Current land policy issues in Ethiopia - B. Nega, B. Adenew and S. Gebre Sellasie

Ethiopian Economic Policy Research Institute, Addis Ababa, Ethiopia

We have used an extensive data set from Ethiopia to study the country's existing land tenure system from the perspective of farmers and to assess the impacts that this system currently has on agricultural development and implications for poverty reduction at large. The data show that access to land has considerable influence on household income and food security, to the extent that small farm sizes appear to be one of the major constraints for farm development and intensification efforts. Farmers placed more emphasis on tenure security rather than on type of ownership per se. A study of farmers' perception shows that they feel largely insecure on their current holdings. Ways must be sought to provide better tenure security of land holdings. Given the growing population pressure, increasing number of landless households and limited access to land, encouraging the development of the nonagricultural employment sector is crucial.

BACKGROUND

The land-holding system in most developing countries is not simply an economic affair. It is very much intertwined with the people's culture and identity. This partly explains why land-related issues usually generate intense emotional reactions. Obviously, for rural residents of most developing countries, land is the primary means of production used to generate a livelihood for a family. It is also the main asset that farmers have to accumulate wealth and, equally importantly, is what they can transfer in the form of wealth to future generations. Accordingly, the size of the land that they own and the level of security they have in their holdings affect a household's income, and their incentive to work and to invest. Accordingly, land policy in developing countries is a crucial, albeit sensitive, part of the overall development policy that governments need to consider if rapid economic growth and poverty alleviation are to be achieved.

Although there is wide recognition regarding the importance of land policy in agrarian development, there is no clear and universally applicable blueprint as to what an appropriate land policy should be. This is partly because the efficacy of land policy in encouraging agricultural development depends on sociocultural and geographical variables that significantly differ from country to country. In a number of cases, even within one country, various types of landholding system exist depending upon historical factors and the farming systems that have prevailed in the different regions. Despite these differences, however, using established theories, behavioral assumptions regarding economic agents and on experience from other countries, researchers have tried to define certain basic principles and thereby achieve a land policy that will generate higher levels of productivity in agriculture while also maintaining considerations of equity. However, even these principles will change with time as more information is gathered from other countries.

If we take the 1975 World Bank land policy paper as a starting point, three basic principles should be considered in informing any land policy. At that time, the World Bank believed that (a) owneroperated family farms were efficient and thus desirable, (b) there should be freely operating land markets to permit land transfers to more efficient and productive users and (c) there was a need for a more equitable distribution of assets (Deininger and Binswanger, 1999). These principles are still considered to be largely valid. However, based on experience from various countries that have subsequently implemented land reforms, a number of amendments were made to this position including: (a) a recognition, under certain circumstances, that communal tenure could be a cost-effective mechanism for land allocation compared with formal titling; and (b) that formal titling, when desirable, should be evaluated in terms of both its potential efficiency benefits and its implications for equity and the significance of expanded land rental markets on productivity and agrarian development in general. Although the benefits of formal land markets have been recognized, its full potential can be secured only if other factor markets are equally and effectively operational in rural areas. Recent research has also recognized the importance of decentralized administration of land-related issues. What is significant in these findings is that whereas there are certain agreed objectives (such as efficiency and equity) and principles (security of tenure) that should inform land policy, the modalities of a landholding system in a specific country could and should take in to account the specific conditions prevailing in that country in order to achieve the stated objectives of providing a decent living to the farming population.

One development in the last few decades that has been recognized is the degree to which the importance of overtly ideological considerations have been relegated to the background in land policy debates. This should certainly help in allowing a more reasoned discussion about appropriate land tenure arrangements that take into account the specific circumstances of each country; in addition this will reduce any politicization of the issues thereby overly complicating the search for a viable solution. Accordingly, with ideology consigned to the background, many countries are now boldly addressing the issue of land reforms. In the last two decades, most countries in sub-Saharan Africa have been variously undertaking land reform in one way or another. The aims of most of these reforms have been to promote productivity in agriculture and thereby generate rapid economic growth, encourage a more sustainable management and reduce poverty. In relation to reducing rural poverty, access to land is recognized in most of these reforms as of special significance in the absence of alternative employment opportunities in the non-agricultural sectors of the economy (Toulmin and Quan, 2000).

Ethiopia is one of the few countries in Africa that has not made significant changes in its basic land policy since the radical land reform of 1975; exceptions to this have been occasional land redistributions to accommodate the growing population and, in some instances (such as the Amhara region's redistribution of 1996), to redistribute the land that the government feels was unjustly acquired by former government cadres. The reasons behind the lack of government interest in alternative tenure regimes are not that the current policy essentially serves the interest of agricultural development better than its alternatives. Neither is it because there are no critics who argue that the existing land tenure system in the country falls significantly short of satisfying some of the basic principles of land tenure outlined above. In fact, many researchers have raised a number of issues regarding the land policy since 1975 concerning the deleterious effect it has had and the potentially disastrous future consequences to the country's agrarian population if it is not reformed urgently (Dessalegn, 1984).

In terms of the performance of the agricultural sector since the 1975 reform, it is well known that the sector has been performing badly. The entire sector grew at an average of 2 percent per annum between 1980 and the fall of the Derge (Ethiopian Military Junta) regime in 1991. Part of this poor performance is explained by the numerous restrictive regulations imposed by the Derge regime, including price fixing, forced creation of cooperatives, and preferential treatment to cooperatives and state farms at the expense of smallholders. Unfortunately, the performance of the sector did not improve much even after some of these egregious restrictions were lifted following the 1992 reforms of the current government. A thorough analysis and evaluation of Ethiopian agriculture recently undertaken by the Ethiopian Economic Association (EEA, 1999/2000) reveals that, despite the commendable numerous initiatives and measures undertaken by the government, the performance of the agricultural sector remained poor. The growth of the sector, excluding the politically turbulent year of 1991/92, increased only slightly to 2.26 percent per annum in the six years between 1992/93 and 1997/98. In an attempt to explain this lacklustre performance, the EEA report cited above and a number of other studies (e.g. Mulat et al., 1998) identified many elements of the existing land tenure system such as declining farm size, tenure insecurity and subsistence farming practices as important causes for the poor performance of the sector.

It is with this background that the EEA/Ethiopian Economic Policy Research Institute (EEPRI) considered the need for a thorough analysis and evaluation of the rural land tenure system as an important factor in shaping the socio-economic structure of the rural sector. This research involved three steps. First was a survey of the literature on the economics of land policy in general and the land tenure systems under different regimes in the country since the period of colonial rule. The second part of the research, which is more relevant to this paper, involved a large survey of farm households in all regions of the country except for Gambella to solicit a wide variety of information from farmers regarding size of holding, farm and non-farm income, their opinion of the current land tenure arrangement and their preferred alternatives, etc. This was an extensive survey that involved more than 8 500 households in eight regions of the country. A complete analysis of this information is provided in the full report made available by EEA/EEPRI (2002). This paper provides results from this survey towards answering questions raised in the current debate on land policy issues in the country. The third part of the report looks at the opinion of experts in the field (researchers, government operatives, extension agents, etc.).

The paper is organized into five sections. The next briefly outlines those land tenure systems that have been in use since the period of colonial rule, followed by a brief discussion of the impact of these policies on resource allocation and agricultural development in the country based on available literature. The fourth and main part of the paper discusses findings from the rural household survey by looking at pertinent variables related to land policy and rural incomes. This section also discusses the various arguments presented in support of or against the current land policy in light of the survey findings. The last section provides some policy recommendations for the future.

OVERVIEW OF ETHIOPIA'S LAND TENURE SYSTEM^[84]

Land tenure system up to 1974

It is generally acknowledged that the pre-1975 land tenure system in Ethiopia was one of the most complex in the world and had not been thoroughly studied (Cohen and Weintraub, 1975; Gilkes, 1975; Dessalegn, 1984; Dejene, 1999). The country's geographical, ethnic and cultural diversity and its historical background were mentioned among those factors that had produced highly varied forms of land utilization and ownership. The complex nature of the system was also noted as playing a major part in hindering any serious progress towards a meaningful reform. These difficulties might have also resulted in a variety of classifications and approaches used to describe the then land tenure system. Among others, rist/kinship, communal, diessa/village, private, state and church land tenure designations were used. However, the most commonly recognized tenure types are rist/kinship, private, church and state holding systems. The rist/kinship tenure system was most prevalent in the northern part of the country whereas private tenure predominated in much of the rest of the country. Government tenure was most prevalent in the lowland and pastoral areas of the country.

The rist system was characterized by the principle of acknowledging access to land (use rights and transfer rights without land alienation) by all descendants of people from a common ancestor and in an ambilineal way (the right to inherit land from father's and mother's line of descent). Gult is a tenure system that can be considered a variant to rist or rist - gult which is often easily confused with rist. The main distinctions are that gult is not a right on the land, rather a right to tax the benefits from land. Gult is not transferable whereas rist - gult is transferable. Reduced landlessness and tenancy were among the positive attributes of this system over private tenure whereas diminution of holdings, land fragmentation and persistent litigation over land access were among its serious problems cited in the literature. With its gult rights over landholders, the rist system had also been an important element mediating the control of the peasantry by the regime through those elites holding such rights (Yigremew, 2002).

Private tenure was recognized as the most dominant system during the final days of the Imperial regime, affecting some 60 percent of peasants and 65 percent of the country's population. It was largely created by means of land granting by the crown to those members of the army who came from the north and those who were loyal to the regime in captured territories. Under this system land was sold and exchanged; however, given that all the land was originally state property and that private holders had no absolute rights, this was different from the general concept of a freehold system. Serious land concentration, exploitative tenancy and insecurity have characterized the private tenure system.

Land tenure during the Derge period (1974 - 1991)

The 1975 land reform by the Derge has been considered by many as a radical measure that has abolished tenant - landlord relationships in Ethiopia. The reform was designed to alter fundamentally the then agrarian relations and make those working the land the owners; increase agricultural production; create employment; distribute land and increase rural income; and provide a basis for agricultural expansion. The provisions of the proclamation (No. 31/ 1975) include: public ownership of all rural lands; distribution of private land to the tiller; prohibitions on transfer-of-use rights by sale, exchange, succession, mortgage or lease, except upon death and only then to a wife, husband or children of the deceased; and in the case of communal lands, possession rights over the land for those working the land at the time of the reform. The power of administering land was vested in the Ministry of Land Reform and Administration (MLRA) through Peasant Associations at the grassroots level. The law provided 10 ha of land as the maximum a family can possess. No able adult person was allowed to use hired labour to cultivate their holdings (Yigremew, 2002).

Since the 1975 land reform, the right to own land is vested in the state. Farmers access land through state-mandated peasant associations. Farmers have openended usufruct rights (the right to use another's property) to land in peasant associations where they reside, but subject to proof of permanent physical residence, and ability to farm continuously and meet administrative dues and obligations. These use rights are inheritable.

The common practice was to allocate land according to the number of household members (Dessalegn, 1984; various case studies in Dessalegn, 1994; Yigremew, 1997; Haile-Gabriel, 2000). Other factors such as quality of land, size of family workforce and ownership of farm assets, which have substantial influence on ability to use land, are not given as much prominence as family size. Hence, there are farmers who hold equal size of land per household, but with significant variation in factor intensity, such as land per adult labour, land per oxen and land per working capital. As new claimants for land arise, these associations have continuously to meet the new demands for land. However, their ability to meet the growing demand for land, especially their capacity to balance other factors at farm level, has been limited (Tesfaye and Bedassa, 2002).

IMPACT OF PAST LAND POLICIES AND PROGRAMMES

As indicated earlier, the major problems of the pre-1975 land tenure in Ethiopia cited by different studies include exploitative tenancy, land concentration and underutilization, tenure insecurity, and diminution and fragmentation of holdings (Fesseha, 1970: pp. 135 - 140). Tenure insecurity was cited as one of the limitations of the prerevolution/reform land tenure system. It manifested in various forms ranging from endless litigation over land rights to complete eviction from holdings. Endless litigation has also been mentioned as one of the features of the rist tenure (Alem-Ante, 1970: p. 218).

Dessalegn (1984) has argued that for all tenants, the major factor for their dependency and the chief obstacle to improved production was the lack of security of tenure. One major feature of the tenure system was dwindling size of holdings. The majority of peasant households both in the prereform (57 percent) and in the postreform periods (72 percent) operated holdings of 1 ha or less. The total number of this group of "mini-plot" holders increased substantially after the reform. The Ministry of Agriculture (1975: pp. 58 - 59) agricultural sample survey (1974/75) regards fragmentation "as a very serious problem in some regions of Ethiopia". Although sometimes "reluctance" of peasants is given as an excuse for not adopting new technologies, it is more closely associated with resource provisioning than with attitude. Mulat et al. (1998) in their study on the determinants of fertilizer use concluded that the most important factor explaining the quantity used by farmers is farm size.

A survey of the literature on land tenure during the Derge regime (1975 - 1990) generally shows that diminution and fragmentation of holdings, tenure insecurity and all its consequences, land degradation, and inefficient allocation of land by way of restrictions on land transfer and to some extent lack of appropriate land use and administration are the most commonly cited problems. Many case studies illustrate these situations.

Land tenure and agricultural development policies of the government since 1991

Rural land policy and Ethiopia's agricultural development led industrialization policy (ADLI)

The Ethiopian Government formulated a development strategy known as the ADLI launched in 1994/95. ADLI is described as focusing on increasing the productivity of "smallholder farmers" through the diffusion of fertilizers and improved seeds, together with the establishment of credit schemes as well as the expansion of infrastructure - the road system, improvement of primary health care, primary education and water supply. The strategy viewed agriculture as the engine of growth, based on its potentially superior growth linkages, surplus generation, market creation, and provision of raw materials and foreign exchange. The success of ADLI depends on the validity of a number of its assumptions. First is the issue of substantially increasing productivity through the provision of improved technology alone without fundamental change in the existing institutional arrangement. Second is the rather tenuous assumption that increased productivity and thus output will not affect prices adversely and thus lead to increased monetary income of rural households and thereby to a demanddriven industrialization. The two recent experiences of 1996/97 and 2001/02 in which bumper crops have led to a precipitous collapse of agricultural prices are examples questioning the validity of this assumption. Accordingly, critics of the strategy argue that ADLI as an agriculture and overall development strategy is facing complex challenges. One such issue is the unresolved issue of land tenure. Moreover, the assumption that agriculture could take a leading role in economic development without concomitant strategies for urban development requires close scrutiny.

Poverty reduction strategy paper in Ethiopia in relation to land policy

Ethiopia completed the preparation of the poverty reduction strategy paper (PRSP). In the process of developing the PRSP document, consultative meetings were held both by the government and by civil society institutions, and similar results were found regarding poverty in Ethiopia and its possible causes. All the findings from these consultations indicate a significant increase in poverty over the past 5 years and in almost all regions of the country. In terms of increasing rural poverty, one of the most frequently cited reasons was the unavailability of land and the declining fertility of the soil (EEA/ EEPRI, 2002). This is clearly related directly to the existing land policy. Unfortunately, the issue of land policy was not vigorously debated in these consultations as the discussions were deliberately focused on the government's development strategy (ADLI), the Interim Poverty Strategy Paper (IPRSP) and the government's five-year development programme. Although there was no prohibition in discussing issues outside this framework, a discussion on land policy certainly was not encouraged. The government also made it clear that the PRSP document would focus on ADLI as it relates to the agricultural sector. It is therefore unlikely that much will come from the PRSP document by way of land policy reform. However, we show here that rural poverty is very much related to the small size of landholdings, insecurity of tenure, poor farm management practices, etc., which are directly or indirectly related to land policy. In particular, reducing poverty in rural Ethiopia by using the agricultural extension programme alone without addressing problems relating to land policy is unlikely to succeed.

Current land tenure

Immediately after the downfall of the Derge, no one was certain what course the new government would take regarding land tenure. The Transitional Government of Ethiopia had declared that the issue of land tenure (then defined as a choice between private and public ownership) would be settled in the process of developing the new federal constitution. When the new federal constitution was adopted in 1995, the issue was settled in favour of public ownership of land and secured as one of the articles of the constitution requiring not only the full agreement of regional parliaments but also a two-thirds majority in a nationwide referendum. In so doing, the government effectively eliminated land policy as a variable instrument that could be used to address the changing circumstances that affect the rural economy.

Article 40 of the 1995 constitution (which concerns property rights) provides that 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 people of Ethiopia. "Land is a common property of the Nations, Nationalities and peoples of Ethiopia and shall not be subject to sale or other means of exchange" (Sub Article 3). Sub Article 4 also states that "Ethiopian peasants have the right to obtain land without payment and the protection against eviction from their possession." Another important provision regarding property rights (Sub Article 7) states that "Every Ethiopian shall have the full right to the immovable property he builds and to the permanent improvements he brings about on the land by his labor or capital. This right shall include the right to alienate, to bequeath, and, where the right of use expires, to remove his property, transfer his title, or claim compensation for it."

The constitution also states (Article 51) that the Federal Government shall enact laws for the utilization and conservation of land and other natural resources. Article 52 also states that Regional Governments have the duty to administer land and other natural resources according to federal laws. Such law was enacted in July 1997 on "Rural Land Administration Proclamation, No. 89/1997". This law vested Regional Governments with the power of land administration (defined as "the assignment of holding rights and the execution of distribution of holdings"; Article 2, Sub Article 6). Holding rights were also defined as "the right any peasant shall have to use rural land for agricultural purposes as well as to lease and, while the right remains in effect, bequeath it to his family member; and includes the right to acquire property thereon, by his labor or capital, and to sell, exchange and bequeath same" (Article 2. Sub Article 3).

There are no fundamental differences between the legal framework of the Derge and the present government on rural land issues. Moreover, if those belated policy changes^[85] made by the Derge following its 1990 "mixed economy" reform are taken into consideration, more has stayed the same than has changed. In practical terms, there are more similarities in land administration between the two regimes than differences (see Dessalegn, 1997; Ege, 1997; Yigremew, 1997, 1999; Mekonnen, 1999; Hoben, 2000).

However, control of land administration has now been taken away from regional governments^[86] and is now directly under the responsibility of political bodies rather than technical ministries. Land redistribution practices, which were also halted by the Derge in 1989, have been started under the present government. The recent land laws mentioned above also contain provisions for land redistribution. However, in some regions it seems that no further redistributions are anticipated (e.g. in Tigray). For many economic issues, current policies are a continuation of those changes introduced towards the end of the Derge government.

Major problems of the existing land tenure system

A group of specialists from universities (local and abroad) had assessed rural land tenure issues in Ethiopia after the fall of the Derge (Bruce and Shem Migot- Adholla, 1993). The major issues identified by the study include: many people are considered as landless by their community; inheritance, sharecropping, cash rentals, disguised land sales and possessory mortgages (antichresis) have also been important means of gaining access to land.

Other researchers have also concluded that there are problems with the current land tenure system. From his recent studies in Amhara, Oromia and Tigray regions, Tekie (2000) considered that the government had only one imperative policy option: a movement away from the existing insecure tenure system towards a more stable and secured system. In a related work, Hoben (2000: p. 7) concluded:

"there was a consensus that the current system, because it does not guarantee security of tenure and undermines incentives, has detrimental effects on agricultural productivity and natural resource conservation... current land policy does not give farmers secure rights over the land they use, does not maintain equitable access to land over time, does not provide incentives for investment in improvements or conservation, and does not encourage farmers' entrepreneurial and experimental efforts to better their lot. From a policy perspective, it does not foster agricultural intensification, improved environmental management, accretion capital formation, or rural development."

The current debate on land tenure and policy

Despite the constitutional provision that securely vested the ownership of land to the state, rural land policy in Ethiopia has remained one source of disagreement and focus of debate among politicians, academics and other stakeholders. An assessment of the land policy debate in present-day Ethiopia (Yigremew, 2001) shows that there is an unfortunate focus on ownership issues and a dichotomy of views on state versus private ownership. The government and the ruling party advocate state ownership of land whereas experts and scholars in the field, Western economic advisors, international organizations such as the World Bank (World Bank, 1992) and opposition political parties favour private ownership. The main plank of the view advocating state ownership is that private ownership will lead to concentration of land in the hands of few people who have the ability to buy, resulting in the eviction of poor peasants and thus aggravating landlessness and potentially leading to massive rural - urban migration of people left without any alternative means of livelihood. The empirical validity of this claim is one of the issues that the survey addresses.

Critics of the current landholding system and those who advocate freehold largely base their arguments on the behaviour of economic agents and familiar property rights arguments, partially backed up by empirical results from Ethiopia and other countries. Because most of the arguments are variations on the same theme, they can be summarized using the more coherent formulation in Barrows and Roth (1989: p. 4):

• Individualization of land tenure (leasehold and freehold ownership) increases tenure security of the landholder, thereby reducing economic costs of litigation over land disputes.

• Individualization increases investment by increasing tenure security and reducing transaction costs. Higher tenure security increases expected investment returns, thereby increasing the demand for capital (including credit) for fixed-place investment. The supply price of credit decreases because the cost of lending is reduced by improved credit worthiness of projects, and higher collateral value. Both supply and demand effects increase investment.

• Individualization will cause a land market to emerge. Land will be transferred to those who are able to extract a higher value of product from the land as users who are more productive bid land away from less productive users.

Apart from a few attempts (Gebru, 1988; Dessalegn, 1992, 1994; Yigremew, 2001) there has not been a thorough, large-scale and systematic study of the patterns, diversity and rationale of alternative views on land tenure in Ethiopia.

THE CURRENT LAND TENURE SYSTEM

Despite the prevailing and continued debate about the Ethiopian land tenure system, empirical investigations from the point of view of the major stakeholders, the farming community, are not adequately documented. The debate was largely between professionals, academics, opposition political groups and the government or the ruling party. The perceptions of farmers regarding the problem and their alternative suggestions were seldom taken into consideration. By contrast, the current study covers almost all parts of the country, with large numbers of samples, and a wide breadth of information has been gathered. A notable exception is that the Gambela region was excluded, because of the lack of enumerators who can speak the local language. Some findings of the research work are given in this paper.

Research methodology

The launching of the land policy project at EEA/EEPRI was followed by a workshop to increase awareness, which discussed the importance of the land tenure problems in Ethiopia, and suggested the nature of the land policy research project to be conducted by EEPRI. In addition, groups of researchers and academics with notable experience and knowledge in the field have given their comments and suggestions on the contents of the fieldwork and types of data to be gathered.

Careful and sufficient thought was given to the sampling framework and the preparation of the fieldwork questionnaire. The draft questionnaire was scrutinized by various experts in the field with different backgrounds: agriculturalists, sociologists, geographers, economists, etc. The questionnaire was designed to collect diverse information on all aspects of rural land tenure, farmers' perceptions and smallholder agriculture in general.

Sample cases for the rural survey were selected based on established criteria to capture as many factors as possible that are thought to have influence on agriculture and rural life in general, with particular relevance for the land tenure system. The criteria used were agro-ecology, population density, market/infrastructure access, farming systems and previous experience in agricultural extension practices.

The Ethiopian Ministry of Agriculture with assistance from FAO and previous studies classify Ethiopia's 525 woredas (areas)/ districts into 18 broad agro-ecological zones (AEZs) (Ministry of Agriculture, 2000). The zones are considered to reflect the broad variation in temperature, soil moisture and altitude of the area they represent. The study conducted by the Ministry of Agriculture also identified 49 subagroecological zones that are more homogeneous in terms of climate, physiography, soil type, vegetation, animals, farming system and land use. It was decided to include at least one woreda from a given AEZ^[87] and maintain the proportion between the number of woredas to be selected from a given AEZ and its percentage geographical share. Forty-one woredas were randomly selected to represent the different AEZs.

Other non-agro-ecological factors considered in sample selection were population density, farming system and better access to markets/infrastructure. Moreover, some woredas known for their previous success in implementing the agricultural extension programmes were considered in the site selection. Although AEZs, which are primarily defined in terms of altitude and temperature, can ultimately influence population density and farming system, these variables (population density, farming system and market) were considered independently to capture their impact in a better and direct way. A further 29 woredas were thus selected based on the non-agro-ecological factors used for the selection. The total number of woredas covered in the study was thus 70.

Following a multistage sampling, an attempt was made to capture regional disparities within the selected variables. Accordingly, the samples were further classified geographically into regions, zones and woredas. The total number of samples to be selected from each group (strata) were distributed to the different areas in proportion to their percentage share within the selected variable.

Two peasant associations (PAs) were selected from each selected woreda randomly; one PA was selected from PAs near the woreda capital or the nearest town (to reflect availability of improved infrastructure) whereas the other was very far from the nearest town. Sixtyone farmers from each PA were selected randomly; 8 540 farmers from the entire country were randomly sampled using the aforementioned methods.

TABLE 1
Average landholding per household and per working age population (percentage of sampled households)

The field survey was carried out during August 2001. Preparation included translation of the questionnaire into one of the local languages (Amharic), training of enumerators, and pretesting of the questionnaire; together with actual data collection this took a period of about 6 weeks. About 140 trained enumerators and ten supervisors participated in the survey. A database program called FoxPro was used for data compilation and organization. A complied database was processed using SPSS statistical software.

The existing land tenure system in Ethiopia

Landholding

One of the key issues raised in relation to land policy is the degree to which the size of landholdings can adequately support the livelihood of the farmer. As discussed earlier, a number of researchers (e.g. Dessalegn, 1984, 1994) have raised the issue of the gradual conversion of Ethiopian agriculture from small-scale agriculture to micro-agriculture that cannot reduce the poverty of the farmers. Farm size in all sample households ranges from zero (landless) to 10 ha, although the latter are very small in number and usually found in the less densely populated regions of Somali and Afar. The mean size is about 1.02 ha per household and this average declines slightly to about 0.96 ha when we remove samples from the Somali and Afar regions (Table 1).

When we look at the figures in relation to holding per active farm labour force (landto- labour ratio) we find that the average holding is significantly reduced to 0.38 ha. The frequency distribution of holdings shows that nearly 75 percent of the sample households have less or equal to 0.5 ha of land to till per active farm labour force in the family whereas about 93 percent of these families have 1 ha or less. Equally importantly, the number of landless farmers (despite the distributive nature of the prevailing land policy) is significant, amounting to about 11 percent of the farming population in the sample.

As expected, farm-holding is not uniform throughout the country and differs significantly from region to region and depending on farming practices and other operating factors. The prevalence of smaller holdings of up to 0.5 ha in the sample reaches 55 percent in Tigray, 40 percent in Amhara region and 34 percent in the Southern region. It is lowest in the Somali region (3.1 percent). Landholding ranges from as low as 0.22 ha per active farm labour force in Tigray to a high of 1.61 ha in the Somali region. The highly populated highland regions of Amhara and SNNPR (South Nation Nationalities and People Region) have an average holding of about one-third of a hectare whereas Oromia has an average holding of 0.40 ha.

TABLE 2
Income, productivity and food security of sample households during the 2000/01 cropping season by region

ME = man-equivalent, measure of household labour force.

*An estimated 1 qt cereal equivalent is added from enset.

Net farm income is the revenue from all farm outputs less all direct and indirect costs excluding family labour input (but accounting for paid labour cost). Household income is all reported income from farming and non-farm activities for the whole household. Per capita household income is calculated as total household income divided by the number of adult equivalents in the family.

The regional variation in farm holdings is to a certain degree replicated in the holding patterns by major farming systems. As expected, size of holding is very low in the enset (false banana)-dominant regions of the south, on average less than 0.25 ha, whereas wheat-dominant farming areas have the highest average size holding, about 1 ha.

Clearly, given this distribution, the claim that the existing landholding system will reduce landholding size to an non-viable proportion with time cannot be discredited. In fact, this is one of the key issues that any land policy (or any development policy) in Ethiopia needs to address. Whether the current size of holdings can provide sufficient income to farmers to enable them to live above the poverty line via increased productivity and the use of modern technology is an issue that will be discussed later. What is clear from these data is that with the expected increase in the farming population in coming years, it is difficult to see how the farming population can escape poverty without a significant creation of non-farm employment in the near future to absorb the additional population.

Farm income

The income of the farming population closely follows the patterns observed for size of holdings. As shown in Table 2, net farm income is higher in regions with high average holding and lower in regions with low average holding. Sample households in Benishangul-Gumuz have the highest net farm income at Br2 175 whereas net farm income is lowest in the Southern region at Br512 and the average for all farms in the five regions is Br1 162. The average household income for all the sample households is about Br1 549. The regional distribution of average household income is slightly different in that households in Tigray have significantly higher incomes not commensurate with average size of landholding. Tigray is the second highest household income earner next to Benishangul. This is clearly because of the availability of more off-farm employment in the region and/or the availability of more food aid.

What is more revealing about the meagre size of income in rural Ethiopia are the numbers per capita household income. Table 2 shows that for the five regions considered the average income per capita (adult equivalent) is about Br333, roughly about US$39 a year or a little more than US$0.1 per day. Tigray has the highest income in this category at Br431, more than 2½ times that earned by the average person in the lowest income region of the south. Per capita household income in Amhara and Oromia are roughly the same whereas Benishangul comes a very close second to Tigray.

In addition to the average income figures given above, a more detailed picture of farm income can be obtained from income frequency distribution (data not shown). The distribution of the net income is skewed to the lowest income categories. Close to one-fifth of the sample cases earned only up to Br100; 46.7 percent of the cases had a net farm income up to Br500. Only slightly more than one-third of the households had a net income above Br1 000. Only 6.3 percent earned above Br4 000.

When disaggregated by farming systems, wheat-dominant areas have the highest per capita household income at about Br602, followed very closely by teff-producing areas (less only by Br10) (data not shown). The enset-producing farmers earn less than one-tenth of the wheat and teff farmers. The extremely low level of income of the enset-producing and densely populated regions of the south is a clear indication of the importance of the availability of land in determining household income.

The rather low level of income of farmers in Ethiopia has obvious implications for the poverty situation that prevails in the country in general and in the rural population in particular. If we consider the low poverty line calculated by the government (MEDAC, 1999), estimated at Br1 075 per annum per household (about US$126 dollars per annum per household), 63 percent of the sample households lie below this national poverty line if we consider only net farm income. When the distribution of total household income is considered (net farm income plus non-farm income) the proportion of the population below the poverty line decreases to 53 percent, but this is still a very large percentage. This figure compares unfavourably with the 47 percent (and declining) estimated by the government in rural Ethiopia (MoFED, 2002) and is commensurate with the increased poverty reported by the public in the recently held PRSP consultations both by the government and civil society institutions (EEA/EEPRI, 2002b; PRSP Secretariat, 2002). This clearly shows that the poverty situation in rural Ethiopia has reached rather desperate levels.

Farm productivity

The low level of income of farm households is a result of both the small size of landholding and the low level of productivity in Ethiopian agriculture. As can be seen from Table 2, average net farm income per hectare of cultivated land (an indicator of land productivity) averages Br2 300 in the 2000/01 cropping season, which is about US$282. This figure is twice the average value added per hectare of agricultural land estimated by the World Bank for 1994 - 1996.^[88] In terms of land productivity, the Southern region has the highest value followed by the Tigray and Amhara regions. The higher productivity per unit of land in the Southern region may be attributed to cash crops such as coffee, which unfortunately is today fetching lower prices.

Average net farm income per labour force (a rough measure of labour productivity) stands at Br413; the highest figure obtained is from Benishangul where the average size of holding is relatively high, and the lowest figure is from the Southern region where size of holding is extremely low. The other three regions have roughly similar returns per labour force. This again is a clear indication of the close link between size of holding and labour productivity in Ethiopian agriculture.

In terms of farming systems (data not shown), land productivity is highest in the teff-dominant farming regions followed by wheat-producing regions, whereas return to labour is highest in wheat-dominant farming systems followed by teff-dominant regions. This seems to reflect the fact that teff farming is more labour intensive than wheat farming, tending to reduce labour efficiency for the former. What is interesting in these figures is the low level of income obtained in the coffee-producing regions during this cropping season. Given the history of relatively high levels of income of coffee farmers, the fact that coffee producers earned only 32 percent of the net farm income per hectare and only 32 percent of the net farm income per labour force earned in teff-producing regions is a clear indication of the degree to which coffee farmers in Ethiopia have suffered because of the precipitous collapse of coffee prices in the international market.

Farm size and household food security

Given the low level of productivity and the small size of holdings, one of the key issues that needs to be addressed is the degree to which farmers can attain food security for their families within the current production system. Respondents were asked if they have enough land to produce sufficient food for their families. Nearly 77 percent of the respondents feel that they do not have enough land to produce sufficient food for their families.

If we look at the food security issue strictly in relation to cereal production and by region, the picture we get is that on average cereal production provides slightly above the minimum requirement for food security. However, regions in the far north and south are facing serious food deficits. Including the landless farmers reported in this study, the proportion of food-insecure households for all the sample households (cases with landholding less than the minimum area required for minimum food production) is 48 percent (Table 2). Variation by regions is also quite significant. A rough estimate of the minimum area required to attain food security, given the current average productivity of cereals, shows that households below this minimum size of land range between 41 percent in Oromia and 75 percent in Tigray. Considering the role of enset as a staple food in southern regions, the prevalence of food insecurity (at least in terms of meeting minimum food needs) is lowest here at 26 percent.

The minimum cultivable area required per household ranges from 0.86 ha in Amhara to 1.01 ha in Oromiya, reflecting differences in average household size. To account for the role of enset^[89] production in food security, particularly in the predominantly enset-producing regions, the amount of enset production per household was considered. If enset production is taken into account in addition to cereals, the food production level of the average household increases and the minimum area required for food production falls to 0.56 ha. This reduces the proportion of households with landholdings of less than the minimum requirement to 26 percent in Southern regions compared with a value of 47 percent if only food grains are considered.

Current land policy, farmers' opinion and attitude

As stated earlier, the debate on land tenure in Ethiopia is characterized by extreme positions revolving around state/public ownership and freehold. Furthermore, this debate has rarely involved the main stakeholders, the farmers themselves. One of the central objectives of this research was to register farmers' opinions on the prevailing landholding system. It was important not only to find out what farmers think, but equally importantly to discuss the validity of the arguments on both sides of the debate. We will come back to the latter issue in due course. First, let us look at farmers' opinion regarding the prevailing land policy.

TABLE 3
Farmers' perception of tenure security, evaluation of the existing tenure system and alternative choices (percentage of sampled households)

*There is a variation in the response of the other regions (Afar, Somali, and Benishangul). In Benishangul 62.3 percent of the cases feel that the current system is not good.

Farmers' view of the current tenure arrangement

As can be seen from Table 3, the majority of farmers (61 percent) in the sample think that the current land tenure system is good and 38 percent believe that the system is not. In this sense, those who argue for state/public ownership of land seem to reflect the majority opinion of farm households. The degree of support for the existing tenure arrangement, however, is not as strong as the government claims and varies significantly from region to region.

In the predominantly pastoral areas of Somali (86 percent) and Afar (68 percent) as well as in the densely populated region of the South (78 percent), support is quite strong. There is strong opposition in the relatively sparsely populated Benishangul region where over 62 percent of the households oppose the current system. In the highland region of Tigray some two-thirds of the households are in favour of the current system whereas nearly 47 percent of the households in Amhara and 43 percent of the households in Oromiya do not like the current tenure arrangement; these do not represent a small minority.

The most important reason given for support of the existing tenure arrangement seems to be related to the user rights granted to farmers, as reported by 37 percent of those questioned, followed by 20 percent of the respondents who emphasized the equity/justice brought by the tenure system as a continuation of the 1975 land reform. Another 19.7 percent of the surveyed farmers feel that with the existing system they do not fear losing their plots.

Farmers opposing the prevailing land policy give their inability to obtain additional land as the dominant reason for their dissatisfaction (44 percent), followed by their inability to buy or sell land (13 percent), fear of losing land (12 percent) and injustice in land administration (11 percent). The reasons provided are closely related to access to land, insecurity of tenure, absence of formal land markets and administrative injustice.

Tenure security issues. Insecurity of tenure can be triggered and affected by a variety of factors. At the centre of the issue is the degree to which the holders feel that their rights to the land will not be arbitrarily violated. In this sense, the most secure tenure arrangement is largely believed to be freehold, which provides a full sense of ownership to the holder provided that there is a properly functioning and fair land adjudication system. Additional insecurity factors include expectations towards further land redistribution, i.e. how long farmers feel they can retain their current holding.

Despite government claims that farmers feel they own the land they cultivate, the overwhelming majority of farmers (84 percent) know that as currently stipulated by laws and in practice, the land belongs to the government. Only a very minor 4.4 percent believe that the cultivator has ownership. Furthermore, only 3.5 percent of the households believe that they can retain their current holding for over 20 years, and a significant majority (76 percent) do not feel secure that their claim towards their existing holding will last over 5 years. Obviously, this has important implications for incentives of farmers to put long-term investment in their current holding. In terms of regional variation, except for Benishangul and the Southern region^[90] where the majority of respondents feel that they will retain their current holdings indefinitely, most respondents in Tigray, Amhara and Oromia are not sure how long they will be able to keep their current holding.

Another indicator of insecurity is farmers' expectations of future land distributions. Despite the fact that most regional governments have publicly dissociated themselves from possible future land redistribution, only a minority of farmers (27 percent) are convinced that this will not occur in the future. By contrast, a significant majority (73 percent) feel either uncertain about the future or are certain that there will be redistribution (Table 3). Of those farmers who think that there will be redistribution, a large majority (70 percent) feel that this could take place before 2005. On average, 10 percent of the sample have experienced land redistribution since 1990.

In terms of winners and losers in a potential land distribution, some 45 percent of the sample households believe that they will benefit if such distribution takes place, whereas the remaining 55 percent are either uncertain of the outcome or believe that they will lose a part of their land from further redistribution (data not shown). Given the fact that the primary underlying motive of redistribution is a concern for equity, it is reasonable to expect farmers' perception of potential benefit from distribution to be a function of their current size of holding. Accordingly, farmers with relatively larger holdings fear a possible loss of land from redistribution whereas those with small holdings are optimistic about the outcomes of redistribution. A related issue is what to do with the increasing landless population in the absence of redistribution. As indicated earlier, land redistribution is becoming less and less viable^[91] as an option to address the problem of landlessness. Both the unpopularity of redistribution among peasants and the acknowledged small size of holdings prohibit redistribution as a viable policy option for the government in many regions of the country. The more viable alternative, as discussed earlier, is to devise a coherent and urgent strategy that would enable the rapid growth in non-farm economic activities in urban centres of the country.

Farmers' alternatives to the current system

The issue of tenure security seems to be a more important consideration for farmers than the particular form of ownership. The proportion of surveyed farmers who preferred public or state ownership with secured use rights was about 48 percent followed by private ownership with full transfer rights at about 31 percent as their ideal choice (Table 3). In both cases, security of tenure is the key concept. If we assume that those who did not give any preference can be persuaded to adopt freehold, the farming population could roughly be split into two equal parts between the two alternatives.

Further information supporting this contention is revealed if we look at

TABLE 4
Farmers' attitude towards current land tenure and their participation in long-term land management practices (number with percentage in parentheses)

farmers' second preferred choice of tenure arrangement (data not reported). Support for private ownership increased by the same percentage of those who were undecided as to their first choice to almost 50 percent whereas support for state/public ownership declined to a mere 15 percent. This should serve as a caution to advocates of state holding regarding the amount of public support they perceive to have. Clearly, farmers seem to take a more pragmatic approach than others involved in the debate. Most farmers are not keen on unrestricted freehold as revealed by the large support for state ownership with secured rights as their first choice. Neither do they seem to be willing to sacrifice security of tenure when they feel that state ownership fails to provide this. Indeed, the data seem to suggest that a more flexible landholding system centred around providing security of tenure and that takes into account local sensibilities and includes a mixture of private, state and communal holding might generate significant support among the farming population.

Land management practices

One of the key issues related to land tenure is the degree to which the tenure arrangement encourages sustainable farm practices. It is generally believed that a more secure tenure system provides the necessary incentives for farmers to manage their land more efficiently and invest in land improvement. Assessment of better land management is evaluated in relation to farm practices such as crop rotation, terracing, fallowing and tree planting. In general, the prevailing land management practice does not encourage agricultural sustainability in the country.

One practice that most farmers use to improve land fertility is crop rotation, which over three-quarters of the households employ. The majority of farmers do not use any other form of sustainable practice. The lowest level for any sustainable practice (23 percent) is fallowing, which can be explained by the small size of plots. Planting trees is done by only 39 percent of households. Terracing, by contrast, is applied by 47 percent of the sample households. Lack of tenure security could be one of the prime reasons for weak land management practices exhibited by Ethiopian farmers (Deininger et al., 2003). Interestingly, however, there does not appear to be any discernible pattern in use of sustainable farm practices between plots of differing sizes. This could strengthen the argument that issues of security of tenure are more important than those of plot size or land availability.

The data show that engagement in longterm land improvement practices is related to farmers' opinion on the current land tenure system, and perceptions regarding security (Table 4). Sixty-two percent of farmers who support the current system, for example, built terraces and planted tress. By contrast, roughly 50 percent of farmers who feel more secure (i.e. those who said "yes" to the question "Do you believe that you will continue to own your current holding?") built terraces and planted trees against the 36 percent of farmers who were uncertain about continuing to own their current holding. Even when we classify the relatively more secure farmers into groups based on their perceived degree of security as reflected by their expected years of continued ownership, those who are more secure (who believe that they will retain their current holding size indefinitely) are better engaged in long-term land improvement practices compared with those who expect to own plots for shorter periods (Table 5).

TABLE 5
Farmers' perceptions of security of holdings and their involvement in long-term land management practices (number with percentage in parentheses)

Evaluation of the argument in support of the existing system

Government's fear of land sales and migration. As discussed in detail in previous sections, excluding ideological considerations, the government's or the ruling party's argument for the continuation of the existing system rests solely on concerns for the alternative. In particular, it is claimed that freehold will lead to massive eviction of the farming population as poor farmers are forced to sell their plots to unscrupulous urban speculators particularly during periods of hardship. One of the most interesting results of this survey is the decisive rejection of this claim. Over 90 percent of the households surveyed indicated that they would not sell their land wholly or partially if given the right to own their plots. Only 4.5 percent of the households would be inclined to sell their land given the opportunity (data not shown).

The reasons given for the unwillingness to sell reveals a rational response on the part of farmers. The overwhelming majority of farmers (70 percent) will not sell their land because they have no viable alternative, and a significant minority (17 percent) said they would never sell their land no matter what the circumstances. How then would they cope when faced with periods of hardship? The response to this question largely supports the recent World Bank position for the important role that formal rural land rental markets have in enhancing efficiency, and increasing access to credit and possibly security (Deininger and Binswanger, 1999: pp. 255 - 267). The data clearly reveal that most farmers would rather rent their land during stressful periods (47 percent) compared with any other alternative. In other words, in addition to all the other benefits of rental markets suggested in the literature, the availability of formal land rental markets will serve as a cushion to enable farmers to withstand unfavourable circumstances by temporarily renting their land rather than selling it. This also suggests one option to Ethiopian policy makers to consider in a possible reform programme without necessarily resorting to the privatization of rural land.

The potential of the extension programme to compensate for the problems of the existing land tenure system. The government's position in favour of the current land policy is not entirely because of a lack of recognition of the problems associated with it. As suggested earlier, it results mainly from a fear of the alternative. A corollary to this is the deeply held conviction that the system's failings can be adequately compensated through rapid increases in productivity of smallholders through the extension programme that the government has been undertaking since 1994. The validity of this claim needs to be measured in relation to the increase in productivity that has been achieved by farmers included in the extension programme compared with those who are not and their resulting increase in household income.

TABLE 6
Performance of farms by size of land holding and participation in extension programme

To evaluate this claim, the survey separated respondents between those who are currently participating in the extension programme and those who are not. An important point that must be mentioned from the outset is that even if the claim is entirely valid, the number of farmers not included in the extension programme is far greater than those who are. For this sample survey, nearly 63 percent were not participating in the programme whereas 34 percent were (Table 6). Therefore, it may be some while before a successful extension programme could make a marked difference in reducing rural poverty and therefore other measures would be required.

Taking farm income and resource productivity as performance indicators, we attempted to compare farmers who participated in the agricultural extension programme and those who did not when this survey was conducted. As can be seen from Table 6, land productivity is higher by only 5.8 percent for those participating in the extension programme as measured by net farm income per hectare. Per capita food production is also higher by about 33 percent. The participating farmers earned on average Br285 more. Although not significant, at least in these two cases, the government's claim that the extension programme could increase productivity does seem valid. Curiously, the same is not true with regard to labour productivity (net farm income per labour force), which is higher for non-participants by about 2 percent. This obviously contradicts the government's argument that the extension programme could compensate for the problems with the land policy.

Looking at the disaggregated data by size of holding, a more complicated picture emerges. The first clear conclusion is that the size of the holding is very important. With very minor exceptions, in almost all categories larger size holdings perform better than smaller size holdings irrespective of participation in the extension programme.

The tentative conclusions that emerge from analysis of the data discussed in this section are not in agreement with those of the government. Clearly, the average performance of those who participated in the extension programme is not better (in fact is usually worse) than those who did not participate. Even for those involved in the extension programme, landholding remains a key determinant for success and better performance. As landholding declines, per capita food production and farm income also decline, indicating that extremely small-sized farms cannot be made productive even with improved technology, and certainly not enough to address rural poverty issues by the extension programme alone. This clearly reveals the need to address the problems associated with the landholding programme rather than fully relying on the extension system to solve rural poverty issues let alone to solve the broader development problems of the country.

CONCLUSION AND POLICY IMPLICATIONS

This study clearly indicates that the problems faced by Ethiopian agriculture are very much related to the existing landholding system, although not exclusively. Poverty and food insecurity are widespread, and these are the main challenges with which the agricultural sector is faced.

Analysis of farmers' preferred alternative choice of land tenure system has revealed that the issue of tenure security is a more important consideration than the particular form of ownership. The results clearly show that farmers seem to have a more pragmatic approach than those involved in the debate. Most farmers are not keen on unrestricted freehold, as indicated by the large support for state ownership with secured rights as their first choice. Neither do they seem to be willing to sacrifice security of tenure when they feel that state ownership fails to do so. Instead, the data seem to suggest that a more flexible landholding system centred around providing security of tenure and that takes into account local sensibilities including a mixture of private, state and communal holding would be a favourable option among the farming population rather than one fixated by the public/private dichotomy that characterizes the current debate in the country.

Further analysis using the survey data differentiating tenure security and transferability to explore determinants of different types of land-related investment and its possible impact on productivity shows that government action to increase tenure security and transferability of land rights can significantly enhance rural investment and productivity.

To some extent, the government's claim that the extension programme can increase land productivity is valid. The findings in this study do not unequivocally support the government's argument that the agricultural extension programme can overcome the problems with land policy. Looking at the impact of size of landholding on farm performance, the first obvious conclusion from the survey data is that the size of the holding is of considerable importance. With minor exceptions, larger size holdings perform (with regard to income and food production) better than smaller size holdings, irrespective of the farmer's involvement in the agricultural extension programme. Room for sustainable intensification is very much limited where peasant agriculture is characterized by mini-plots of uneconomic size.

The farmers' de facto involvement in land transactions despite restrictions by law suggests that a suitable land policy will facilitate the operation of formal land markets to enable better allocation of this important resource. In a separate work using survey data from Ethiopia that empirically assessed the determinants of participation in land rental markets compared with those of administrative land reallocation (Adnewe et al., 2003), the results indicate that land rental markets outperform administrative reallocation in terms of efficiency and poverty, i.e. it has been found to favour the poor.

The following policy issues and processes are tentatively suggested for consideration:

• The land tenure policy and strategy must be governed by a socio-economic rationale that includes efficiency and equity considerations.

• The land tenure policy, programmes and measures must give due attention to the farmers' point of view.

• A more flexible landholding system centred around providing security of tenure that takes into account local sensibilities including a mixture of private, state and communal holdings might generate significant support among the farming population and resolve the structural problems facing the sector.

• A conducive environment must be created to encourage rural land markets in view of their role in improving resource allocation, efficiency, productivity and mobility.

• Policy-makers and agricultural development experts must give serious consideration to the fact that there is limited room for intensification with the prevailing mini-plots operated by subsistence farms. Ways and means must be sought to create viable farm sizes that will enable economic use of farm input in a way that leads to increased intensification in a dynamic and sustainable manner.

• Although the issue of rural land tenure is a major component of agricultural and rural development in Ethiopia, the problem of agrarian transformation should be looked at from a wider and long-term perspective.

• Given the high rate of increase in the rural population and the already small size of holdings, poverty reduction needs input beyond the farming sector. There is a clear need to devise a coherent and urgent strategy that will enable the rapid growth in non-agricultural sectors.

• Creation of strong rural - urban links that encourage mobility of the farming population should be addressed as a matter of priority.

• Given these findings, the government needs to revisit its land policy in order to devise an appropriate policy framework that addresses the problem of the agricultural sector. There should be free and informed dialogue to prepare the ground for constructive input from all sectors of the population, particularly those in the farming population. It is thus recommended that the government establish an independent commission composed of all stakeholders to study the problems of the current tenure system and suggest appropriate policy measures.

What can be said for certain is that as daunting as the current situation is, the future for the rural population could be catastrophic unless the situation is boldly and urgently addressed. The current solution provided by the government to the agrarian crisis rests on the conviction that land productivity and rural incomes will be increased substantially through the agricultural extension package, thereby reducing rural poverty. The government also sees this policy as the main engine of growth for the whole economy through the demand-driven industrialization it will engender. This claim for rural Ethiopia is seriously questioned in the empirical analysis provided in this paper.

Cour (2002) has recently discussed the interrelationship between urbanization and rural development, and has seriously questioned the validity of the primacy given to agriculture as the engine of growth for the whole Ethiopian economy. Cour argued that given the current level of urbanization (the urban-to-rural ratio is 0.14; roughly seven farmers supplying one urban consumer), even if output and productivity substantially increases as the strategy (ADLI) hopes, the result will be a decline in prices as a result of shortage of demand, leaving the farmer possibly worse off. This is essentially what we have observed in 2001/2002. Two consecutive bumper harvests (largely caused by favourable weather conditions) have led to a precipitous collapse in cereal prices. Cour's recommendation was that policy makers need to give urgent attention to measures to increase urbanization, including lifting the restrictions (which are inherent in the current land policy) on farmers moving to urban areas. According to Cour's projections, if current urbanization and population trends continue, the Ethiopian population will increase to about 110 million in 2025, of which the rural population will account for 88 million, or a further 30 million more mouths to feed for an already overcrowded rural population.

This is clearly not sustainable. What this and other related studies have shown is that Ethiopia has to make some hard choices in the near future if a widespread disaster is to be avoided. What is needed is a careful and deliberate initiative, backed by research, to identify an appropriate and pragmatic policy mix that could address the looming crisis of land tenure and its consequences. We hope this research could serve as a starting point for further informed policy dialogue.

BIBLIOGRAPHY

Adnewe, B., Deininger, K., Demeke, M., Gebre- Selassie, S. & Jin, S. 2003. Market and nonmarket transfer of land in Ethiopia: implications for efficiency, equity and non-farm development. World Bank Working Paper No. 2992. Washington, DC, World Bank.

Alem-Ante, G.S. 1970. Legal aspects of agricultural land disputes, pp. 218 - 234. MLRA.

Barrows, R. & Roth, M. 1989. Land tenure and investment in African agriculture: theory and evidence. LTC Working Paper 136. Madison, USA, Land Tenure Center, University of Wisconsin- Madison.

Bruce, J. & Shem Migot-Adholla, eds. 1993. Searching for land tenure security in Africa. Iowa, USA, Kendal Hunt Publishing Co.

Cohen, J. & Weintraub, D. 1975. Land and peasants in Imperial Ethiopia: the social background of a revolution. Assen, The Netherlands, Van Gorkum.

Cour, J.-M. 2002. A demo-economic analysis of long term structural changes in Ethiopia. In Berhanu Nega, eds. The role of urbanization in the socio-economic development process: demographic challenges, urban growth, spatial development and agriculture. EEA/EEPRI Publications.

Deininger, K. & Binswanger, H. 1999. The evolution of the World Bank's land policy: principles, experience, and future challenges. In Research Observer, Vol. 14, No. 2. Washington, DC, World Bank.

Deininger, K. et al. 2003. Mechanisms for land transfer in Ethiopia: implications for efficiency, equity, and non-farm development.

Dejene, A. 1999. In Tegegne Gebre Egziabher et al., eds. The role of women in rural development in Ethiopia, pp. 329 - 353.

Dessalegn, R. 1984. Agrarian reform in Ethiopia. Uppsala, Sweden, SIAS.

Dessalegn, R. 1992. The land question and reform policy: issues for debate. Dialogue, 1(1): 43 - 57.

Dessalegn, R. 1994. Land policy in Ethiopia at cross-roads. In R. Dessalegn, ed. Land tenure and land policy in Ethiopia after the Derg. Norway, Reprocentralen AVH.

Dessalegn, R. 1997. Manufacturing poverty: rural policy and micro-agriculture. Paper presented at the Workshop of the Land Tenure Project of the Institute of Development Research, Addis Ababa.

EEA. 1999/2000. Annual report on the Ethiopian economy, Vol. 1, edited by Befecadu Degfe and Berhanu Nega. Addis Ababa, EEA.

EEA/EEPRI. 2002. A research report on land tenure and agricultural development in Ethiopia. Addis Ababa, EEA/EEPRI.

Ege, S. 1997. The promised land: the Amhara land redistribution of 1997. Working Papers on Ethiopian Development No. 12. Norway, Reprocentrlen AVH, ANTU.

Fesseha, E. 1970. The need for progressive taxation on unutilized lands in Ethiopia, pp. 129 - 148. MLRA.

Gebru, M. 1998. Privatization of rural land: to protect whose security? Paper presented at the Inaugural Workshop of the Forum for Social Studies. Addis Ababa.

Gilkes, P. 1975. The dying lion: feudalism and modernization in Ethiopia. London, Julian Friedman.

Haile-Gabriel, A. 2000. In A. Mekonnen & D. Aredo, eds. Thriving informal land markets and patterns of entitlement redistribution among peasant households, pp. 67 - 86.

Hoben, A. 2000. Ethiopian rural land tenure policy revisited. Paper presented at the Symposium for Reviewing Ethiopia's Economic Performance 1991 - 1999. Organized by Inter-Africa Group. Addis Ababa.

Kidanu, A. & Alemu, T. 1994. In A. Dessalegn, ed. Rapid population growth and access to land: coping strategies in two PAs in North Showa, pp. 35 - 53.

MEDAC. 1999. Poverty situation in Ethiopia. Addis Ababa, Welfare Monitoring Unit.

Mekonnen, L. 1999. Land reform and its impact on the environment: the case of Gidan Wereda. In Proceedings of the First National Workshop of NOVIB Partners Forum on Sustainable Land Use. Addis Ababa.

Ministry of Agriculture. 1975. Agricultural sample survey 1974/75, Vol. I. Addis Ababa.

Ministry of Agriculture. 2000. Agro-ecological zones of Ethiopia. Addis Ababa, Natural Resources Management and Regulatory Department.

MoFED. 2002. District Level PRSP Consultation Report. Addis Ababa.

Mulat, D., Kelly, V., Jayne, T.S., Said, A., Le Vallee, J.C. & Chen, H. 1998. Agricultural market performance and determinants of fertilizer use in Ethiopia. Working Paper 10. Grain Market Research Project. Addis Ababa, MEDaC.

Provisional Military Government of Ethiopia. 1975. Proclamation No. 31 of 1975, a proclamation to provide for the public ownership of rural lands. Addis Ababa.

PRSP Secretariat. 2002. Public voice, Vol. 1, No. 2. Addis Ababa, EEPRI.

Robert, S. & Chernet, E. 1996. A linear model for predicting Enset plant yield and assessment of Kocho production in Ethiopia. UNDP Emergency Unit for Ethiopia (available at www.Africa.upenn.edu/eue_web/enset96.html).

Salazar, A., Branao, P. & Feder, G. 1995. Regulatory policies and reform: the case of land markets. In C. Frischtak, ed. Regulatory policies and reform. A comparative perspective. Washington, DC, World Bank.

Tekie, A. 2000. Farmers' willingness to pay for tenure security. In A. Mekonnen & D. Aredo, eds. Institutions, resources and development in Ethiopia, pp. 87 - 112. Proceedings of the Ninth Annual Conference on the Ethiopian Economy. Addis Ababa, Ethiopian Economic Association.

Tesfaye T. & Bedassa, T. 2002. Access to agricultural land, farm size, quality and productivity relations in Ethiopian Highlands. Kalamazoo, USA, Western Michigan University Department of Economics.

Toulmin, C. & Quan, J. 2000. Evolving land rights and policy in sub-Saharan Africa. In C. Toulmin & J. Quan, eds. Evolving land rights, policy and tenure in Africa. London, DFID/IIED/NRI.

World Bank. 1992. Ethiopia: toward poverty alleviation and a social action program. Report No. 11306-ET, November. Washington, DC.

Yigremew, A. 1997. Rural land holding readjustment in West Gojjam, Amhara Region. Ethiopian J. Develop. Res., 19(2): 57 - 89.

Yigremew, A. 1999. The rural land tenure system in Ethiopia since 1975: some observations about its impact on agricultural production and sustainable land use. In Tegegne Gebre Egziabher et al., eds, pp. 205 - 225.

Yigremew, A. 2000. Preliminary assessment of the impacts of the post-Derg land policy on agricultural production: a case study of land redistribution in two communities in north-west Ethiopia. In Workneh Negatu et al., eds, pp. 173 - 190.

Yigremew, A. 2001. Some queries about the debate on land tenure in Ethiopia. In M. Demeke & T. Woldemariam, eds. Proceedings of the Tenth Annual Conference on the Ethiopian Economy. Addis Ababa, Ethiopian Economic Association.

Yigremew, A. 2002. Review of landholding systems and policies in Ethiopia under the different regimes. Addis Ababa, Ethiopian Economic Association/Ethiopian Economic Policy Research Institute.