,100,
20%

Chart 6: Land allocated by regional governments

As noted, regional figures are complicated because unlike land allocated by the federal MOA,
significant amounts of land are allocated below the 1000 hectare cut off point used for this
research. If the cut off point were lower then the area of land allocated by Amhara, Oromia and
Tigray would increase substantially, and the share of these regions in all regional land
allocations would increase, as would the total number of projects in these regions. Secondly,
the 2005 cut off date used for this project is also significant. Many small amounts of land were
granted to coffee investors in SNNPR before 2005, constituting a large percentage of
investment licences from 1998 onwards. However these are not included here. Finally, as
discussed above information on regional land allocation is the least precise. Record keeping
with many regional governments has not been good, and until recently at least there has been
a lack of clarity about how much land has been allocated. Large amounts of land leased by
regional governments in all regions are not being farmed. At present many regions are trying to
assess exactly how much land has been allocated and how much is being put to productive
use.

Some regions have allocated relatively little land despite being identified as having high-
potential for commercial agriculture. Very little land is allocated in Somali, for example, despite
this being area with high irrigation potential. Conflict in the region and difficulties ensuring
agreements with clans are key reasons for limited land investment cited by officials during
interviews. Additional factors include an unclear boundary with Oromia region and the absence
of a regional land bank.

As noted above, in Ethiopia, there are separate systems for allocation of land for highly
strategic crops. The main example is sugar. When sugar lands are added to figures for total
commercial land investment then far more land is allocated in SNNPR than other regions –
335,000 hectares (see chart 7). Map 2 indicates the main sugar producing regions of the
country.

27
Sugar production area by region (hectares)

Tigray, 45,000
SNNPR
Afar
Amhara, 75,000 SNNPR, 175,000 Benishangul-Gumuz
Amhara
Benishangul- Tigray
Gumuz, 20,000
Afar, 20,000

Chart 7:: Allocation of land for sugar production

Map 2: Ethiopian Sugar Corporation map of major sugar production areas.

28
6.3 Profile of investors

The profile of investors differs for federal and regional land allocations. At the federal level,
where allocations are much larger, international investors are more prominent in terms of area
of land acquired, but not projects if diaspora investors are grouped with Ethiopian investors. At
the regional level domestic investors dominate. The state is also a major investor in land
through its allocation of land for sugar plantations.

Of 31 lease agreements with the MOA from the federal land bank, 12 are with Ethiopian
investors, 8 with diaspora investors and 9 with Indian investors (see chart 8 and table 2). In
addition there is one Turkish and one Saudi investor40. As noted, there was a federal
agreement with a Chinese company but this was cancelled. The importance of diaspora
investors is a particular feature of land investment in Ethiopia, which may be less relevant in
the inward land investment profile of other low income countries.

Federal leases by nationality of investor

300,000 275,012

250,000
Area (hectares)

200,000

150,000

100,000 72,223

50,000 22,568
8 12 9 1 6,000 1 10,000
0
Diaspora Ethiopia India Saudi Arabia Turkey
Number of leases, area

Chart 8: Federal land bank investors by nationality or diaspora status

Origin of investor Number of projects Total area Average size (ha)

Diaspora 8 22568 2821
40
As noted elsewhere the Saudi company is owned by Mohammed Al-Amoudi, a Saudi citizen, who
was born in Ethiopia with an Ethiopian mother, and is the most important private sector actor in many
sectors of the Ethiopian economy.

29
Ethiopia 12 72223 6019
India 9 275012 30557
Saudi Arabia 1 6000 6000
Turkey 1 10000 10000
TOTAL 12445
31 385803

Table 2: Origin of investors with agreements with the federal government

Regional government land allocations, Ethiopian and foreign investors

Foreign , 11, 11%

Ethiopia
Foreign

Ethiopia, 93, 89%

Chart 9: Regional government land allocations by area, Ethiopian and federal investors

At the regional level data is harder to disaggregate by nationality. For the agreements where
we have data 93 agreements are with domestic investors and only 11 with foreign investors.
These foreign investors come from Israel, Italy (two agreements), Malaysia, Netherlands (two
agreements), New Zealand, Saudi Arabia (two agreements), Turkey and USA (see chart 9).
Around 23% of land allocated by regions is to foreign investors, a higher percentage than that
for the number of deals with foreign investors with the 12.

While we have not made a specific comparison we anticipate that this profile of nationality of
investors would be very different to that for other African countries, where international
investors would be more significant. Investor profile illustrates the importance of the role of the
state in controlling access to land and in determining the terms on which private and
international capital is able to access land resources.

Most investors are private individuals. Certainly all investors with the federal government are
private companies, as opposed to sovereign wealth funds, or national companies (it would
seem unlikely that sovereign wealth funds are funding any of the private companies in question
either). Most funding for land deals appears to be corporate capital, and commercial bank loans
in some cases.

30
6.4 Crop focus for large-scale land investment

The focus of land investment varies depending on whether land is allocated by regional
governments or federal governments and depending on the comparative advantages of the
region.

For many regions data on crops being grown on land allocated by regional governments is very
limited. One exception is Afar where all land investments over 1000 hectares are for cotton
production. Reasonable data also exists for Somali and SNNPR. Much of the data for other
regions only indicates that the land use is crop production as opposed to livestock, without
specifying categories of crops, or particular crops. Sometimes data indicates that several crops
are grown and the relative amounts are unclear. Even where crops are specified this does not
necessarily match with what is actually being cultivated. It should also be noted that flowers
and coffee are important investment crops in Ethiopia, but as they are usually grown on areas
below the 1000 hectare lease investment threshold used for this report they do not show up in
the data, even though the total area planted is significant.

In general, for regional allocations crop choice seems to be up to the investor, although tougher
review processes for investors may in the future mean that crop choice has to be more
carefully justified in terms of suitability of land use and strategic relevance. In Benishangul-
Gumuz region, it was suggested that investors were sometimes planting crops unsuited to the
agro-ecological conditions, but that in the future soil testing services by the regional
government would help ensure that more suitable choices were made.

For strategic crops managed by the federal government land allocations are relatively clear. As
noted sugar is the major crop in this category and amounts to around 30 per cent of all large-
scale land allocations.

Federal land bank allocations by crop type

152,360
160,000
140,000
Area (hectares)

120,000
100,000 80,000
80,000
60,000 50,000 50,000
35,000
40,000
15,000
20,000 3,012
0
Cotton Edible oil Biofuels Sugarcane Tea Cereals Other
crops
Crop type

31
Chart 10: Federal land bank allocations by crop type

While there are targets for other industrial crops such as biofuels (excluding sugar which is
being grown both for edible consumption and for ethanol), they are not subject to such
systematic land allocation processes. For biofuels, it is claimed that 85,000 hectares are
currently being grown (Davison, 2012a), although this is difficult to verify. There is only one
specific biofuel project among the 31 federal Ministry of Agriculture land leases and this is
50,000 hectares allocated for pongomia cultivation, although one sugar and one oil palm
cultivation project could also be for biofuel feedstock production. If this were the case then 22%
of land allocated by the Federal Land Bank would be for biofuels, this needs further verification
however. Interviews suggest that the S and P project is still in early stages and pongomia is not
yet being cultivated on the land allocated. Other biofuel producers such as Sun Biofuels
(operating in SNNPR) have had their license cancelled.

Federal land lease documents usually specify what crop is to be grown, although there appears
to be some flexibility with investors testing soils to choose which crop is most appropriate in
some cases. Analysis of the database of the 31 federal land bank projects suggests that
investments fit closely with the priorities of the government as set out in the GTP to encourage
strategic industrial crops. 40% of all land allocated from the Federal Land Bank is for cotton
cultivation, or cotton in combination with other crops. This amounts to 68% of all leases. Edible
oils, excluding palm oil are 16% of the land allocated (see chart 10).

The policy now appears to be to discourage production of food crops on land allocated from the
federal land bank (the large Saudi Star farm for very high-grade rice production seems to be
the big exception). Food crops are only grown on 30% of the land allocated from the land bank.
Leases for cereals alone only account for 9% of the land area41. Saudi Star also claim that they
plan to farm on 300,000 hectares producing 1 million tonnes of rice per year, although the
additional land to make this possible has not yet been leased42.

6.5 Implementation of land leases

Ethiopian government representatives at both regional and federal level acknowledge that
while considerable amounts of land have been allocated to investors, performance to date in
terms of production, employment, and development of land has been disappointing for the most
part. Ethiopia has attracted foreign investors, but has not attracted the very highest quality
international agricultural land development companies, despite the high quality of land
resources and favourable investment climate. The performance of domestic investors has also
been poor in many cases. The land investment situation led the federal government to declare
a moratorium on further land leases in 2011 while the situation was investigated. A similar
process has happened in Benishangul-Gumuz region.

One of the clearest features of large-scale land investments in Ethiopia as they currently stand
is that only a small percentage of land allocated has been developed. According to the MOA of
2.2 million hectares of land allocated (discussed above) only 17.6% has been developed, a
total area of 372,088 hectares (Eskedar, 2012). Similarly, Dessalegn (2011) cites MOA (then
MOARD) data showing that over 3 million hectares were allocated by the regions between
1996 and 2008 to 8000 applicants, and that only 20% of investors had begun project
41
This includes a breakdown of the 100,000 hectare Karuturi land allocation, based on interview data
42
Interview with company technical expert

32
implementation43. Background research for the World Bank Rising Interest in Global Farmland
report found that of 46 cases looked at only 16 were being implemented as intended (Imeru,
2010).

Research for this study suggests that none of the land leases in Somali region had started
work. In SNNPR, of 79 land leases documented at the regional level over 1000 hectares, 43
were active, 21 cancelled and 8 non-functional. In Afar, 9 of 15 projects were farming only 30%
or less of the land leased.

In the course of this research several datasets (from investment agencies and regional land
bureaux) on regional land allocation were examined. Using these sources it is difficult to be
precise about how much land is being developed. Generally tables indicate whether an
investment is operational (or in implementation stage), pre-implementation, and non-
implemented/non-functional or cancelled. Some regions have data by investor on how much
land is being farmed, but these are the exception. In some cases an investment is marked as
operational, but interviews with a company revealed that only a small percentage of land was
actually being cultivated.

Interviews suggest that regional governments are developing better systems for recording how
much land is being developed, but that this is still a considerable challenge given the vast
areas that need to be covered, poor infrastructure and lack of staff and vehicles to carry out
assessments.

In Gambella, 32 licences for land investors have been cancelled following a federal and
regional monitoring review. In SNNPR, 21 projects have been cancelled. In Benishangul-
Gumuz, a recent review resulted in 11 projects being cancelled due to non-development,
breaking the terms of the lease, or because the investor had disappeared. In Afar, it has not
been easy to cancel agreements due to lack of development, particularly where agreements
are made between investors and clans. New regulations and improved capacity in the regional
government are helping to address this.

There are several reasons why land is not developed as expected. These include: high costs
and difficulty of developing land, security issues, poor capacity of investors, misuse of
investment licenses and limited monitoring and evaluation capacity.

A common complaint of investors is that the remoteness of investment locations, poor

infrastructure such as roads, power, telecommunications, and lack of skilled labour, and
services and accommodation and transport for workers, make it difficult and costly to begin to
develop land. The absence of good infrastructure likewise means that it is expensive to market
produce. This is a major disincentive for many investors. In some cases government has
promised that infrastructure would be in place (for example a bridge over the Omo River), but
has not yet completed the task, delaying development by investors.

Some investors also underestimate the costs of developing land, particularly in regions like
Gambella where there is a need to clear forest, and level and prepare land. Saudi Star, for
example, suggest that land preparation costs for its 10,000 hectare irrigated rice farm in
Gambella are around US$ 17,000 per hectare. Representatives of Karuturi put the cost at
around US$ 3000 per hectare. Even if capital is available land preparation also takes significant

43
Note that one third of these 8000 projects were under 100 hectares.

33
amounts of time. Saudi Star for example has only cleared 6500 hectares of 10000 hectares
leased in 2008, with only 350 hectares cultivated.

Unpredictable environments mean that returns on investments are not guaranteed. In 2011,
Karuturi, for example, lost an expected 60,000 tonne first maize harvest on 12,000 hectares of
land due to flooding (Davison, 2011b).

Security is another contributing factor to poor land investment performance. In Somali, security
concerns are cited as the major reason for lack of investment, despite availability of land and
water resources. In Gambella, security is a major consideration for investors, particularly
following the attack on the Saudi Star farm in April 2012 in which 5 employees of Ghulam
Rasool and co, a Pakistani engineering company died and also a bus attack outside Gambella
city in which 19 people were killed (Kirubel, 2012). Security imposes costs in terms of needing
to guard equipment and protect staff, it also makes it difficult to attract and retain workers. At
one farm interviewed for this research, 18 military personnel are deployed permanently to
protect those working on the farm investment, and the farm is not even engaged in commercial
production.

Poor infrastructure and the high costs of land development are particularly significant in
developing regions. This has led the Federal government to explore options for state financing
of development of land resources and relevant infrastructure, to speed up the process of land
development and attract higher quality investors. Ato Tefera Derebew, Minister of Agriculture,
commented: ‘The performance of commercial farms is not up to our expectations. To speed up
their performance, we are planning to develop ready-made clusters of agricultural land. Such
an activity will reduce the time required to prepare land and develop infrastructure like roads.
When agricultural investors request a plot of land, we will promptly give it to them’ (Pawlos,
2012).

As part of this process, 67,000 hectares of land have been identified in Benishangul-Gumuz
(with a proposed target of 100,000 hectares) and 100,000 hectares in Gambella. The federal
government proposes to provide infrastructure such as feeder roads and irrigation access
points, and to engage in land development (clearing, levelling and managing the soil).
However, while these plans have been discussed in national newspapers (Mahlet, 2012), there
was no knowledge of them in discussions with bureau heads and other staff in Benishangul-
Gumuz and Gambella.

Another factor is the size of land leases, some informants felt that some allocations are simply
too large to be managed effectively. Even if a company has capital, shortages of labour and
associated services will be a problem. One regional land bureau chief commented that 10 or
20,000 hectares should be the maximum size leased to investors. However, as discussed
below, it is by no means the large investors who always have the least capacity to develop
land, there are many investors with small land leases who have made limited progress 44.

Research in the regions45 shows that it is common for investors to seek licenses for land
investment often with no clear intention to develop the land. Licenses are attractive to investors
because they allow access to capital (following down payment of part of a loan), which is then

44
In the Saudi Star case plans are in place to create 9 semi-autonomous units on the 160,000 hectares
that would be suitable for rice cultivation on a 300,000 hectare lease to encourage competition and
efficiency
45
This was identified during fieldwork in Afar, Benishangul-Gumuz and SNNPR.

34
put to other investment uses (in some cases the investor even disappears). Licenses also
entitle investors to import machinery and vehicles duty free, in some case these are used for
other purposes unrelated to the land investment business46.

In some cases, land is being used, but for purposes other than those specified in the lease, for
example, in Benishangul-Gumuz land has been used for charcoal production (which is illegal),
or has been rented out to others. These type of activities led government in the region to place
a moratorium on further land leases while relevant bureaux assess the status of investments
and verify the exact size of plots.

The website of the Agricultural Investment Support Directorate has published highlights from a
review of investor performance to date. One finding is that:

When it comes to foreign investors, though they showed to have a huge capital before
investment they were tardy in engaging in tangible tasks according to their words.
Limitations, failure to adopt international trends and failure to import the required
medicaments and inputs were conspicuous on their part. A desire to getting land
without collecting sufficient capital was the problem seen among the Diaspora [MOA,
2012b]

While these comments target foreign investors, it is clear that domestic investors are equally
culpable of lacking capital and skills, and being disingenuous when seeking licenses or signing
a land agreement.

One problem is that land at the regional level was given out with very limited scrutiny of
investors. In Gambella, it was suggested that there was a policy of attracting investors at all
cost, with little thought in the first instance about the implications of this. Investors now produce
statements of financial and technical capacity but there are limits as to how far these can be
probed for accuracy47.

As discussed above, after land is allocated capacity to monitor and evaluate is limited. In most
regions it was evident that very little monitoring has been carried. Monitoring at the regional
level seems to be primarily focused on whether investors are using the land and not abusing
their licenses. Monitoring of how well the land is being developed, and wider impacts is limited.
In some cases it is evident that woreda governments are aware of the problems with particular
investors but lack authority to address problems.

46
The report on investor performance by MOA notes: ‘In Gambela and Benishangul States, with a rent
seeking spirit, acquiring land via corrupt practices against the suspension of land transfer was
witnessed.’ (MOA, 2012b)
47
Dessalegn asserts that capital and business plans are not checked for accuracy (2011: 14). This has
probably been the case in many regions, but scrutiny is now more rigorous both at regional and federal
levels.

35
7. OUTCOMES

This section indicates some of the key issues in relation to land investment outcomes. For
many issues it is either too early, or there is not enough baseline data to allow for strong
conclusions about outcomes.

7.1 Revenues from land investment

One argument in favour of land investment is that it has the potential to generate significant
revenue for government from fees and taxes, which can be used to fund national and regional
development activities.

Revenue from rental fees is not high, and these appear to have been set more to attract
investors than to capture revenue from them. Rental fees for land acquisitions are specified in
land lease agreements. The federal government has set out guidelines for land rental fees
based on two criteria, distance from Addis Ababa and whether land is rainfed or irrigated. Over
700 km from Addis Ababa a uniform rate is applied of 111 ETB (US$ 6.1) per hectare for
rainfed land, and 158 ETB (US$ 8.7) per hectare for irrigated land. The rate is based on a
hypothetical rate of 2100 ETB per hectare at 0 km from Addis, decreasing by 4.5 Birr per km
away from Addis Ababa for rainfed land and 4.17 for irrigated land.

The federal government has been concerned that there has been too much diversity in land
rents across regions, and that relatively cheap land rental fees mean that land acquisition is
being misused to take advantage of tax arrangements such as import duty exemptions.
However, regions are not obliged to follow federal guidelines (for example, in Benishangul-
Gumuz rents are in the 30-70 ETB (US$ 1.7–3.9) per hectare range, SNNPR and Amhara also
set their own rates). Gambella follows federal guidelines, and rates have accordingly been
revised, for example, Saudi Star started out paying 30 Birr per hectare in 2008 but now pay 110
Birr. In Somali rates vary between 16 and 65 Birr. In Oromia they are 27-135 Birr depending on
various criteria and will soon be revised. In some cases there are holidays on payment of rental
fees for up to three years. Land rental fees still seem to be very low even after revision by the
federal government, although we have not made a comparison with other rates in Africa.

Foreign investors are offered various incentives to encourage investment, these include
exemption from income tax for a period of between 3 and 5 years, depending on agricultural
value added and proportion of exportable products, and also 100% import duty exemption for
most capital items. Most products are also not subject to export taxes (one exception is semi-
processed hides and skins). Regions have different rules on tax exemption. In Gambella, for
permanent crops such as palm and fruit trees tax exemption is for five years, but for annual
cereal crops it is just one year.

36
Foreign investment regulations specify that investors are required to invest a minimum of US$
100,000, or 60,000 if in partnership with a domestic investor. If profits or dividends are
reinvested or 75% of output is exported then there is no minimum capital requirement.
Proclamation 280/2002 specifies that profits and dividends can be taken from the country in
convertible foreign currency.

Data for 89 leases collected for this research shows a projected total investment US$ 4.724
billion. Across the 89 leases this gives an average of over US$ 53 million per land lease. This
gives an average investment of US$ 10,496 per hectare. Chart 11 below gives an idea of the
range for investment per hectare. 11 of the 93 cases have projected investment of less than
US$ 200 per hectare, 17 have projected investment of over US$ 20,000 per hectare.

Projected Investment for Selected Land Leases (n=89)

17 17
18
Number of leases

16
14 12 12
12
10 8 7
8 6 5
6 3
4 1 1
2
0
00

00
00

0
0

+
00

00

00
00

00
20

50

0
,0
,0
r1

00
0,

0,

0,
1,
0-

0-

5,

00
10
de

0-

0-

-2

-3

-5
10

20

0,
0-

-1
00

00

00
50

00
Un

10
00
00

,0

,0

,0
1,

,0
5,

10

20

30

50

US$ per hectare

Chart 11: Projected investment for selected land leases

Authorised foreign investors may be able to access finance from the Development Bank of
Ethiopia. Investors are expected to submit 30% of the value of the loan, with the remaining
70% then made available by the bank.

In the early years of land investments it appears that fiscal and other flows to the government
may be quite limited, making it difficult to argue that land investment results in increased
revenues for development activities. State sugar investments are a different case as
government directly controls revenue streams from sugar plantations (although it is possible
that they may be privatised in the future). For private investments, returns to government could
become more favourable, as grace periods come to an end, and if land investments are able to
make effective use of land

Some critics have also argued that looking beyond fiscal and other revenues, large-scale land
investment may not always be the most economically beneficial use of land. Behnke and
Kerven (2011) for example, analyse returns to cotton and sugar plantation investments (before
processing) in the Awash Valley and find that they yield less economic benefits than pastoralist
livestock production systems on the same land (they also do not factor in the risk of long-term
resource degradation from irrigated agriculture in these environments). At present there are no
publicly-available economic analysis studies looking at returns to different land use choices,

37
where land investment is one of the options. This is an area where more work could be carried
out, together with discussion of assumptions, risk and trade-offs.

7.2 Employment

Land investment has the potential to create significant amounts of employment on farms,
whether preparing land, planting, weeding, harvesting crops, managing facilities, or providing
security or other services. Jobs may also be created in transport, hotel, restaurant and other
sectors as a result of land investments. Regional officials in Benishangul-Gumuz also saw
employment as one of the major ways in which land investment would contribute to food
security in the region.

According to data from Ethiopian investment agencies, projected direct job creation through
land investment is substantial – for Afar, Somali and SNNPR our data suggest at least
projected 120,000 jobs for regional investments. This figure amounts to less than one job per
hectare. Furthermore, given the currently low overall levels of implementation of investment
contracts actual job creation is much lower than projected figures. Chart 12 shows the total
number of projected job per 1000 hectares for 53 different investments. Projected job creation
for 12 cases is less than 100 jobs per 1000 hectares. 14 cases projected 200-500 jobs, and 11
cases, 500-1000. Only 7 of 53 cases predicted over 1000 jobs per 1000 hectares, ie one job
per hectare.

Protected jobs per 1000 hectares (n=53)

16 14
14 12
Number of leases

12 11
10 9
8
6 4
4 2
2 1
0
0
0 to 100 100 to 200 200 to 500 500 to 1,000- 2,000- 5,000 - Over
1,000 2,000 5,000 10,000 10,000
Number of jobs

Chart 12: Projected jobs per 1000 hectares

The proportion of total jobs that are skilled is small, and most work is casual and temporary
rather than permanent. In developing regions, skilled jobs are most likely to go to those from
outside the region (or outside of the zone in SNNPR).

38
Some of the major farm investments claim to have created large numbers of jobs. Saudi Star
(with a 10,000 hectare lease in Gambella) claim that at present they are employing 2000
people with 500 permanent and 1500 semi-permanent staff. When the 10,000 hectares are
fully operational they expect that 7500 people will be employed (excluding processing). Karuturi
claim that they are currently employing 1000 people, but that this will increase to 25,000 when
their 100,000 hectare concession is fully operational.

Sugar plantations and factories in SNNPR are projected to create 118,000 jobs (Davison,
2012b), although we have not been able to find a breakdown of types of employment for this
high figure. Concerns have been raised however about the cultural impacts of swamping of
local populations by migrant labourers.

In developing regions concerns have been expressed that most jobs go to people coming from
outside the region. In Benishangul-Gumuz informants suggested employment, particularly
harvesting work, often goes to migrants from Amhara or Oromia and not local ethnic groups. In
one case according to an official in Benishangul-Gumuz local people even destroyed the crops
on an investment project because they were angry about the lack of employment opportunities.
In some cases local people are more likely to get work which is ongoing such as weeding work,
and labour is brought in for harvesting periods. Officials in Benishangul-Gumuz were very
aware that local regulations suggest that jobs should go to people from the region unless there
is no demand or availability of labour. In Afar, interviews suggested that local people were
rarely employed on farm projects.

Field research with SEKA farm in Bench Maji, SNNPR revealed that skilled jobs went to ethnic
groups from outside the region, but people from the local ethnic group were involved in manual
labour on the farm. At the Omosheleko farm investment in South Omo, SNNPR, positive
discrimination policies have been employed to encourage recruitment of people from local
ethnic groups, even when their educational qualifications were below the normal required
standard for a position.

In Gambella, there seem to have been active efforts to create employment for Anuak and Nu’er
people in their respective areas, and companies interviewed for this research claim that they
have training programmes to create opportunities in areas such as tractor driving or pump
operation. It was not possible to verify this on the ground however. It should also be noted that
in some cases taking a job on a farm may feel very alien for members of ethnic groups in areas
where land investment is situated, particularly if they do not have a tradition of settled
cultivation.

In some cases wages on farms are seen as too low. In a case study in Somali region local
people felt wages were unreasonable and even tried to block outside workers from coming to
the farm site. We lack good baselines and data however to assess whether wages are really
low compared to national or regional standards, and how they might compare with incomes
from other livelihoods making use of land before land investment.

People are also sometimes unwilling to take up jobs because of difficulties travelling to the farm
location, or because of lack of services such as clinics in these areas. Communities interviewed
at an investment site in SNNPR also complained of lack of temporary insurance to cover the
costs of accidents. Concern about absence of contracts was also expressed in interviews.

39
In Benishangul-Gumuz, officials noted that there were many cases of Development Agents
(village-level government agricultural extension agents) leaving government service to work on
investment projects. This was not viewed as a problem however as it was seen as a possible
route to transfer knowledge from investments. With high levels of unemployment among those
with technical skills, the creation of new DA positions as a result of these job transfers was
viewed as a positive outcome.

7.3 Food security

One issue raised in relation to land investment is how large-scale commercial farming will
impact on food security. The Ethiopian government promote land investment as a strategy to
improve food security at the national level through foreign exchange earnings generated by
farm outputs, by increased production of crops in the country, and by improved incomes
through jobs created on farm. However, the core of food security strategy is to rely on
intensification of smallholder farming in the highlands. In this respect this paper slightly
disagrees with the interpretation of Lavers (2012) that commercial land investment amounts to
a fundamental and highly-risky shift to a trade based food security strategy. The basic policy is
that that smallholder intensification will deliver food security and not trade based on outputs of
the large-scale sector. The feasibility of this is, of course, another matter.

At the regional level officials cited the possibility of technology transfer to local farmers, either
through inputs that are made available as a by-product of farm investments, or through transfer
of agronomic expertise. There is some evidence of this happening in SNNPR, although it may
well be the exception rather than the rule. Field research in Bench Maji zone, SNNPR, revealed
that the SEKA farm had distributed 15,000 mango and coffee seedlings to local households,
and had built a 23 kilometre road that significantly improved market access for local
households. 150 farms were trained in agronomic techniques and 50 households received
supplies of fertiliser. In Gambella, Karuturi representatives say that they have used company
tractors to plough lands of local smallholders. In Gamo Gofa and South Omo zones, local
officials claim that 10,000 hectares of irrigated land have been made available to smallholders
as a consequence of investment in irrigation for large-scale commercial farming. A statement
by Ethiopian Sugar Corporation, claims that 2250 households resettled for sugar development
were given 1400 hectares of irrigable land and a grain mill (Davison, 2012b).

The Omosheleko farm in South Omo Zone, SNNPR appears to have had positive impacts on
local food security. Benna, Arbora, Tsemay and Braile communities in the have been highly
food insecure, regularly facing drought and hunger. They need to travel substantial distances to
Jinka and Konso to buy cereals such as maize and sorghum. Now thanks to irrigation and other
resources provided by the farm they are self-sufficient in cereals and are producing a
marketable surplus. According to a local Tsemay elder interviewed for this research:
‘Omosheleko is benefiting us a lot. It gives us canals to water our crops. We have been
suffering by hunger, but now we are producing many quintals of maize and we also started
selling it to the market in order to get money.’48

48
Interview for case study of Omosheleko Farm, South Omo Zone. August, 2012.

40
In Afar, 40 per cent of the income from land investments where contracts are agreed with clans
(which has been the case for all private deals) goes to clans in the relevant area, according to
informants for this work. This income may only go to clan leaders, and not benefit all members
of the clan or their food security status.

These examples are only used for illustrative purposes, more intensive research with a wider
number of farms and communities would be needed to confirm how representative they are.

Critics of land investment argue that prime farmland is used to produce crops for export,
making the country dependant on export markets, with less food available locally. Indeed, as
discussed, foreign investment regulations encourage exports with 5 year tax holidays available
where 50% or more of output is exported. Only 2 years are granted if the export percentage is
lower than 50%. Likewise there is no minimum capital requirement if 75% of production is
exported. Profits can also repatriated. As noted above, in the short run these regulations mean
there may be limited tax benefits to government but they are likely to improve as grace periods
come to an end, although this is dependent on farms becoming fully operational.

Good data on the final destination of food produced on farms is very limited. For example, the
Saudi Star farm in Gambella is criticised by some for producing rice for export for Middle
Eastern markets, but company informants in interviews suggest that a significant proportion will
be marketed in Ethiopia, reducing imports of high-grade basmati rice. Rice is not however a
staple crop for the Ethiopian poor.

As noted above the argument about prime farmland is not so applicable for much of the
Ethiopian case as many lands that are being given over to large-scale farms were not
previously used for farming on any significant scale. There may be issues of greater
vulnerability to food insecurity at the local level in some cases where communities no longer
have access to resources that they used as part of their strategies for coping in difficult periods,
and for pastoralists land investment may result in further erosion of the viability of rangelands
and food security of these communities over the longer term. Careful baselines and monitoring
are needed to assess where these risks are most acute.

Large-scale investment needs to be seen in the context of all policies for the agricultural sector.
As noted, the GTP aims to dramatically increase productivity in the Ethiopian Highlands and
the main grain producing areas, through improved varieties, better access to markets,
improved input supply and micro-irrigation (FRDR, 2010a and b). It this part of the Ethiopian
agricultural sector that is viewed as the engine to turn Ethiopia into a net food exporter, rather
than the large-scale sector.

It should also be noted that a significant amount of the land allocated for large-scale production
is for industrial crops. Critics would argue that these do not contribute to food security per se,
only to government revenues. Again the argument is that for the most part these are grown in
areas of new farmland, and the main food security linkages are through employment creation
and improved incomes (including in expanded processing industries), although there will
undoubtedly be pockets of increased vulnerability where communities lose access to
resources, especially if alternative livelihood options or natural resources are not made
available.

7.4 Resettlement and loss of land

41
In addition to employment and food security there are other social impacts of land investment
that need to be considered. These include resettlement, loss of access to resources, and
provision of new social services as a consequence of land investment projects. Given the
limited scope of this project it is difficult to comment on these matters beyond indicating what
the key themes are. Interviews with key informants and a very small number of case studies
suggest that generalisations are risky and that there are a range of both positive and negative
experiences in different places. Again this is an area where more baseline data collection and
monitoring is needed.

In terms of resettlement explicitly linked to land investment, there are different views. Some
informants in Benishangul-Gumuz argued that there was no resettlement, others said that there
was, and that compensation was paid as set out in Ethiopian law. In general it was felt that
resettlement was fairly limited given the very low population densities in the region. The same
is possibly the case in Gambella, particularly in areas where there are less commercial farms.
For large-scale sugar plantations in SNNPR and Afar there has been more criticism of
resettlement. Resettlement is voluntary according to the government, although this is contested
by rights-based groups such as Human Rights Watch (HRW, 2012b). Research for this project
suggested that for the sugar development in Afar compensation of 2500 Birr (US$ 140) per
hectare had been provided, with 60 million Birr (US$ 3.3 million) paid out already.

Consultation about land investment seems to have happened most clearly in Afar, where deals
were with the clans. In other areas such as SNNPR consultation has been very limited.

Another criticism of resettlement is that in some locations, government has been slow to
provide promised services (clinics, schools, water and so on), or has inadequately considered
livelihood options. In addition, some groups simply do not want to move due to attachment to
ancestral lands, and promises of jobs or development are not necessarily persuasive in these
cases. Again, it is not possible to comment further on this here, and this report is only able to
point out that these issues are contested.

In relation to villagisation, the Ethiopian government’s scheme to move scattered communities

into centralised villages in order to facilitate better provision of services (particularly in
Gambella, Benishangul-Gumuz and Somali regions)49, there are again assertions by
international rights-based groups that villagisation is an explicit policy to make land available for
investors (HRW, 2012a). The balance of opinion from informants for this work is that there is no
direct link between villagisation and land investment, as pressure on land resources in these
regions is not so acute as to require this kind of intervention. It is possible that land vacated as
a result of villagisation could be made available for land investment, but this research team did
not see specific evidence of that.

A wider issue is that it is sometimes asserted that large-scale investments only take place on
lands that are vacant or idle. For example for the Kuraz sugar development in SNNPR, the
Ethiopian Sugar Corporation asserts on its website that: ‘At Kuraz Sugar Development Project
where no one will be displaced due to the project, farmers and ranchers are made beneficiaries
of irrigated water and potable water.’(ESC, 2012)50. Officials in the Ministry of Agriculture

49
43,000 people have been resettled in Benishangul-Gumuz
50
This was repeated in an interview with Ato Yilma Tibebu, a representative of Ethiopian Sugar
Corporation: ‘There is no one to be relocated at all, let alone forced relocation, due to the sugar
development project’ (Davison, 2012b). It is evident that people are being resettled in areas around the

42
commented that all lands allocated from the land bank are uncultivated and are not part of local
livelihood systems51. It is beyond the scope of this research to comment on this in detail, save
to say that it is contested, and it appears that in many cases as with sugar in SNNPR,
resources that are being made over to investors are part of livelihood systems, either as
pastoralist grazing lands, or forest resources that are used for a variety of activities (timber, wild
foods and other non-timber forest products), including as part of coping systems (this point was
made in relation to lands in Benishangul-Gumuz that were used by Bertha and Gumuz people
to reduce vulnerability). Part of the problem is that mapping of traditional rights and certification
of group rights in Ethiopia to date has not been adequate. This is gradually being addressed as
part of a new round of land certification programmes (for example, in Benishangul-Gumuz
government will identify certified areas for shifting cultivators, a form of rights protection they
have not had in the past). Also the initial satellite mapping of the federal land bank took land as
being empty, because it was only possible to identify lands that were clearly cultivated. These
limitations have been acknowledged and successive rounds of mappings, in consultation with
regional and woreda governments, are specifying in much greater detail, using GIS systems,
where villages, forests and grazing lands are, and taking these into account in demarcation of
land boundaries.

It should be also noted however that large land allocations do not necessarily have the largest
social impacts, in densely populated areas such as Oromia and Amhara where land is already
intensely cultivated a relatively small deal for 500 hectares or even 100 hectares could
conceivably have impacts on more people than a significantly larger deal in a less-populated
region.

Case studies in SNNPR suggested that there could be considerable conflict with communities
(with alleged loss of life in both cases) when these traditional rights were not respected, but
that some measure of consensus and acceptance of the investment could be achieved if
certain key issues were handled sensitively. Sometimes particular grievances that could have
been avoided become flashpoints for conflict. For example, if investors provide social benefits
such as roads and clinics, and inputs for farming systems, and if they allow continued access to
water points, and design corridors for passage of livestock to grazing lands. Also, if patches of
forests remain with clear rights of access for communities, then some concern about insecurity
will be addressed. Government has a role to play here supporting careful dialogue and
understanding between investors and communities. Large-scale investors are often felt to be
very remote, as they may be based in Addis Ababa or overseas, decision-makers only visit
periodically, and often only technicians are available on the farm for long periods – this can limit
opportunities for dialogue and better mutual understanding.

Corporate social responsibility type activities are an important way for investments to have
benefits at the local level. While research was not able to collect data on this issue in great
detail, informants suggested that good practice by investors was the exception rather than the
norm. Usually actions like provision of schools or clinics are not specified in contracts, and are
certainly not checked as part of the monitoring system. However, in some regions where
government is moving to a land bank system with more systematic assessment of investors,
they could become one criterion for judging between different investors.

sugar farms, these are presented as part of the government’s resettlement programme to promote
development, rather than a consequence of sugar development per se, hence the government argues that
sugar land investment is not technically causing resettlement.
51
Interview with MOA, AISD, technical expert, July 2012.

43
7.5 Land use and environmental impacts

Land use planning is important for ensuring that land investment is considered in the overall
context of how to make best use of different resources to achieve a range of social, economic,
environmental and other outcomes. It inevitably involves trade-offs between objectives.

In Ethiopia, land use planning is weakest in developing regions. Development has happened
without clear strategic analysis of optimal allocations of land in the regions for industry, forestry,
crop agriculture, fisheries, tourism and wildlife. Early iterations of the federal land bank for
these regions also ignored rival land uses, for example the land bank areas in Benishangul-
Gumuz covered important bamboo forests. In Oromia, in contrast there has been much clearer
demarcation of land use zones, and which areas are for specific crops, with guidance provided
to investors accordingly.

Part of the explanation for this is that there is a lack of good information on the status of natural
resources and livelihoods in these areas, on vegetation, topography and other relevant
subjects. Benishangul-Gumuz and Gambella in particular are now developing land use
planning frameworks and guidelines (a framework exists for Benishangul-Gumuz, with
guidelines under development). In Gambella there has also been a remapping of the National
Park, which was never previously gazetted. Benishangul-Gumuz does not have a national park,
but discussions are underway to create one. While these developments are important, they
come after major areas of land have already been allocated for large-scale farming. In the case
of the Karuturi land investment, adjustments to the land allocation have recently been made,
with land reallocated to the Gambella National Park which was formerly in the Karuturi area.

As noted above, land was to some extent allocated in these regions at a time when they were
paid little attention in policy or media discourse. Now the situation has changed somewhat and
developing regions are seen as areas with considerable potential (wildlife tourism in Gambella,
or hydropower development in Benishangul-Gumuz), and policy-makers are attempting to
create better systems for categorising land resources and making choices about how they
should be used.

Environmental impact assessments of land investments to date have been weak or non-
existent. In Benishangul-Gumuz region government officials were frank that land had been
given out without any kind of environmental impact assessment (EIA). Only now are investors
being forced to develop environmental management plans as part of on-going monitoring of
investments. Research in Afar and Somali regions suggested that no EIA had ever been
conducted.

New investors at the federal level need to deliver an EIA within three months of signing a lease.
It is also not possible to access funds from the Development Bank of Ethiopia without an EIA. It
is generally unclear how comprehensive these EIAs are, or how well they are checked, as this
information is not in the public domain. EIAs are also reviewed by MOA at the federal level
rather than the Environmental Protection Authority, the agency which should have more
expertise in this area.

An example of EIA in practice is given in Boxes 3 and 4. Box 3 contains details of possible risks
identified for the Saudi Star land investment in Gambella. These are comprehensive, but with

44
some exceptions rather generic environmental risks that might apply to an investment in many
locations. Box 4 identifies measures taken by the company to mitigate risk.

a. Adverse effect in the life support function (biodiversity of fauna and flora)
of the complex natural ecosystem in the project sites and surrounding
areas
b. Deforestation, loss of forest and grass cover and plant biodiversity
including unique and/or endemic forest tree species
c. Loss in populations and species diversity of high profile and endemic
wildlife resources including large mammals in the project sites and the
surroundings due to changes in habitat, proliferation of poaching and
disturbance.
d. Causing or aggravating climate change and desertification following
continued or sustained reduction in the biological productivity of the land.
e. Land use change induced overall soil and/or land degradation
f. Intensive mechanized cultivation-induced soil physical, and eventually
soil biological and chemical degradation leading to reduction in soil
health and land productivity.
g. Continuous and intensive cultivation-induced mining of non-fertiliser
plant nutrients in soils leading to soil infertility
h. Possibility of secondary salinisation/sodification and alkalization of soils,
and even acidic fertilisers-induced soil acidification.
i. Depletion of water resources in the areas with time and increased flood
hazards
j. Excessive accumulation of farm applied agricultural chemicals caused
pollution and/or toxicity leading to damage of soil health and
contamination of surface and groundwater resources (environmental
pollution) due to safe disposal of farm wastes, and effluents and
byproducts of the rice processing plant
k. Land use conflicts related to protected forest, wildlife reserve, cultivation,
grazing and/or browsing and livestock watering grounds
l. Risk on employees health and safety
m. Increased/aggravated prevelance and outbreak of vector/infectious
human diseases notably malaria and typhoid, and favoured growth of
insect population including agricultural vermin (insects and rodents) in
the irrigated fields
n. Possibility of more burden on women over men in the production
(planting or transplanting, weeding, bird control, harvesting and drying),
processing and trading of rice where women play a major role in all
aspects of rice value chain than men

Box 3: Environmental impacts identified for Saudi Star rice farm

45
 Maintaining trees on farmlands (while clearing the natural vegetation for land
development to retain an acceptable density of trees on the site about 15-20% for
increasing and sustaining ecological functions and yield of crops grown both within
the farmlands and as borders of the farm.
 Reforestation and/or reafforestation of activities (the natural vegetation cover lost
during land clearing and development for rice production particularly on the
command areas shall be compensated to a certain extent through afforestation
and/or reforestation program of on 10 per cent of the project land
 Wild animal population and diversity (recommended alternative development to
supplement in the proposed areas with cattle breeding, wild animal farming, forest
development, exotic wildlife reserves, hunting and even development of adaptable
perennial fruit tree plantations in some areas, and
 Soil erosion control (in order to combat soil erosion and soil physical degradation
through soil structure deterioration and compaction accelerating as a result of
clearing of the natural vegetation with bulldozers followed by intensive mechanized
cultivation, use of integrated biological, mechanical and cultural) soil and water
conservation practices will be adopted.

Box 4: Environmental protection measures taken by Saudi Star

Officials interviewed in Benishangul-Gumuz were very concerned at the effects of land

investment on natural resources, particularly loss of high-value forests, and the danger that
cultivation in areas with shallow soils would result in land becoming exhausted after a few
years. Trees are generally cut and burnt during land clearance resulting in release of carbon to
the atmosphere as well as a loss of ecosystem services. Regulations now specify that investors
must leave 60-70 indigenous trees per hectare, and 100 on slopes. In addition cultivation within
60 metres of a slope is prohibited52.

In some cases investors argued that they were transforming the soil in a positive way, for
example, a company growing cotton in SNNPR, claimed that they had planted leguminous
crops to build up sub-soil, allowing cultivation where none had been possible before. The wider
sustainability of water use for this cotton cultivation was not discussed. In Afar, there is concern
about the sustainability of use of the Awash river for large-scale cotton production. In general
there is no policy on charging for water for large-scale investments, but there is a difference in
federal land rental price guidelines with a higher price specified for irrigated land.

As noted elsewhere better baselines are needed on environmental conditions and land use
when land investments get underway. This would make it more possible to rigorously assess
environmental impacts over the medium term.

52
Officials also noted that studies of tree resources linked to the federal land bank were based on
checklists of trees that were irrelevant to Benishangul-Gumuz, none of the key species in the region
were listed.

46
8. Conclusions

Ethiopia is an important test case for global agricultural land investment in a poor developing
country. As the inventory for this research has presented, with over 1 million hectares allocated
to large-scale commercial agriculture in the last 8 years, the granting of land concessions has
been both a rapid and sizeable process. It is likely to continue to grow in the years ahead.

Ethiopia also challenges assumptions about global land investment, namely that it is primarily
an agenda driven by global corporations and governments of the rich world, with developing
country governments reluctantly pressured to accept liberal investment policies as part of wider
development conditionalities. In Ethiopia this is resolutely not the case and land investment has
been primarily driven by the EPRDF government, which sees itself as managing a
developmental state, governing the market to encourage investment to meet carefully
determined development goals and objectives.

Africa is seen in analysis of international food security and agricultural development as a

continent with unfulfilled agricultural potential and some of the most underutilised land in the
world. The Ethiopian government largely subscribes to this view in relation to its own country,
and identifies large areas of land in the country as virtually unused and of high potential for
agricultural production. This has driven the creation of a federal land bank and promotion of
Ethiopia as an investment destination for international agribusiness.

This process of investment in the lowlands can be seen as part of an ongoing project of state-
building in hitherto peripheral parts of Ethiopian territory. This has entailed some transfer of
power over land allocation from regional to federal governments, possibly weakening regional
autonomy.

Pursuit of national development goals has social and environmental impacts at the local level,
as forests are cleared and in some cases communities are resettled, or lose access to
rangelands, forest or water resources. This research has only been able to touch on these
issues, and this is an area where more publicly-shared data collection by government is
needed, as well as independent research. There are undoubtedly cases where problems at the
local level could be avoided by more careful choice of land allocation, and by encouraging
companies to allow better access to resources where possible, to create training opportunities
for local populations, and to provide social benefits such as schools and clinics.

Large-scale land investment is one part of agricultural strategy. The Ethiopian government also
places great emphasis on intensifying production in smallholder areas. Greater efforts could be
made to build linkages between the two sectors, for example, in terms of joint marketing,
access to inputs and sharing of technologies.

The long term impact of land investment on pastoralist rangelands, and the economic returns to
different land uses also needs more analysis.

On the right terms investment can be welcomed by communities if carefully crafted. Likewise
there are possibilities for spillovers to local smallholder farmers through sharing of
technologies, skills, provision of low-cost inputs and help with marketing, through contract
farming and other arrangements, although much more evidence is needed on this. There
should be incentives for investors to pursue these kind of activities and they should be
monitored as they implement them.

47
Land investment in Ethiopia proceeded initially in a chaotic fashion. This is especially the case
in some regions where land has been given out without proper scrutiny of investors,
environmental impact assessments or monitoring of performance. Even with federal land
agreements progress on implementation has been slow. Government at both regional and
federal levels is now working to address some of these difficulties with temporary moratoriums
on further land deals at the federal and some regional levels, and a plan to put in infrastructure
in remote regions to speed up implementation of agricultural production plans and attract
higher-quality investors.

New land bureaux in many regions are seeking to assess how much land really has been
allocated, how much is being farmed and cancelling inappropriate leases. New land
agreements will involve more careful assessment of investors both in terms of environmental
impacts and capacity to farm, with allocation of pre-identified land. New donor-supported land
certification processes could also help protect the rights of local land-users alongside the
interests of investors. However, capacity to scrutinise and monitor investors in the developing
regional states in particular is still very limited, especially where investments are spread over
very large areas.

Ethiopia has set itself ambitious development goals with agricultural transformation at the
centre of its strategies. Large-scale land investment that is productive rather than speculative
has a role to play, although careful consideration of social and environmental impacts,
alongside economic analysis of costs and benefits of different land use options, based on
quality information shared between different stakeholders, is needed to ensure this.

48
References

Behnke, R. and C. Kerven, 2011. Replacing pastoralism with irrigated agriculture in the Awash
Valley, north eastern Ethiopia: counting the costs. Paper presented at the International
Conference on the Future of Pastoralism, Addis Ababa, 21 st - 23rd March 2011.

Butler, E. 2010. Land grab fears for Ethiopian rural communities. BBC News. 16 th December
2010. Available at: http://www.bbc.co.uk/news/business-11991926

Eskedar Kifle, 2012. Commercial farms’ performance dire. Capital. Sunday May 13 th.

Cotula, L., Vermeulen, S., Leonard, R. and J. Keeley, 2009. Land grab or development
opportunity? Agricultural investment and international land deals in Africa. London:
International Institute for Environment and Development (IIED) and Food and Agriculture
Organisation (FAO).

Davison, W. 2011a. Ethiopia builds 10 new sugar plants as it aims to become leading exporter.
Bloomberg News. September 13th. Available at: http://www.bloomberg.com/news/2011-09-
13/ethiopia-builds-10-new-sugar-plants-as-it-aims-to-become-leading-exporter.html

Davison, W. 2011b. Karuturi’s First Corn Crop in Ethiopia Destroyed; Will Replant. Bloomberg
News. 6th October.

Davison, W. 2012a. Ethiopia Identifies Million Hectares for Biofuel Crop Investors. Bloomberg
News. 11th April

Davison, W., 2012b. State Sugar Project Harming Ethiopian Tribes, Rights Group Says.
Bloomberg News. 18th June.
Available at: http://www.bloomberg.com/news/2012-06-18/state-sugar-project-harming-
ethiopian-tribes-rights-group-says.htm

Dessalegn Rahmato, 2011. Land to investors: large-scale land transfers in Ethiopia. Addis
Ababa: Forum for Social Studies.

ESC, 2012. Community around new sugar development projects become most beneficiary.
Ethiopian Sugar Corporation website. Posted. October 6th. Available at:
http://www.etsugar.gov.et/corp/index.php/news

Eyasu Elias and Feyera Abdi, 2010. Putting pastoralists on the policy agenda: land alienation in
southern Ethiopia. Gatekeeper No. 145. London: International Institute for Environment and
Development

FDRE, 1994. Constitution of the Federal Democratic Republic of Ethiopia. Federal Democratic
Republic of Ethiopia

FDRE, 2005. Federal land administration and use proclamation. Federal Democratic Republic
of Ethiopia. No. 456/2005

49
FDRE, 2010a. Growth and Transformation Plan 2010/11– 2014/15. Volume I. Main Text.
Federal Democratic Republic of Ethiopia. Ministry of Finance and Economic Development.
Addis Ababa.

FDRE, 2010b. Growth and Transformation Plan 2010/11– 2014/15. Volume II. Policy Matrix.
Federal Democratic Republic of Ethiopia. Ministry of Finance and Economic Development.
Addis Ababa.

Flintan, F., 2011. ‘Broken lands, broken lives?’ Causes, processes and impacts of land
fragmentation in the rangelands of Ethiopia, Kenya and Uganda. Report for the Regional
Learning and Advocacy Programme for Vulnerable Dryland Communities (REGLAP), Nairobi.
June.

Hammond, L. 2008. Strategies of Invisibilization: how Ethiopia’s resettlement programme hides

the poorest of the poor. Journal of Refugee Studies. 21.4. pp 517-536

Hammond, L. 2011. Governmentality in motion: twenty-five years of Ethiopia’s experience of

famine and mobilisation policy. Mobilities. 6.3. pp 415-432

HRW, 2012a. “Waiting here for death”: Displacement and “Villagization” in Ethiopia’s Gambella
region. Report by Human Rights Watch. New York: Human Rights Watch

HRW, 2012b. “What will happen if hunger comes?”: Abuses against the indigenous peoples of
Ethiopia’s Lower Omo Valley. Report by Human Rights Watch. New York: Human Rights
Watch.

Imeru Tamrat, 2010. Governance of large-scale agricultural investments in Africa: the case of
Ethiopia. Paper presented at the World Bank conference on land policy and administration.
Washington, D.C., April

Kirubel Tadesse, 2012. Unknown gunmen attack Al-Amoudi’s commercial farm, 5 dead.
Capital. Sunday May 6th.

Lavers, T. 2012. ‘Land grab’ as development strategy? The political economy of agricultural
investment in Ethiopia. Journal of Peasant Studies. 39.1. pp. 105-132.

Mahlet Mesfin, 2012. US experts join agro ministry for zone study. Fortune. 12: 625. April 22nd.

MAI, 2012. Agricultural Operations in Ethiopia. Morrell Agro-Industries Corporate Presentation.

Available at: http://morrellagro.com/wp-content/uploads/2012/05/MAI-Corporate-
Presentation.pdf

Makki, F. 2012. Power and property: commercialization, enclosures and the transformation of
agrarian relations in Ethiopia. Journal of Peasant Studies. 39.1. pp 89-104

Markakis, J. 2011. Ethiopia: the last two frontiers. London: James Currey

MFA, 2012. ‘The Oakland Institute’s campaign to perpetuate poverty’. A Week in the Horn.
Ministry of Foreign Affairs press release. 9th March 2012.

50
MFAF, 2012. Socio-economic baseline study and assessment of the impact of villagisation.
Ministry for Foreign Affairs of Finland, Unit for Eastern and Western Africa. Responsible and
Innovative Land Administration in Ethiopia programme. Final Report. May, 2012.

MOA, 2012a. Policies and strategies. Ministry of Agriculture website. Available at:
http://www.moa.gov.et/node/65

MOA, 2012b. A performance assessment report on agricultural investors. Ministry of

Agriculture website. Available at: http://www.moa.gov.et/node/212

MOARD, 2009a. Agricultural investment potential of Ethiopia. Ministry of Agriculture and Rural
Development. March

MOARD, 2009b. The Ethiopian Agricultural Investment Areas. Ministry of Agriculture and Rural
Development. Agricultural Investment Support Directorate. Public Relation Bureau. November.

Mosley, J. 2012. Peace, bread and land: agricultural investments in Ethiopia and the Sudans.
Chatham House Briefing Paper. January.

Muluken Yewondwossen, 2012. Rubber plantation closer to fruition. Capital. 14: 697. Sunday
April 22nd.

OI, 2011. Understanding Land Investment Deals in Africa. Country report: Ethiopia. Report by
The Oakland Institute. Oakland, CA.: The Oakland Institute. Available at:

Pankhurst, A. and F. Piguet. 2004. People, space and the state: migration, resettlement and
displacement in Ethiopia. Proceedings of the workshop held by the Ethiopian Society of
Sociologists, Social Workers and Anthropologists and the United Nations Emergencies Unit for
Ethiopia. 28-30th April, 2003. Addis Ababa: Addis Ababa University

Pawlos Belete, 2012. Making agricultural investment more appealing. Capital. June 18 th.

Pearce, F. 2012. Land Grabbers: Africa’s hidden revolution. Sunday 20th May. The Observer.
Available at: http://www.guardian.co.uk/world/2012/may/20/land-grab-ethiopia-saudi-
agribusiness?INTCMP=SRCH

SOS Sahel, undated. Pastoralism in Ethiopia: its total economic value and development
challenges

World Bank, 2010. Rising Interest in Global Farmland: can it yield sustainable and equitable
benefits? Washington D.C.: World Bank.

51