CASE STUDY Shortcomings to overlapping land rights and a way forward: the case of Mailo land in Central Uganda DANIEL ALI AND MARGUERITE DUPONCHEL INTRODUCTION Uganda has a complex land tenure system. Four tenure systems officially recognized by the 1995 Constitution coexist: freehold, Mailo, leasehold, and customary tenure. While undocumented customary tenure dominates in the country (UNPS 2013/2014 data estimates that 84% of parcels are held customarily), Mailo tenure is predominant in the central and parts of the western regions. Around 10% of Uganda’s land (in surface) is held under Mailo tenure, possibly more1. MAILO LAND: HISTORICAL OVERVIEW The Mailo system is rooted in the 1900 Buganda Agreement, in which the British returned appropriated land to the Buganda Kingdom. The king and his nobles were awarded large blocks of land, occupied by smallholders, known as Mailo land, in reference to the square mile unit in which they were measured (West 1972). In practice, land under Mailo tenure continued to be governed by Buganda law and custom, including being transmitted exclusively to male heirs. The Figure 1: woman working her land Mailo system introduced private property for customary authorities, but ignored the rights of Act. Under that law, bona fide lawful occupants overlapping land rights are substantial. This the occupants (tenants or kibanjas holders) who have the right to assign, sublet, pledge, sub-divide, underlines that failure to clarify land rights on were actually cultivating the land, as very few bequeath, or create third party rights in the land. The Mailo land can be a significant impediment to rural chiefs engaged directly in farming. Therefore, 2010 amendment to the Land Act further clarified development in Central and Western Uganda. the agreement was amended in 1928, introducing that tenants can only be evicted following a court busuulu (ground rent) and envujju (tribute) laws, order and only for non-payment of busuulu provided DATA that aimed to recognize and strengthen the rights they had occupied the land unchallenged for more of occupants by putting a limit on the rent to be than 12 years before the 1995 Constitution. In paid and protecting tenants against eviction To address the issue of overlapping land rights in a addition, tenants can acquire certificates of without compensation. way that increases land-related investment and occupancy (CoOs) with payments of UGX 5,000, boosts agricultural productivity of smallholder provided they obtain the consent of the landlords Following Amin Dada’s coup d’état, the Land and prove there are no arrears in the payment of farmers, the German Corporation for International Reform Decree (1975) declared all land public. Cooperation (GIZ, for its German initials), with ground rent. In practice, however, no CoO has been The Mailo system disappeared on paper, and funding from the European Union and the German issued to date, possibly because a CoO translates tenants became tolerated occupants with no Government and in partnership with Uganda’s into a written encumbrance on the landlords’ Mailo tenure/occupancy rights, since there was no Ministry of Lands, Housing and Urban title. mention of them in the law. The 1975 decree Development, have developed a pilot program was, however, not implemented in practice. In “Improvement of Land Governance in Uganda to The overlap of ownership and use rights on Mailo 1986, after Museveni’s government assumed land has been identified as a key impediment to Increase Productivity of Small-Scale Farmers on power, a reform of the country’s land laws was Mailo-Land (ILGU)� to test low cost approaches investment and improved productive efficiency in engaged. The result was the re-institution of the for systematic demarcation of Mailo land occupied Uganda’s agricultural economy (Deininger and Ali, Mailo system in 1995, at the insistence of the by tenants. The pilot consists of land inventories in 2008). Because the productivity reductions are Buganda Kingdom. The rights of tenants, eight sub-counties in Mityana and Mubende significant, and the amount of land under Mailo type however, were only clarified by the 1998 Land districts. The World Bank, in partnership with the tenure is substantial, the overall losses from SHORTCOMINGS OF OVERLAPPING LAND RIGHTS AND WAY FORWARD Uganda Bureau of Statistics, will measure the Table 1: Household characteristics owned and kibanja parcels). Intercropping is a impacts of this pilot project at the household Female head 0.27 common practice, and the tendency is to mix level. To achieve that, baseline data for the Age head 45.59 annual crops with perennials more often on owned impact evaluation was collected from July to Head has no formal education 0.88 November 2017 with a random sample of 2485 parcels compared to kibanjas and short-term rentals. The incidence of hired labor is significantly Any wage employment in the past tenant households, cultivating 3590 parcels, in 12 months 0.19 191 enumeration areas.2 In addition, 64 resident higher in owned and short-term-rental parcels, but No. Of days (8hours) worked for and non-resident Mailo landlords were differences are not observed in terms of family wage in the 12 months 46.25 interviewed during the data collection exercise. labor and value of crop output per unit of land Number of ag parcels (own/used) 1.60 This policy note uses data from that survey. across parcels by ownership types. Total size of ag land (own/used) 3.63 Number of hh members 4.94 KNOWLEDGE OF LANDLORDS AND Number of adult male 1.07 Out of the 2485 tenant households interviewed, BUSUULU PAYMENT IS NOT A GIVEN Number of adult female 1.14 just over a quarter are female headed and 7% are in a polygamous marriage (Table 1). Household Homestead parcel under kibanja While kibanja holders have been using their land ownership rights 0.96 heads are on average 46 years old, with average on average for 14 years, only half of the tenants No document on homestead family size of 5 members. Most household heads know their landlord, among which 20% report the property 0.91 (88%) have no formal education and are mainly landlord being a relative (Table 3). Only a tenth of Semiformal/informal document on involved in subsistence agriculture. Only 19% of property 0.47 the landlords reported by kibanja holders are households report having at least one member Experience of shocks in the last 12 women. Despite the amount being modest (about working for a wage, reporting, on average, about months: UGX 20,000), payment of busuulu is found to be 1.5 month of paid work over the past 12 months. Drought 0.39 low. Sampled households reported that they have Climatic and agricultural price shocks are quite Irregular rains 0.21 ever paid the required ground rent for only 26% of common in the study areas. Among sampled Erosion 0.08 the parcels, and apparently, receipts were issued for households, 39% reported that they have faced High level crop pests 0.47 96% of the cases. The reported reasons for not droughts, 21% irregular rains, 47% high level of Livestock disease 0.10 paying busuulu include: not knowing the landlord crop pests (possibly largely army worm), 13% High costs of ag. Inputs 0.13 (36%), the landlord never asked (23%), not feeling high costs of agricultural inputs and 19% low Low prices of ag. Outputs 0.19 like they owed the landlord anything (18%), not prices of outputs over the past 12 months No. Of obs. 2485 preceding the survey. Almost all the homestead parcels are under Mailo tenancy (96%). Informality is high, with 91% of the households Table 2: Parcels characteristics and inputs by type of rights reporting that they do not have formal Other use Total Ownership Kibanja documentation for their homestead property. right Parcel size, acres 2.22 7.67 2.13 *** 1.76 *** KIBANJAS AND SHORT-TERM RENTALS Parcel cultivated, % 83.93 68.62 83.02 *** 90.56 *** ARE CHARACTERIZED BY LOWER LEVEL Homestead 0.47 0.71 0.53 *** 0.20 *** OF INVESTMENT AND INPUT USE Primary use: own cultivated (annual crops) 0.76 0.69 0.73 0.92 *** Primary use: own cultivated (perennial Table 2 provides parcel characteristics by type of crops) 0.24 0.31 0.27 0.08 *** ownership rights. Kibanjas constitute 82% of Rain fed 0.98 0.99 0.99 0.97 ** agricultural land owned/used by the interviewed Any erosion problem in last season 0.12 0.15 0.11 0.16 *** households. While the size of agricultural Erosion control facility: parcels is on average 2.22 acres, kibanjas are None 0.65 0.55 0.64 * 0.73 *** significantly smaller than owned parcels (2.13 Terrace 0.20 0.30 0.20 ** 0.15 *** vs. 7.67 acres), but larger than rented-in parcels Trees 0.04 0.06 0.04 0.03 (1.76 acres). Although there are very few owned Drainage ditches 0.11 0.07 0.12 0.08 *** parcels in the sample, households that do own Plot burnt in preparation for planting 0.14 0.08 0.14 * 0.17 ** land tend to leave it uncultivated (31%), Use of org. Fertilizer 0.25 0.46 0.27 *** 0.15 *** compared to households under kibanja (27%) Use of chem. Fertilizer 0.18 0.18 0.17 0.24 *** and short-term rental arrangements (9%). Use of pesticides 0.40 0.53 0.37 *** 0.50 *** Similarly, households allocate relatively more Hired labor for land preparation 0.31 0.45 0.29 *** 0.36 *** land to perennial crops on their owned parcels Hired labor for input application 0.10 0.23 0.09 *** 0.14 *** (31% as primary use) than kibanjas (27%) and Hired labor for wedding and pruning 0.28 0.40 0.27 *** 0.32 ** short-term rentals (8%). These observations Hired labor for harvesting 0.23 0.32 0.21 *** 0.31 *** have implications for the relationships among Total family labor days, per acre 119 112 123 100 land rights, land-related investment and intensity Mixed stand 0.52 0.63 0.53 * 0.45 *** of land utilization. In fact, the incidence of Total crop area, in acres: erosion control facilities is significantly higher Parcel planted with: Dry maize 0.59 0.72 0.55 * 0.75 *** on owned parcels (45%) than kibanja (36%) and Dry beans 0.26 0.44 0.24 *** 0.30 *** short-term rental (27%) parcels. Similar patterns Sweat potatoes 0.04 0.15 0.04 *** 0.03 are also observed in land management and input Green cassava 0.03 0.05 0.03 0.03 use: use of organic fertilizer is significantly Bananas 0.15 0.45 0.16 *** 0.05 *** higher in owned parcels than those under Green coffee 0.05 0.07 0.05 0.03 ** tenancy arrangements (46% vs. 27% for kibanja Dry coffee 0.18 0.65 0.19 *** 0.07 *** and 15% for short-term rentals). In contrast, use Value of total harvest per acre, UGX* 904130 732486 908052 914996 of chemical fertilizer is relatively higher in short-term rentals (24% vs. 18% and 17% in No. of obs. 3590 102 2859 625 *value trimmed of the 1st and last percentile. note: t-test performed from one categories against the others. DANIEL ALI AND MARGUERITE DUPONCHEL Table 3: Descriptive statistics on kibanja parcels Table 4: Relationship to tenants by gender of No. of years used 14 landlord and residence status Know the landlord 0.50 Landlord cultivates 0.83 Landlord is a relative 0.20 Perc. cultivated by landlord 27 Female landlord 0.11 No. of kibanja tenants on the parcels 71 Ever paid the busuulu 0.26 Amount paid (last busuulu), UGX 19333 Perc. of known kibanja holders 0.63 Busuulu payment receipt issued 0.96 Has collected busuulu 0.70 Reason for not paying busuulu: Amount of busuulu collected in past Don't know the process 0.08 year, UGX 324364 Too expensive 0.03 Amount of busuulu entitled in past Landlord didn’t ask 0.23 year, UGX 1053571 Landlord refused 0.05 Reason of not collecting busuulu: Don't know the landlord 0.36 Tenants refused 0.62 Don't owe the landlord anything 0.18 Tenants unknown 0.08 Figure 2: mailo tenant farmer Others 0.06 No. obs. 2859 Don’t want to collect 0.30 * data was trimmed dropping the 1st and last No. of obs. 63 percentiles Table 5: Conflict history and perceptions of risks formal way may thus go a long way in reducing Panel 1: kibanja parcels knowing the process (6%) and the landlord incidence of disputes as well as improving perceived refused (5%). Given that payment of ground rent Conflict in the last 12 months 0.09 tenure security, an important prerequisite to investing Reason for the conflict: is required as proof of lawful/bona fide tenancy, on the land. the current status quo will stand against tenants in Boundary 0.37 Inheritance 0.10 any court of law. BINDING CREDIT CONSTRAINTS RESTRICT Sales 0.12 PRODUCTIVITY IMPROVEMENT AND LAND Expropriation 0.04 These patterns are also illustrated in the landlords’ RELATED INVESTMENTS Other ownership 0.25 responses (Table 4).3 While the interviewed Rental 0.03 landlords are not a random sample, only 63% of Tenant households have low level of access to credit: Payment of busuulu 0.07 them report that they know their tenants. About only 11%, 12% and 2% applied for a loan from Other 0.02 70% have collected busuulu in the past from some informal, semi-formal and formal lending institutions, Conflict was with: of their tenants (on average, interviewed landlords respectively, in the past 12 months preceding the date Neighbor 0.35 have 71 tenants each), yet with a substantial gap of the survey.5 As can be seen in Table 6, most of the Landlord 0.35 between entitlement (estimated to be UGX 1M per loans were taken to purchase land and agricultural Family 0.18 annum) and collection rate (estimated to be UGX inputs (36%, 41% and 62% from informal, semi- Village elders 0.04 300,000). Interestingly, when asked why they did formal and formal sectors, respectively), followed by Local government 0.00 not collect ground rents, 62% of the interviewed starting and running off-farm businesses (15%, 18% Central government 0.02 landlords responded that tenants refused, and 30% and 20% respectively from the three sources). Private company 0.01 of them did not want to collect ground rents.4 Other 0.05 Borrowing from informal sources is also used to cover Conflict was resolved 0.38 educational (16%), health (20%), and ceremonial, e.g., HIGH LEVELS OF CONFLICTS AND No likelihood of conflicts on user rights 0.77 wedding and funeral (20%) related expenses. About INSECURITY CHARACTERIZE MAILO No likelihood of losing land if left fallow 0.77 22% of educational expenses are covered through No likelihood of losing land if landlord TENANCY loans from semi-formal institutions. Collateral is were to sale 0.59 required particularly to borrow from semi-formal No. obs. 2589 Kibanja holders report land-related disputes on (74%) and formal (90%) institutions. Given that 62% about 9% of parcels under Mailo tenancy (Table of the loans were secured using land as collateral, land Panel 2: landlords 5, panel 1). The types of conflicts reported range appears to be the most important asset required by the Any dispute, d 0.30 from boundary (37%), other ownership (25%), Null likelihood of dispute in the future, d 0.57 formal sector. But only about a third of the loans in the sales (12%) and inheritance related (10%). No. of obs. 63 semi-formal sector required land as collateral. The Despite its low level, payment of busuulu was the reported figures in Table 6 also show that tenant main reason for conflict in only 7% of the cases. households face high levels of credit constraint across Conflicts arose mainly with neighbors (35%), all the three types of credit.6 In the semi-formal sector, WHAT ARE THE OPTIONS TO REDUCE THE landlords (35%) and relatives and family households report being constrained by risk (69%), INCIDENCE OF OVERLAPPING RIGHTS? members (18%). Given that Mailo tenancy is transaction costs (60%) and quantity of credit characterized by overlapping land rights, the available (28%), while in the formal sectors The reported willingness to pay for obtaining a reported level of perceived tenure insecurity is as constraints are associated with transactions costs legal certificate to secure the occupation of expected. Kibanja holders expect potential (37%), risk (31%) and quantity (11%).7 Although kibanjas is high, with an average of about UGX disputes or loss of their land (if left fallow) in the obtaining formal land documentation that could be 400,000 per parcel8 (Table 7). This relatively high next 5 years on about 23% of the parcels being accepted by lending institutions, such as CoOs, can amount for modest-income families highlights the held under Mailo tenancy. The prospect of losing provide households with a more easily collateralized value they attach to more secure land rights. Yet, land in the next 5 years if the landlord were to sell asset, the effect of improving households’ access to knowledge of the type of documentation available is even more prevalent, at 41% of Mailo parcels. credit is far from certain as they face a multitude of to tenants, i.e. CoOs, and knowledge of the Similar figures are reported by the interviewed bottlenecks. In fact, the observed high incidence of process to obtain it, remain low, with only 19% of landlords: about a third of the landlords reported a transaction costs and risk constraints point towards the male respondents and 10% of female respondents dispute in the last 12 months on their holdings, and need to address the functioning of markets for risk and reporting being aware of the existence of CoOs. 43% expected land-related disputes in the future insurance at a broader level. However, those who are aware report that having (Table 5, panel 2). Clarifying kibanja rights in a such a document will somewhat (6%), likely (8%), 4 SHORTCOMINGS OF OVERLAPPING LAND RIGHTS AND WAY FORWARD 4 very likely (19%) and certainly (39%) decrease the Table 6: Access to credit and loans chances of being expelled from their land without semi- informal formal compensation. A surprisingly large number of formal landlords interviewed (60%) say they would sign the Applied for a loan in the past 12 months 0.11 0.12 0.02 document authorizing the tenants to obtain CoOs from Total amount applied for, UGX 492810 565058 14500000 the land administration authorities. The most Purpose of the loan: important reasons reported by the landlords are to: (i) Buy land 0.15 0.13 0.22 reduce conflict with tenants (53%), (ii) reduce conflict Purchase ag. Inputs 0.19 0.25 0.36 among tenants (16%), (iii) collect ground rent (13%), Buying livestock 0.02 0.03 0.04 and (iv) clarify parcel boundaries (11%). Reasons for Business 0.15 0.18 0.20 not agreeing to issue authorization include the burden Housing 0.03 0.07 0.12 Food 0.03 0.02 0.00 of encumbrance on their title (20%), and fear that it Educ 0.16 0.22 0.06 will dampen land prices (20%). Health 0.20 0.05 0.00 Buy a vehicle 0.02 0.01 0.00 In addition, 71% of tenant households expressed high Ceremonies (wedding, funerals etc.) 0.21 0.05 0.00 willingness to buy kibanja parcels from landlords if No guarantee required 0.89 0.36 0.10 they were to sell. However, tenants are willing to buy Land was required as a guarantee 0.05 0.27 0.62 the land at an average price of UGX 2.2M per acre, Household is unconstrained 0.36 0.26 0.06 which is significantly lower than the asking price of Household is price rationed 0.12 0.16 0.06 the 65% of landlords who indicate they are willing to Household is quantity constrained 0.40 0.28 0.11 sell to the tenants (UGX 6.8M). Yet, when tenants Household is risk constrained 0.73 0.69 0.31 were asked for how much they could sell their kibanja Household is constrained by transaction costs 0.08 0.60 0.37 rights with no reference to the landlord, the reported No. of obs. 2485 2485 2485 average price is UGX 8.4M per acre, which is higher than the asking price of landlords. This implies that there is a room for negotiations between parties if the market were to be opened up. Note that about 35% of the interviewed landlords are not willing to sale Mailo land to kibanja holders, mainly because they do not Table 7: Knowledge, willingness and perceived benefits of documentation want to subdivide their holdings (41%) and they do not Panel 1: parcel level (kibanja) want to sell land at all (45%). Willingness to pay for a legal certificate, UGX * 389681 Willingness to buy from landlord 0.71 Therefore, the data suggest that issuance of land Willingness to buy from landlord, UGX per acre * 2170470 certificates to tenants and buying land from the Estimated sale price, UGX per acre * 8378288 landlords both could be viable options. Sensitization No. Of parcels 2589 campaigns and establishment of negotiation platforms Panel 2: household level during the implementation of the project would help Female create the right environment in searching for long term Male respondent respondent solutions to the problem of overlapping rights on Knows what a CoO is 0.19 0.10 Mailo land. CoO decreases likelihood of expropriation w/o compensation: Not at all 0.28 0.36 Somewhat 0.06 0.05 Likely 0.08 0.06 CHANGING STRONG PATRIARCHAL Very likely 0.19 0.16 TRADITIONS THAT DISCRIMINATE AGAINST For sure 0.39 0.37 WOMEN AND GIRLS SHOULD BE A PRIORITY No. of respondents 1709 1986 Panel 3: landlord level Emphasis on female rights to land is critical to ensure Would sign for issuance of CoO 0.60 that they are not left out during adjudication and Reason for agreeing to sign CoO: regularization processes. As can be seen in table 8, Identify empty land 0.03 traditional patriarchal practices could put women in a Clarify boundaries 0.11 disadvantaged position in terms of inheritance and co- Collect busuulu 0.13 ownership of land. When asked about hypothetical Reduce conflict with tenants 0.53 scenarios of inheritance, only 31% of male Reduce conflict among tenants 0.16 respondents stated that the wife would inherit the land Other 0.05 of her dead husband. In addition, 39% expected that Reason for not agreeing to sign CoO: only the son would inherit the land if both parents were Encumbrance 0.20 to pass away, 44% expected it to be divided among Reduce land price 0.20 children, and only 7% stated that even if the son would Increase conflict 0.08 inherit he has a duty to look after his sister. Illegal tenants 0.08 Other 0.44 The numbers are slightly different for female Willingness to sell to kibanja holder 0.65 respondents, 41% of whom indicated that the spouse Price willing to sell per acre 5756098 inherits the land of her deceased husband. In addition, Reason for not wanting to sell: 32% of females interviewed reported that the male Doesn’t want subdivision 0.41 Doesn’t want to sell land 0.45 child would inherit all the land in case of the death of Other 0.14 both parents, 50% expected it to be divided between No. of landlords 63.00 * Data were trimmed dropping the 1st and last percentiles DANIEL ALI AND MARGUERITE DUPONCHEL 5 children, and 10% expected the male child to inherit with a responsibility of looking after his Table 8: Perceived gender rights on inheritance and registration sister. Male Female Regarding joint registration, 64% of male respondent respondent respondents but only 38% of female respondents If husband dies - married couple with two children, male and women think that only the husband’s name should be Spouse owns all the land 0.31 0.41 included in official land certificates. Only 34% of Spouse owns 50%, other to children 0.13 0.15 male respondents and 56% of female respondents Equal shares to children and spouse 0.08 0.08 expected that both the husband’s and the wife’s Son owns all the land 0.28 0.18 name would be on the certificate of occupancy. Children own all the land 0.18 0.16 Relatives take back the land 0.01 0.01 CONCLUSIONS Other owns the land 0.00 0.00 If both parents die: Clarifying overlapping rights on land could be a Son inherits the entire land 0.39 0.32 significant opportunity to increase land Son inherits the entire land but looks after sister 0.07 0.10 productivity and food security in the context of Daughter inherits the entire land 0.02 0.02 rapid population growth in Central and Western Land is divided between children 0.44 0.50 Uganda. The EU/GIZ intervention implemented in Land is passed on to others 0.04 0.01 partnership with the Ministry of Lands, Housing Don’t know 0.05 0.05 and Urban Development aims at strengthening Registration of land - which names should be on certificate land rights of kibanja holders. Baseline data Only husband's name 0.63 0.38 indeed emphasize the potential to improve Only wife's name 0.01 0.03 investment and sustainable land management on Husband's and wife's names 0.34 0.56 parcels held under Mailo tenancy, reducing land Indifferent on which one’s name 0.01 0.02 disputes, as well as easing certain constraints in Do not know 0.02 0.02 improving access to credit. Gender specific Registration of land (in the case of polygamy) - which names should be on certificate interventions are strongly advisable to lower the Husband's name on all plots certificate 0.59 0.38 risks that patriarchal norms could potentially Husband and first wife name on all plots certificate 0.08 0.14 exclude women from benefiting from the Husband and second wife name on all plots certificate 0.00 0.01 formalization of land rights. Transforming land Husband with first wife name some plots and with second wife on others 0.09 0.15 inventories to CoOs, or title (through land sales), Husband with first wife & children on some plots and with second wife & could be encouraged considering the observed children on some plots 0.12 0.18 willingness of both parties, and the apparent Husband with children of 1st wife some plots with of 2nd wife children inefficiency of the existing status quo which other plots 0.09 0.10 includes a high level of insecurity for tenants and Others' names 0.01 0.01 non-payment of ground rent for the landlords. The No. obs. 1709 1986 intervention may create the space for the establishment of platforms for dialogue among the parties that will be involved in the process. 1 Originally about 10% of the country was under a Mailo type of tenure: 9000 sqm were allocated by the British protectorate administrators to the Buganda Kingdom under Mailo, and 323 sqm and 268 sqm under Native Freeholds to the Toro and Ankole Kingdoms, respectively. In addition, land given to the Catholic Church has been administered similarly to that of mailo land (West, 1972). The Ministry of Lands, Housing and Urban Development estimates a larger surface, 30,000 sqm, under mailo tenancy, affecting 15% of the total population (Oput, 2013). 2 Ali,. D. A., K. Deininger, M. Duponchel (2014). Enumeration areas are selected on each side of the borders of the sub-counties selected for the pilot intervention. Credit Constraints and Agricultural Productivity: 3 The statistics on landlords should be regarded as suggestive informative considering small number of interviews and possible Evidence from rural Rwanda. Journal of bias in the sample. Development Studies. 50(5), 649-665. 4 This could be arguably linked to the refusal to recognize tenants as lawful by refusing to accept their payment of the busuulu. Deininger, K., & Ali, D. A. (2008). Do overlapping 5 Formal credit sector refers to formal banks while the semi-formal sectors refers to micro-finance institutions, NGOs, land rights reduce agricultural investment? ROSCAs, farmers’ groups etc. Informal sector refers to credit supplied by friends/family or money lenders. Evidence from Uganda. American Journal of 6 Agricultural Economics, 90(4), 869-882. We used a direct elicitation approach as described in Ali et al. (2014). Note that for the semi-formal sector, we report constraints based on multiple source of credit, hence households can face more than one type of constraint. Oput, R. (2013) Uganda. FIG- WB Forum on Land 7 Quantity constrained qualified households are those who borrowed but whose loan was partially or wholly rejected, and Administration in Sub-Saharan Africa. households who lacked collateral to borrow or to borrow additional amounts. Risk constrained households identify as fearing Presentation. to have their demand rejected or fearing to lose their collateral, or who don’t want to be (more) indebted as a reason for not borrowing or not being able to borrow more when they would have like additional resources. Households constrained by West, H. W. (1972). Land policy in Buganda. transactions costs do not have access to supplier, do not know how and where to apply, have no bank accounts or are not a member to the institution. 8 This represent half of the harvest value for a season, and 5% of the estimated sale price of the parcel per acre. This case study was prepared by a team from the Development Economics Research of the World Bank, in collaboration with the Ugandan Bureau of Statistics. The German Cooperation Agency (GIZ) provided generous financial support. Photo credit: GIZ/Martina Bacigalupo. The findings, interpretations, and conclusions expressed are entirely those of the authors. They do not necessarily represent the views of the Government of Uganda, and of the World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.