Command and Control: How Does It Work? The Case of Land Market Law Restrictions in Ethiopia

First draft! Not for citation. Comments welcome! Command and Control: How Does It Work? The Case of Land Market Law Restrictions in Ethiopia By Stein Holden 1 and Hosaena Ghebru 2 1 Centre for Land Tenure Studies, Norwegian University of Life Sciences, 1432 Ås, Norway. Email: stein.holden@umb.no 2 International Food Policy Research Institute, Washington, D. C., USA. Email: h.g.hagos@cgiar.org Abstract This study assesses the de jure and de facto land market law restrictions in Tigray Region in Ethiopia and the extent and speed of enforcement of the new land market law restrictions that were introduced in 2006 and stating that not more than 50% of the farm can be rented out and that land should be confiscated from households that have been away from their community for more than two years. Both law restrictions are unpopular in the rural communities in the region. Only the land taking restriction for migrant households had been enforced through local Land Administration Committees by 2010, while the rented out area share restriction is commonly violated and not yet enforced. 1. Introduction Property rights to land and the role of land markets for redistribution of land rights has been a central and controversial political issue in the history of many countries. With population increase and economic development land markets have developed despite many attempts to prohibit or suppress their development. The fear of markets and especially land markets has been strong and central in many radical ideologies while neo-liberal ideologies have advocated the advantages of leaving development to unregulated markets as most attempts to regulate markets have achieved mainly unintended negative effects. Historically there have been many attempts at prohibiting or regulating land markets, such as prohibition of land sales markets in many countries, land-to-the-tiller policies attempting to transfer property rights to tenants, and various contract regulations driven by efficiency, equity or sustainability objectives. Careful examination of many of these interventions have revealed that the interventions had unexpected unintended and counterproductive effects. For example, land-to-thetiller policies that aimed to enhance the access to land for landless and land-poor households have in many cases created tenure insecurity and have not only reduced their access but have also resulted in less efficient land use (Otsuka 2007). The recent global financial crisis has created a new awareness and political realization that state interventions are essential to mitigate financial crises and to protect vulnerable groups against the damages from market forces out of control and that can also lead to social unrest and political

instability. The recent land rush ( land grab ) that in particular has hit Africa has also created fears that unregulated land markets rapidly can lead to landlessness and increased food insecurity for large vulnerable populations in the African continent. On the other hand, higher food prices and growing demand for land could also be a business opportunity that could stimulate agricultural production on underutilized land in Africa, promote economic growth, employment and development. However, the legal frameworks and national policies in most African countries are far from adequate to ensure such a positive development. A recent study (Arzeki et al., 2011) has shown that the most rapid expansion takes place in the countries with the weakest institutional frameworks, increasing the risk that the feared negative outcomes can materialize. This study focuses on land tenure reforms wrt regulation of land markets; their political rationale, efficacy, and grassroots support in one African country, Ethiopia, which had a radical land tenure reform in 1975, is highly dependent on the land as the livelihood basis for its population of more than 80 million, is facing chronic food insecurity for a substantial share of its population (6-8 million depend on food aid regularly), and has at the same time opened up for investments by international investors on about 5 million ha of land. Critical voices argue that this contributes to the food insecurity of the country and has led to violations of human rights in the areas opened for investors. Supporters of this new policy argue that the areas opened for international investors are underpopulated and underdeveloped, and international investors can bring in investment funds and expertise that can help develop these areas and promote economic development in the country in the longer run. Yet there are massive contrasts with the Ethiopian highlands where land scarcity is growing and landlessness is emerging and youth unemployment is a growing threat. The core of this study focuses on the land reform and attempts at land market regulations in the Ethiopian highlands and will towards the end discuss these findings in the broader perspective of the national development strategy and investment policy. First we provide a historical back-drop on the Ethiopian land policies since the radical land tenure reform in 1975. We try to provide the rationale for the policies, including the changes in land laws and land tenure reforms that have happened since. We provide empirical evidence on the efficacy of the most recent reforms in one region of the country, Tigray, from which the political leadership in the country the last 20 years originates. The efficacy depends on the ability of the land administrative system to disseminate information about new laws and enforce their content. Such command-and-control law enforcement should have the highest chance of success in an area that is politically well connected to the political power. However, this may depend whether the law changes have grassroots support. Our working hypothesis is that socially acceptable parts of the law are much easier to implement than laws that involve social dilemmas that meet resistance among the people. Customary norms, perceptions and attitudes towards the reforms at the grassroots level are therefore of public interest. They have implications for the identification of intended and unintended effects of new laws and the need for law and policy revisions. Our study also has high relevance for similar and diverging ongoing reforms in other regions of the country. It also has relevance for the laws that are developed to promote commercial agriculture in the country by allocation of land to large international and national investors. The findings should also have broader relevance at a time when there is a high need for land tenure reforms in many countries with rapidly growing demands for land and weak institutional structures. 2. A Historical Backdrop Ethiopia is an exceptional country in Africa in the sense that it was never colonized. It therefore has a unique culture that has been less influenced by colonial powers. The economy has been described as

a feudal system up to the revolution and overthrow of Haile Selassie in 1974. Particularly southern parts of the country had a system with absent landlords and landless tenants while the rist system dominated in the north (Rahmato 1984). The rist system provided usufructuary rights to land to residents in a community whose ancestors also came from the same community. Land sales or mortgaging of land was not allowed but the broad possibility of claiming land that ancestors have used resulted in multiple claims from many relatives on the same land. This created tenure insecurity, many land disputes, and resulted in land fragmentation (Hoben, 1973). It also implied that absent landlords did not exist in the north and land rental activity was at a much more modest level than in the south. Cohen and Weintraub (1975) estimated that 7-16% rented land in the north against 37-73% in the south in the period 1967-70. In the south the few owners of the land were also allowed to buy and sell land. 2.1 The Radical Land Tenure Reform of 1975 and the Derg Period The 1975 land tenure reform in Ethiopia made all land state land. An egalitarian distribution of land to households based on their need was then implemented at community (Peasant Association) level. The maximum farm size was set at 10ha. None of the earlier landlords in southern parts of the country were allowed to have leadership positions in the Peasant Associations, meaning that they lost all power and almost all their land. Households only received user rights to the land as land sales and rentals were prohibited. Even hiring of labor was prohibited. It was therefore an extreme antimarket reform. In order to retain the egalitarian land distribution as new households formed and household sizes changes, land redistributions were introduced partly distributing undistributed land to new households or partly and increasingly redistributing land from households that had relatively more land than others related to their household size. Some land renting took place in this period in form of sharecropping contracts which were not considered to be subject to the restrictions on land renting. 2.2 Change in Government and More Market Friendly Reforms The Derg regime fell in 1991 and the new leadership originating from Tigray constitutionalized state ownership of all land and continued prohibition of all land sales. The new constitution granted access to land for all residents and opened for short-term land renting arrangements and bequeathing of land to children, while mortgaging of land remained illegal. The new land laws that were introduced also stopped the land redistribution policies that have created tenure insecurity and undermined investment incentives. An exception was the 1997 land redistribution in the Amhara region that aimed to take land from local leaders under the Derg regime. 2.3 Land Registration and Certification Ethiopia implemented broad-scale low-cost land registration and certification from the late 1990s that provided land certificate recipients perpetual user rights to land, rights to lease out the land in short-term contracts, and rights to bequeath the land to children and grand-children. The land laws were revised in the following decade and local land administrations were established at the community level with a responsibility for implementing the new laws. It has remained illegal to sell or mortgage land and the new laws also restricted the extent of land renting not only in form of the duration of land rental contracts but also set a maximum share of the farms (max. 50%) that could be rented out. Several studies have shown that the land registration and certification increased tenure security and this enhanced the allocative efficiency of the land rental market (Holden et al., 2011; Deininger et al., 2011). Particularly female-headed households were found to have become

more willing to rent out their land and this also enhanced their welfare up to 2006 when the revised laws were introduced (Holden et al., 2011; Holden and Ghebru, 2011). 2.4 New Land Law Reforms in Tigray 2006-2010 As a follow-up to the successful land registration and certification, and based on some identified weaknesses in the existing land law (proclamation), a new land law proclamation and regulation was enacted from 2007. Some of the new additions to the law include:. a) Establishment of Land Administration Committees (LAC) at tabia (peasant association=municipality) and kushet(village) level, each with 7 members of which 2 should be women, and 2 should be young landless households. These are in charge of implementing the law at local level. Their work is unpaid community service. b) Households that have migrated from their house/farm and have been away for more than two years should have their land taken and redistributed to young landless households. c) All land rental contracts should be ratified and registered at the tabia LAC. d) Land rental contracts related to traditional land use should not be more than 2 years. e) Land rental contracts related to use of modern technologies can be for up to 10 years. f) Lease contracts between the state and investors can be made up to 50 years. g) Households should not be allowed to rent out more than 50% of their land h) The law also prohibits households from using a rural farm land if their regular monthly income is more than 1000 ETB (mostly targeting government employees and traders) i) Tabia level land courts were established for the handling of land-related disputes. The land courts should consist of three locally elected judges (TNRS, 2006; 2008). This study investigates the status of the land market restrictions drawing on new data collected in 2006-2011 to assess how the new land laws have been implemented with a particular focus on the effects on the land rental market. We combine data from rural households belonging to a household panel that has been used for the earlier studies on the land reform in Tigray region, with data from local land administration committees, and local land conflict mediators. We try to answer the following research questions: 1. What is the underlying rationale of the land market restrictions in Ethiopia? 2. To what extent has the knowledge of the law reached out to the newly established land administration committees, traditional conflict mediators and rural households in general? 3. What are the opinions among the different local stakeholder groups (LACs, local conflict mediators, landlords, tenants and other rural households) about the parts of the land rental restrictions in the law? 4. What is the nature of land markets and land rental contracts in the study area? 5. To what extent are the land law restrictions on land renting violated? 6. To what extent have the recently established community-level land administration committees started to implement the law and what are the perceived impacts? 7. To what extent does the implementation of the land law and administrative reform adjust to customary norms and local variations in opinions about the law?

3. Sampling and data The data comes from three main sources: a) A stratified random sample of 400 households covering 17 communities in Tigray surveyed in 2010. This was the fifth round survey to the same sample of households and communities (with the exception of two of the communities for which this was the second round survey). The stratification was based on market access, population density, access to irrigation, and geographical distribution and was limited to the highland areas of the region where the majority of the population lives and where land registration and certification has been implemented. b) A sample of 400 conflict mediators in close to 100 villages (kushets) in 27 communities (tabias) in 9 districts (woredas) covering the 5 zones in Tigray. This sample therefore has a broader geographical coverage than the household survey. The same conflict mediators were interviewed in 2007, soon after the new law was introduced, and in 2011. c) A sample of Land Administration Committee members in the same communities as the household survey covered. These interviews were carried out in 2010 during the household survey. 4. What is the underlying rationale of the land market restrictions in Ethiopia? The prohibition of land sales appears to come from the high reliance on land as the livelihood basis for the large majority of rural population in the country. This is evident from the established constitutional right to land for all residents in a community that do not have another source of livelihood. Land is seen as an important safety net to ensure access to basic needs while landlessness is seen as a threat to food security. The strong emphasis on equitable land distribution under the Derg regime was also founded on a radical socialist political ideology that saw capitalistic development as a threat to the poor and markets as the instruments for exploitation of the proletariat. Land-to-the-tiller policies therefore had strong grassroots support in many parts of Ethiopia. The introduction of modern agricultural technologies such as tractors in Ethiopia in the late 1960s and early 1970s created fears and political support for the radical reform. Duration restrictions for land rental contracts have some of the same rationale. Long-term rental contracts are closer to sales and could be converted to sales under a land-to-the-tiller umbrella. Such long-term rentals or leases could also lead to absent landlordism and facilitate migration. Prevention of excessive rural-urban migration appears to be the second strong motive for restricting the duration of rental contracts as well as the recent restriction on the share of the farm holding that can be rented out. This has been made explicit in the latest land proclamation from Tigray which states that land should be taken without compensation from households that have been away from their land for two years or more. A further evidence of this is that in relation to the Ethiopian Millennium celebration in September 2007 (the Ethiopian Calendar is about 7 years behind the Gregorian Calendar and New Year is on September 11) beggars in Addis Ababa were given money to go home to their villages and stay there. Excessive rural-urban migration could lead to rapid growth in urban slums, urban poverty and unemployment. Urban employment growth is far behind the rural

population growth that it is increasingly difficult to absorb in rural areas without rural nonagricultural transformation and agricultural development. The recent land law reform is to some extent also a return to the land redistribution policy that was so central under the Derg regime. An important reason for the taking of land from migrant households is to obtain more land to distribute to young landless households. The objective is therefore the same as for the land redistribution policy under the Derg regime, minimize landlessness and provide land to those who want to farm, while those who do not want to farm should not have land. The law also specifies that only landless people should be allowed to inherit land and that persons with a permanent job should now be allowed to have or inherit land. One may question whether the current land tenure policy contributes to strengthen a rural poverty trap that ties people to the land? One may contrast this Ethiopian policy with the policy in Uganda where land sales are allowed. In Uganda this facilitates that people can get a starting capital by selling their land in one place and use the capital to invest in land or other business somewhere else. It also facilitates redistribution of the population from densely populated areas where land prices are high to other rural areas with lower population densities and where land prices are lower. While there are intra-regional resettlement projects also in Ethiopia these may require more government support. 5. Has the knowledge of the new land law reached rural communities? A basic requirement for a land law reform to be successful is that the content of the law is disseminated to the parts of the population that it is relevant for. While this is a necessary requirement for a law to have an impact, it is not a sufficient requirement as people may, even though they know the new law, choose to ignore it or they may even adopt various forms of opportunistic behavior, leading to unintended effects of the law. We assessed the knowledge of the law related to land market restrictions among local conflict mediators, Land Administration Committee members and men and women in a random sample of rural households in Tigray region. Local conflict mediators are elder respected persons in the communities that are asked to mediate internal land-related disputes. They may be seen as carriers of the customary norms but also influential people that would need to learn about new laws and need to adopt them in their practicing of conflict mediation if the law is to be implemented at the grassroots level. Land Administration Committee (LAC) members are representatives from the rural community that have been elected by the community to take responsibility for land administrative issues and are doing the work without payment or compensation. Their knowledge of the law may therefore depend on the amount of training they have received from district-level Land Administrations as well as their own motivation and ability to take on such responsibilities. Table 1 provides information about the knowledge of the law among conflict mediators in 2007 and 2011 and among LAC members in 2010 while Table 2 provides information from the random sample of households. Table 1 shows that conflict mediators knowledge of the law has improved from 2007 to 2011. However, the level of knowledge remains low when it comes to the parts of the law that deal with land rental market restrictions. In particular many of the mediators do not perceive that the legal rules regarding land renting also apply to sharecropping. Since sharecropping is the dominant rental arrangement in the region this implies that most conflict mediators perceive there to be no

restrictions on how much of the land a household can sharecrop out. While the law allows longerterm contracts in relation to use of modern technology by a tenant, the law does not specify what type or how much of modern technology has to be used before a contract qualifies as a longer-term legal contract. It is therefore interesting to note that the question about an 8-year contract, where fertilizer and improved seeds are used, is considered legal by 95% of the mediators in 2011 while only 40% thought so in 2007. The awareness about the new law that states that land can be taken from households that have migrated out of the community for more than two years has also increased substantially from 2007 to 2011. This part of the law has also been given quite a lot of focus in the region and has priority for enforcement according to the regional Land Administration (Abera Hadera, 2010, pers. com.). The land obtained as above is primarily distributed to young landless households. Each of these gets minimum 0.25 ha of land and gets a land certificate for this land. Table 1. Conflict mediators and LAC members knowledge of the law related to land rental market law restrictions Knowledge of the law Are you familiar with the new land proclamation for your region and its content? 1=Yes, 2=Some of it, 0=No, % by code What is the maximum number of years for which households can lease/rent (or sharecrop) their land to others who will use modern technology? -=Don t know (10 years is the correct answer), % correct responses What is the maximum number of years for which households can lease/rent (or sharecrop) their land to others who will use traditional technology? -=Don t know (2 years is the correct answer), % correct response Do the same restrictions apply to sharecropped out land as to rented out land (fixed cash rent)? 1=Yes (correct answer), 2=No restriction on length of contract for sharecropping, 3=Other. % correct responses CM 2007 (n=403) CM 2011 (n=396) LAC 2010 (n=157) 1: 52.9 0: 22.9 1: 31.0 1: 60.5 2: 10.2 2: 29.2 4.9 8.0 1.3 28.2 45.8 30.6 30.1 37.7 46.5 How large share of the farm holding can be rented out maximum? 1=One quarter, 2=Half (correct answer), 3=Three quarter, 4=Depends on family needs for food, 5=All, % by code 2: 19.4 4: 41.2 5: 31.9 2: 32.1 4: 60.7 5: 0.5 2:31.9 4: 36.3 5: 18.5 A female-headed household sharecrops out all her land but will still stay in the village. Is this allowed? 1=Yes, 0=No (correct answer), 2=Don t know, % by code 0: 11.8 1: 74.8 0: 7.1 1: 92.6 A household has a 8-year fixed-rent rental contract with a neighbour who uses fertilizer and improved seed (may be interpreted as modern technology that opens for longer-term contracts) on the land. Is this allowed? 1=Yes, 0=No, 2=Don t know Can a wife deny her husband to rent out family land? 1=Yes (correct answer), 0=No, 2=Don t know Can a 18 year old son deny his father to rent out the family land if he wants to farm on the land and is still living in the household? 1=Yes (correct answer), 0=No, 2=Don t know, 3=Other, specify: Can the land be taken from a household (without compensation) that has been out of the community for 3 years but does not have a permanent job? 1=Yes (correct 1: 39.9 1: 95.1 1: 95.5 1: 77.5 1: 75.2 1: 58.4 1: 44.6 1: 47.8 1:34.9 0:61.0 1:63.9 0:35.6 answer), 0=No, 2=Don t know The knowledge of the law appears to be as good among conflict mediators in 2011 as among LAC members in 2010. The limited knowledge among LAC members is a sign of limited training capacity but also varying emphasis on different elements of the law in the training. 1: 63.1 0: 33.1

In Table 2 we provide an overview of the knowledge of the law among men and women in our rural household survey in 2010. At this point in time, about three-four years after the new law was introduced and Land Administration Committees were established we may expect that knowledge of the law also has reached rural households. The pattern we see is not too different from that of conflict mediators in 2011 and LAC members in 2010. Table 2. Knowledge of the law related to land renting law restrictions among males and females in rural households in Tigray in 2010. Percentage of men and women who responded to the questions in the main sample of households Men Women What is the maximum number of years for which households can lease/rent (or sharecrop) their land to others who will use modern technology? % Correct responses (10 years is correct) What is the maximum number of years for which households can lease/rent (or sharecrop) their land to others who will use traditional technology? % Correct responses (2 years is correct) 2.2 3.8 22.4 Do the same restrictions apply to sharecropped out land as to rented out land (fixed cash rent)? 1=Yes (correct response), 2=No restriction on length of contract for sharecropping, % by code How large share of the farm holding can be rented out maximum? 1=One quarter, 2=Half (correct response), 3=Three quarter, 4=Depends on family needs for food, 5=All, % by code. 1: 54.9 2: 43.2 2: 47.3 4: 31.2 5: 6.7 1: 47.7 2: 45.1 2: 41.2 4: 33.3 5: 9.8 Can your wife/husband deny you to rent out your family land? 1=Yes (correct response), 0=No, 2=Don t know, % correct Does your 18 year old son have the right to deny you to rent out your land if he wants to farm on the land and is still living in your household? 1=Yes (correct response), 0=No, 2=Don t know, % correct 1: 83.9 1: 71.0 1: 65.0 1: 53.6 A household has rented out all its land and has been away for 3 years but has no permanent job: can the land be expropriated and distributed to other households in the community? 1=Yes (correct response), 0=No, 2=Don t know, % by code 0: 60.5 1: 33.2 2: 6.3 0: 54.2 1: 21.9 2: 23.9 6. Opinions among local stakeholder groups about land market law restrictions One may wonder how much grassroots support there is for the land market law restrictions in Ethiopia. The extent of such support may also be important for the efficacy of enforcement of such reforms. Below we provide opinion and perception responses from conflict mediators interviewed in 2007 and 2011 and rural households (men and women separately) in 2010. We should add that 73.5% of the conflict mediators stated in 2011 that they were party members of the ruling party. 26.1% of the LAC members stated that they were party members. Table 3 provides the responses from the conflict mediators and LAC members.

We see from Table 3 that the prohibition of land sales has a very strong support but there appears to have been a reduction in support for prohibition of mortgaging of land among conflict mediators. The new regulation allows investors to use their land lease contracts as collateral. We see a noticeable increase in the restriction that only half of the farm area can be rented out. However, this needs to be seen in connection with the next question where more than half of the conflict mediators do not consider this restriction to apply to sharecropping arrangements, the dominant form of land renting in Tigray. A majority (57%) considered it legal to sharecrop out all the land in 2007 and 72% were of the opinion that this should be allowed. By 2011 there is a reduction to 50% who consider it legal and to 53% who are of the opinion that it should be allowed. When we asked the more specific question whether female-headed households, orphan households and other poor households should be allowed to sharecrop out all their land when they lack resources to cultivate it themselves, the support for this was as high as 91% in 2007 and even higher, 95%, in 2011, among conflict mediators. Arguments in support of this were that this is important to secure their livelihood, they cannot farm the land efficiently themselves, and the land can be made available for more efficient land users. Table 3. Conflict mediators and LAC members opinions and perceptions about land market restrictions in 2007 and 2011. Questions CM 2007 (n=403) CM 2011 (n=396) LAC 2010 (n=157) Do you think that selling of land should be illegal? 1=Yes, 0=No 1: 94.0 1: 98.2 1: 82.2 Do you think that mortgaging of land should be illegal? 1=Yes, 0=No 1: 86.8 1: 42.3 1: 70.7 Do you agree that only half of the farm holding should be allowed rented out? 1: 50.8 1: 77.3 1: 70.1 1=Yes, 0=No, % by code Do you perceive sharecropping as land renting and to be subject to the same regulations as land renting? 1=Yes, 0=No, sharecropping is not land renting, % by code Do you consider it legal for a household to sharecrop out all its land? 1=Yes, 0=No, 2=Don t know, % by code Do you think that households should be allowed to sharecrop out all their land? 1=Yes, 0=No, % by code Do you think that female-headed households, orphan households and other poor households should be allowed to sharecrop out all their land when they lack resources to cultivate it themselves? 1=Yes, 0=No, % by code If yes, why? 1=It secures their livelihood, 2=They cannot use the land efficiently themselves, 3=The land can be made available for more productive farmers, 4=Other, specify, % by code If no, why? 1=They should farm it themselves, 2=They should follow the law, 3=They should give away the land to others if they fail to farm it, 4=Other. % by code Do you agree that land should be taken from households that have been out of the village for more than 2 years but do not have a permanent job? 1=Yes, 0=No, 2=Don t know Do you think that the landless in the community should be given more of the land through redistribution? 1=Yes, 0=No Do you think there will be any new land redistributions in your tabia within the next ten years? 1=Yes, 0=No * Multiple responses allowed in 2011. 0: 52.3 1: 46.5 0: 29.2 1: 56.9 0: 27.4 1: 72.3 0: 8.7 1: 91.1 1: 52.3 2: 39.4 3: 5.1 1: 4.3 2: 3.0 1: 36.5 0: 61.5 1: 67.5 0: 32.5 1: 43.4 0: 55.4 0: 55.9 1: 44.1 0: 47.6 1: 49.6 0: 46.8 1: 52.9 0: 3.1 1: 95.4 1: 58.0 2: 52.5 3: 18.0* 1: 2.3 3: 2.3 1: 41.9 0: 58.1 1:35.0 0: 65.0 1: 19.9 0: 80.1 0: 42.0 1: 53.5 0: 40.8 1: 52.3 0: 43.3 1: 52.3 0: 8.3 1: 87.9 1: 42.7 2: 49.7 3: 36.9 1: 5.1 2: 2.6 3: 2.6 1: 73.3 0: 20.4 1: 79.0 0: 17.8 1: 43.3 0: 51.6 We also see that a majority are against that land should be taken from households that have been away for more than two years both in 2007 and 2011. At the same time we see a change in the

attitude about providing land to the landless. While 67.5% supported to provide land to the landless in 2007, this has reduced to 35% in 2011. The responses from LAC members in Table 3 contrast substantially with the conflict mediators responses in some cases. It is notable that a majority of LAC members (73%) support taking of land from migrant households while a majority of the conflict mediators was against this. Similarly, a majority of LAC members (79%) think that more land should be given to the landless through redistribution. A higher share of them, 43%, also believe that there will be a new land redistribution in the community within the next ten years. This may be because they feel more responsible for implementing the current law and have received training to do so and because taking land from migrants has been a priority area for law enforcement by the regional and district land administrations. Although we do not know what the opinions of these same LAC members were before they became LAC members, it is probable that these were closer to that of the conflict mediators and ordinary rural households at that time. On the other hand, their opinions about restrictions on renting out of land were not very different from that of conflict mediators and ordinary households. A large majority of LAC members (88%) perceive that female-headed, orphan and other resource-poor households should be allowed to rent out all their land and this is almost as high as among conflict mediators and ordinary households (see Table 4 ). The responses from our stratified random sample of rural households male and female representatives are presented in Table 4. For some of the variables we only have the response from the head of the household (that usually is male). Table 4 shows a remarkably similar pattern as that of conflict mediators. 64% of the men and 74% of the women considered it legal for a household to sharecrop out all its land and 68% of the men and 74% of the women were of the opinion that this should be legal. When asked whether femaleheaded households, orphan households and other poor households should be allowed to sharecrop out all their land when they lack resources to cultivate it themselves, 94% of the men and 90% of the women agreed. There is therefore no doubt that there is strong disagreement with the recent land law restriction that only 50% of the farm can be rented or sharecropped out. However, there is some support for this restriction in relation to fixed-rent contracts. Sharecropping contracts may require more presence to monitor the work of the tenant and participation in the harvesting of the crop. There is more opposition towards taking land from migrant households and more support for giving land to the landless but obviously not by taking land from migrating households. There is also a higher uncertainty in this sample whether there will be another land redistribution the next ten years than among the conflict mediators.

Table 4. Perceptions and opinions about land renting among males and females in rural households in 2010: Percentage of those that responded to each question Questions Males Females Do you think that selling of land should be illegal? 1=Yes, 0=No 1: 94.0 Do you think that mortgaging of land should be illegal? 1=Yes, 0=No 1: 86.8 Do you agree that only half of the farm holding should be allowed rented out? 0: 40.9 1=Yes, 0=No, 2=Don t know, % by code 1: 54.4 2: 4.8 Do you perceive sharecropping as land renting and to be subject to the same regulations as land renting? 1=Yes, 0=No, sharecropping is not land renting, % by code 0: 33.0 1: 66.3 0: 39.1 1: 47.4 2: 13.5 0: 34.3 1: 59.6 Do you consider it legal for a household to sharecrop out all its land? 1=Yes, 0=No, 2=Don t know, % by code Do you think that households should be allowed to sharecrop out all their land? 1=Yes, 0=No, % by code.do you think that female-headed households, orphan households and other poor households should be allowed to sharecrop out all their land when they lack resources to cultivate it themselves? 1=Yes, 0=No, % by code If yes, why? 1=It secures their livelihood, 2=They cannot use the land efficiently themselves, 3=The land can be made available for more productive farmers, % by code 0: 32.7 1: 64.0 2: 3.4 0: 32.3 1: 67.7 0: 6,1 1: 93.9 1: 33.4 2: 36.9 3: 27.0 0: 19.7 1: 74.2 2: 6.1 0: 22.6 1: 74.0 0: 5.4 1: 89.9 1: 37.5 2: 41.3 3: 21.1 Do you agree that land should be taken from households that have been out of the village for more than 2 years but do not have a parmenent job? 1=Yes, 0=No, 2=Don t know, % by code Do you think that the landless in the community should be given more of the land through redistribution? 1=Yes, 0=No, 2=Don t know, % by code Do you think there will be any new land redistributions in your tabia within the next ten years? 1=Yes, 0=No, 2=Don t know, % by code 1: 19.8 0: 73.0 1: 69.3 0:22.0 1: 21.4 0: 48.1 2: 30.5 7. The nature of the land rental market and land rental contract preferences It is important to view the findings above in relation to the nature of the land rental market in the study area and the preferences of the agents operating in these markets. We therefore provide a brief review of earlier studies that provide such insights in combination with new data regarding contract choices and contract preferences of landlord and tenant agents interviewed in 2006 and 2010 in relation to our rural household survey in Tigray. Ghebru and Holden (2008) use the same household panel data as used in this study to assess the rationale of land rental markets in Tigray. They identified a typical reverse-tenancy system with poor landlords and less poor tenants. Female-headed households are the majority among landlord households while male-headed households dominated on the tenant side of the market. No tenant households were landless, they used the land rental market to gain access to additional land.

Sharecropping was the dominant contract and the market was characterized by rationing from the supply side. Transaction costs caused entry barriers for potential tenants. Sharecropping is an important way for resource-poor female-headed households to access food by contracting tenants with surplus of non-land resources to cultivate their land. Holden et al. (2011) showed that land certification has enhanced tenure security and the efficiency of the land rental market in Tigray Region. Deininger et al. (2011) has shown a similar effect in Amhara Region. Particularly female-headed households had expanded their renting out activity following land certification in Tigray. Holden and Ghebru (2011) found that this has enhanced household welfare, particularly of female-headed households. Gebrehaweria and Holden (2011) used a different data set from Tigray to assess land rental contract choices in six locations with access to irrigation land. They assessed how rainfall shocks affected the extent of land renting and contract choice. They found evidence that the land rental market is used as a coping response to such shocks. Such shocks were found to increase the probability that households chose fixed-rent contracts rather than sharecropping contracts in order to obtain immediate cash to satisfy their urgent needs. However, they also found that fixed-rent contracts were more likely to be used when the ex ante production risks were lower, such as on irrigated land, while sharecropping was preferred when ex ante risks were high. Access to irrigation is still very limited in this part of Ethiopia where rainfall risks are high and this may be the main rationale for the continued strong dominance of sharecropping contracts over fixed-rent contracts. We provide further evidence on this from our data below. Table 5 provides information about the land rental contracts of landlord and tenant households in 2006 and 2010. The 2006 sample includes responses from rental contract partners of our main sample households in addition to land rental market participants in our main sample. The sample consists of close to 500 land rental market agents in 2006 and about 260 such agents in 2010. The majority preferred oral rental contracts among the partners only but there is an increase in the demand for written contracts that are reported to the tabia. Written contracts are preferred in relation to longer-term contracts, fixed-rent contracts, and contracts with strangers. Many of the landlords and tenants perceive that land registration and certification has enhanced the willingness to participate in the rental market and explain this by higher tenure security and easier access to land to rent in. Only about four percent of the landlords and tenants have experienced conflicts related to their contracts. Still, surprisingly many seem to prefer to have written contracts reported to the tabia in 2010. Table 6 provides additional information from tenant households only. We see that sharecropping is the dominant contract type and also the most preferred contract.sharing of risk is one of the most important reasons for this. Contracts longer than two years (the maximum length when traditional technology is used, according to the law) are common and open-ended contracts are also common. Many prefer longer-term contracts and the main reason for that is that they can invest more on the land. Many would have liked to rent in additional land, demonstrating that sharecropping leads to rationing on the tenant side due to the lack of a market-clearing price. Table 7 cross-tabulates contract choice preferences with stated advantages of contracts. One may be puzzled by the many tenant respondents stating that one of the advantages of sharecropping is that it gives more food after harvest. A possible explanation may be that the reference point may not be a fixed-rent

contract but an input sharing contract where the tenant gets a lower share than in a pure output sharing contract. Table 5. Landlords and tenants in 2006 and 2010: Number of responses Question 2006 (n=495) Which type of land rental contract do you prefer? 1=Oral contracts among partners only, 1: 432 2=Oral with witnesses, 3=Written contract, not reported, 4=Written and reported to tabia 2: 25 leaders, 3: 7 If more than one type is preferred, explain when and why. 1=Long-term contracts prefer ed to be written. 2=Fixed-rent contract preferred to be written, 3=Sharecropping contracts preferred to be oral, 4=Prefer oral contracts with relatives, 5=Prefer written contracts with strangers Has land registration and certification had any impact on whether you participate in the land rental market (including sharecropping)? Yes=1, No=0 If yes, are you more or less willing to rent in or out your land after you received the certificate? 1=More willing/able, 2=No difference, 3=Less willing/able If yes, why are you more or less willing? 1=Feel more tenure secure, 2=Easier to rent in land, 3=More difficult to rent in land, 4=Other, specify: Has receiving a land certificate affected the type of land contract you prefer to use? 1=No, 2=Yes, prefer fixed-rent more, 3=Prefer longer-term contract after I received certificate Have you had any land disputes in relation to some of your land contracts? Yes=1, No=0 Do you prefer to have written sharecropping and rental contracts that are registered at the tabia/land Administration Committee? 1=Yes, 0=No, 2=Only for contract partners that I do not know well 4: 28 1: 39 2: 15 3: 180 4: 59 5: 19 0: 267 1: 228 1: 213 2: 1 3: 16 1: 81 2: 131 3: 3 1: 378 2: 78 3: 27 0: 474 1: 20 2010 (n=280) 1: 131 2: 60 3: 2 4: 88 1:13 2: 31 3: 74 4: 2 5: 12 0:184 1: 79 1: 66 2: 12 3: 2 1: 59 2: 13 3: 2 1: 177 2: 20 3: 15 0: 252 1: 10 0: 79 1: 142 2: 21

Table 6. Tenants: Number of responses Question 2006 Which land rental arrangement do you currently apply? 1=Sharecropping, 2=Sharecropping with advance payment, 3=Fixed-rental contact, 4=Input/cost-sharing contract: Landlord pays cash inputs, 5=Cost-sharing where landlord advance input costs, 6=Cost-sharing with equal sharing of cash inputs, 7=Cost-sharing where tenant advances input costs, 8=Other, specify: Which land rental arrangement do you prefer? 1=Sharecropping, 2=Sharecropping with advance payment, 3=Fixed-rental contact, 4=Input/cost-sharing contract What is the advantage of the type of contract you are preferring? 1=It reduces risk (risk sharing), 2=It enables me to share input costs, 3=It gives me incentive to produce more, 4=It is the only available contract type, 5=It gives me more food after harvest, 6=I do not have to pay cash in advance, 7=I can ensure optimal input use and yield on the land Do you have any renting/sharecropping contracts that are for more than one year? Yes = 1 No =0 Duration of contracts: 0=Less than a year, 1=1 year, 2=2 years, 3=3 years, 4=4 years. 5=5 years, 6.>5 years, specify: years, 7=Open-ended (n=265) 1: 209 2: 13 3: 10 4: 15 5: 9 1: 190 2: 16 3: 14 4: 31 1: 69 2: 7 3: 41 4: 53 5: 66 6: 5 7: 15 0: 37 1: 227 1: 21 2: 22 3: 12 4: 9 5: 8 6: 14 7: 12 Do you prefer contracts that last for more than one year? Yes = 1 No = 0 0: 33 1: 228 If yes, why do you prefer longer-term contracts? 1=I can invest more in the land, 2=I can apply 1: 175 more inputs, 3=I do not have to search for other partners so often, 4=Other, specify: 2: 24 3: 31 If no, why do you not prefer longer-term contracts? 1=Only need to rent for one year, 2=Do not know whether I want to rent another year 1: 28 2: 9 If yes, what do you do to obtain longer-term contracts? 1=Work hard on rented land to get 1: 180 contract renewal, 2=Negotiate long-term contracts from the beginning, 3=Select landlords that 2: 34 are willing to give long-term contracts, 4=Identify particularly poor landlords that have weak 3: 13 bargaining power, 5=Offer fixed up-front payment 4: 3 Would you like (be able) to rent in some more land? Yes = 1 No = 0 0: 105 1: 159 If your answer is yes, How much more? Median: 0.75 ha Have you attempted to rent in the additional land you wanted over the last 2 years? 1=Yes, 0=No 0: 82 1: 69 2010 (n=111) 1: 97 2: 8 3: 6 1: 97 2: 8 3: 6 1: 45 2: 4 3: 6 4: 13 5: 20 6: 21 7: 1 0: 26 1: 85 1: 30 2: 22 3: 12 4: 8 5: 3 6: 7 7: 25 0: 10 1: 101 1: 53 2: 10 3: 28 4: 8 1: 4 2: 10 1: 84 2: 7 3: 4 0: 52 1: 59 Median: 0.75 ha 0: 62 1: 39

Table 7. Land rental contract choice versus tenants stated advantages of contracts in 2006 Contract advantage Sharecropping Sharecropping with advance payment Preferred contract Fixedrensharing Input/ cost- contract contract Total It reduces risk (risk sharing) 59 4 0 5 68 It enables me to share input costs 0 0 0 7 7 It gives me incentive to produce more 11 6 6 15 38 It is the only available contract type 42 1 2 1 46 It gives me more food after harvest 53 3 4 3 63 I do not have to pay cash in advance 4 1 0 0 5 I can ensure optimal input use and yield 14 1 0 0 15 Total 183 16 12 31 242 Table 8 provides additional information from landlord households. Their stated most important reasons for renting out land were shortage of oxen, labor and personal illness. Most landlords also prefer sharecropping contracts and important reasons for this were sharing of risk and food security by sharing of the food output. Relatively few landlords perceived there to be a problem with shirking by tenants. Shirking is apparently controlled primarily by selecting good tenants in the first place, that are trustworthy, are good farmers, and/or blood relatives, and in addition monitoring is used and possibly eviction if performance is poor. 8. To what extent are the land rental law restrictions enforced or violated? We assess the distribution of the share of land rented out drawing on the landlord part of our household panel data from Tigray. We measure the share of land rented out by male and female landlord households separately. We assess this share in 1997 (before land registration and certification) was implemented and in 2006, at the time the new restriction that only half of the farm can be rented out. The distributions are shown in Figure 1 for female landlords and in Figure 2 for male landlords. Earlier studies (Holden et al., 2011; Holden and Ghebru, 2011) have shown that particularly female-headed households have benefitted from the land registration and certification as it made them more tenure secure and more willing to rent out their land and this has enhanced their welfare. Female-headed households typically are poor in non-land resources such as labor and oxen (needed for ploughing) and therefore prefer to rent out their land. This is illustrated in Figure 1 where we see a sharp increase in the area rented out by female landlords from 1997 to 2006. While there has been an increase also for male landlords this effect is much smaller than that of female landlords as can be seen from Figure 2. Figure 1 demonstrates that a large share of female landlords violate the restriction that only 50% of the land can be rented out. If the renting out restriction has been enforced between 2006 and 2010 we should see a decline in area share rented out from 2006 to 2010, especially for female-headed households. Figures 3a, 3b, and 4a, 4b, compare the areas rented out in 2006 and 2010 for male and female landlords separately. Figures 3a and 4a compare the distributions of rented out shares of land for all male and female landlords in the two years while Figures 3b and 4b compare the distributions for the male and

Table 8. Landlord data, number of responses Questions 2006 Why do you rent out land? 1=Shortage of labour, 2=Shortage of oxen, 3=Personal problem (illness, aged, etc.), 4=Poor/no access to credit, 5=Seed/Fertilizer problem, 6=Off-farm job, 7=Other, Specify: What type of land contract do you prefer? 1=Sharecropping, 2=Sharecropping with advance payment, 3=Fixed-rent contract, 4=Input/cost-sharing contract, 5=Advance input costs yourself, 6=Let tenant advance input costs, 7=Pay input costs yourself without refunding, 8=Other, specify: Why do you prefer this type of contract? 1=Share risk with tenant, 2=It enables me to share input costs, 3=It provides me with cash, 4=The only type of contract that I am offered, 5=It provides food after harvest (food security), 6=Other, specify: Do you think that the tenant shirks (deliberately avoid to work hard) in sharecropping? 0=No, 1=Yes, 2=Some tenants do, 3=If I do not monitor them, 4=If I do not use threat of eviction, 5=Other, specify: If yes, what mechanisms are you using to motivate the tenant to work hard? 1=Eviction when performance is poor, 2=Increase the share to the tenant, 3=Increase intensity of monitoring and supervision, 4=Provide inputs for production, 5=Nothing, 6=Threat of eviction, 7=Other, specify What criteria do you use to select your tenant? 1=Trustworthy, 2=Good reputation as farmer, 3=In-laws claim the tenancy, 4=Blood-related relatives, 5=Good neighbors, 6=The one that offers better contract, 7=The first who comes and asks to rent in (n=240) 1: 48 2: 77 3: 39 4: 2 5: 1 6: 1 7: 26 1: 196 2: 14 3: 12 4: 5 5: 6 1: 63 2: 7 3: 11 4: 50 5: 111 0: 175 1: 29 2: 32 4: 3 1: 3 2: 6 3: 15 4: 5 5: 10 6: 3 How much land have you rented/sharecropped out? Median: 0.75 ha Would you like (be able) to rent out some more land? 1=Yes, 0=No 0: 220 1: 10 If you have a land certificate, does having the certificate give you any advantages in 0: 41 relation to your contracts with your tenants? 1=Yes, 0=No 1: 180 If yes, how? 1=Improved bargaining power, 2=Contract fulfillment, 3=More tenure security, 4=More long-term contracts, 5=Better performance by tenant, 6=Tenant takes more responsibility for land conservation, 7=Other, specify: Does having a certificate make you more willing to rent out the land to strangers? 1=Yes, 0=No If you do not have a land certificate, what are the disadvantages, if any, in relation to land renting out that you perceive? 1=No disadvantages, 2=Fear land grabbing by tenant, 3=More land disputes with tenant, 4=Harder to enforce tenant to work hard, 5=Less bargaining power in relation to contract choice, 6=Other, specify If you face such difficulties, how do you respond? 1=Rent out less land, 2=Use oneyear contracts only (without renewal of contracts), 3=Use one-year renewable contracts only, 4=Rent out to relatives only, 5=Rent out only to tenants that you trust. 1: 47 2: 22 3: 59 4: 10 5: 39 6: 2 0: 134 1: 86 1: 38 2: 60 3: 86 4: 6 5: 5 1: 10 2: 22 4: 127 5: 51 2010 (n=135) 1: 81 2: 65 3: 30 4: 3 5: 4 6: 2 7: 5 1: 98 2: 31 3: 5 1: 53 2: 6 3: 19 4: 32 5: 29 0: 76 1: 34 2: 20 3: 4 1: 15 2: 17 3: 14 4: 2 5: 6 6: 2 1: 31 2: 34 3: 2 4: 65 5: 7 7: 8 Median: 0.5 ha 0: 129 1: 3 0: 43 1: 76 1: 37 2: 10 3: 29 4: 2 5: 4 6: 9 0: 78 1: 43 1: 46 2: 44 3: 21 4: 2 5: 11 1: 24 2: 13 3: 6 4: 40 5: 31