81679 CASE STUDY WORLD BANK | AGRICULTURE AND DEVELOPMENT OCTOBER 1, 2013 Are Poor Slum-Dwellers Willing to Pay for Formal Land Title? Evidence from Dar es Salaam DANIEL AYALEW ALI, MATTHEW COLLIN, KLAUS DEININGER, STEFAN DERCON, JUSTIN SANDEFUR, AND ANDREW ZEITLIN were hailed as among the most advanced land laws in Africa (Alden-Wily 2003). For urban areas, the Land Act defines Certificates of Right of Occupancy (CROs), i.e. leases with duration between 33 and 99 years, as the main type of land ownership. A short term lease known as a residential license (RL) was also introduced as two-year rights which are renewable but non-transferable.1 However, rapidly increasing demand for land, especially in urban areas, has exposed many weaknesses in Tanzania’s urban land management (World Bank 2009). Effective implementation of legal provisions is under- mined by over-reliance on the central government and top-down planning, which stifles local government initiative; this results in an often incoherent regulatory framework and unrealistic standards (Kyessi and Kyessi 2007). Although the country has close to 40 ABSTRACT million inhabitants and estimates suggest that —————–——————————————————————–———————— in Dar es Salaam alone, 15,000-20,000 new High levels of informality, from 70% to 80% in many developing country cities, parcels are added each year, the land registry increase poverty and make urbanization unsustainable. But it is not clear if this is has records for only around 223,000 parcels. due to lack of interest in more secure land rights or the high cost of traditional With more than 80% of plots in Dar es Salaam institutions. To find out, and see if formalization can help increase female being informal, levels of urban informality in empowerment, we conduct an information campaign about the importance of Tanzania are among the highest in Africa female land ownership in poor informal neighborhoods of Dar es Salaam and allow (Kironde 2006). randomly selected residents to get title at a discounted price, with possible additional discounts if a woman is on the title. We find that demand for title at The difficulties involved in obtaining CROs and affordable prices is very high and that, after the information campaign, many the shortcomings of RLs have impeded land males added females to title applications even if they did not receive large fiscal tenure formalization in the country’s urban incentives. These results could have far-reaching impacts for the most appropriate areas. The complex, costly, and time- way to support urban infrastructure upgrading. consuming process for obtaining a CRO puts BACKGROUND from measures to broadly award such rights, secure title beyond the reach of even relatively arguing that doing so may be affluent households (World Bank 2009). The Land tenure formalization is frequently unaffordable, that informal arrangements can system also exhibits a striking bias against recommended as a policy tool for substitute for formal recognition, or that females, with less than 20% of the land developing country governments to spur securing rights will hinder city expansion. registered in the name of women (Deininger et savings and investment by the poor, and al. 2011). While the government attempted to extend access to credit by unlocking In Tanzania, the focus of this study, a increase tenure security by issuing RLs on a collateral. Nevertheless, very few sweeping reform of the country's land laws in large scale to dwellers in informal developing countries have made significant 1999 created a legal pathway for rural and settlements, demand for these was limited; to progress in expanding formal land tenure to urban households to acquire formal tenure date less than 92,000 RLs have been issued. the poor. Policy-makers often shy away rights. The Land Act and the Village Land Act Possible reasons for this low demand include ECON.WORLDBANK.ORG/RESEARCH 2 ARE POOR SLUM-DWELLERS WILLING TO PAY FOR FORMAL LAND TITLE? 1. perceived few benefits from a document that is only valid for 2 years without transferability1; 2. non-trivial costs (an initial payment of TSh. 5,600 and an annual license fee of TSh. 3,000, along with property taxes); and 3. limited inter-institutional coordination and technical defects. This study explores whether poor occupants of informal areas are interested and willing to expend resources for more secure property rights by conducting a randomized experiment in Dar es Salaam’s informal settlements. Specifically, it examines two issues involved in expanding formal land tenure: cost and willingness to “co-title,” i.e. include women on land titles. DESIGN Control and Treatment Blocks in the Barafu Sub-ward The setting for this study is Kinondoni, one DESCRIPTIVE EVIDENCE data at the plot level, illustrating that most of the three municipalities of Dar es (75%) parcels were purchased rather than Salaam. The purpose was to induce house- Basic characteristics of the intervention obtained via inheritance (18%), government holds in both communities to purchase areas, in particular the Barafu Sub-ward, programs (3.5%), or squatting (2.4%).2 In line CROs in order to subsequently study their show low formal, permanent employment, with other studies, land rights seem quite impact. In 2010, a baseline survey of all property size value, and access to utilities. secure and less than 10% of parcels have 2,500 property owners in the selected sub- Monthly income in Barafu is roughly TSh. limited transferability. wards was conducted. Subsequently, ward - 4,100, and very few of the households have level meetings were held by a local NGO to applied for credit, with most being credit Information on RLs also highlights some of the explain the overall intervention and the constrained. The table on Barafu presents gaps inherent in past efforts to award more modalities under which individual house- Parcel Level Characteristics of Barafu Sub-Ward by Gender holds could obtain CROs. The campaign Total Head’s gender was combined with strong sensitization Female Male regarding the importance of women to be registered as joint land owners, as Parcel Size and Mode of Acquisition an improvement in women’s property rights Parcel size in square meters 249.884 256.242 231.715 ** can lead to female empowerment and other Parcel was purchased (%) 75.9 63 79.9 *** desirable outcomes in child health and Acquired through inheritance/gift (%) 18 29.2 14.8 *** education. Government programs (%) 3.5 3.7 3.1 Parcel was squatted on (%) 2.4 3.3 2.1 We divided land parcels (2 sub-wards with Access to Utilities 1,000 parcels each) into approximately 40 Parcel with piped water connection 3.6 1.6 3.9 * “blocks” (contiguous groups of parcels), Access piped water from a neighbor 15.5 18.9 14.6 randomly assigning half of these into Has access to electricity 61.1 56.4 62.2 treatment and control groups. All parcels in Land rights treatment blocks were subject to a Land can be sold w/o approval 46.6 43.6 47.5 cadastral survey and owning households Land cannot be sold 8.3 13.1 7 were invited to participate in the program to Land can be rented w/o approval 50.6 49.2 51 obtain CROs for TSh. 100,000 over roughly Land cannot be rented 7.0 10.2 6.1 a six month period. To trace out demand Land can be bequeathed 49.1 47.5 49.6 and estimate the price elasticity of demand Land cannot be bequeathed 7.8 13.6 6.1 for CROs, we offered discounts by Ownership and Legality of the Parcel randomly providing individuals with Head is the owner of the parcel 95.5 97.1 95.8 vouchers ranging in value from TSh. 0 to Women own land (alone or jointly) 26.5 97.9 6.1 *** TSh. 80,000 (i.e. households could face Parcel has residential license (RL) 36.7 31.3 38.4 * subsidies between 0-80% of the total cost Spouse's name is on the RL 4.0 0.0 5.0 of a CRO). In addition, the main, Women in the RL (alone or jointly) 20.2 84.2 5.3 *** unconditional voucher was complimented WTP for CRO (TSh, median) 50,000 50,000 50,000 with a “pink” conditional voucher that could Number of parcels 1092 243 842 only be redeemed if a female was listed as Source: 2010 Baseline household survey. co-owner on the CRO. The size of the total Significant levels are reported for t-tests of the equality of the means for each of the variables discount that could be received by any between parcels owned by female and male headed households. household was limited to TSh. 80,000. * significant at 10%; ** significant at 5%; *** significant at 1%. 3 ARE POOR SLUM-DWELLERS WILLING TO PAY FOR FORMAL LAND TITLE? secure land rights, especially with respect conditional vouchers, extolling them to co- providing CROs, it is much higher than to gender. While owners have acquired RLs title. This suggests that land tenure what is currently charged for RLs. To the for some 37% of parcels and 20% of them formalization programs and awareness extent that past initiatives were judged to have women in the RLs (alone or jointly), campaigns that creatively aim to empower have been less than fully successful, the spouses’ names are recorded on 4% of the women can make an important contribution underlying reasons are unlikely to be licenses overall. This suggests that, in to overcoming long-standing gender bias. informal residents’ lack of knowledge about Barafu, RLs either reinforced or helped to the importance of secure land tenure or create a relatively gender-based pattern of Finally, the results of our study suggest that their unwillingness to pay for it. Instead, land access whereby women failed to re- it may be worth considering decomposing limited uptake of RLs seems to have more ceive equal access to housing property the cost of CROs in order to make them to do with the limited incremental benefits even in cases where they were considered affordable. Neighboring countries such as they offer. This suggests that, in the joint owners. Rwanda have been able to title urban plots context of Dar es Salaam, policy initiatives at significantly lower costs with reduced to make CROs more accessible—for RESULTS standards. As cadastral survey accounts for example by lowering the required survey some two thirds of the cost of a CRO, standards—or to enhance the rights Using the experimental variation in the price systematic surveying, possibly at slightly Net Effect of Vouchers on Co-titling of CROs, we constructed a demand curve lower standards, could reduce cost Pooled for formal land titles. In Barafu, a majority of dramatically. All Dual-headed residents are willing to pay up to TSh. CONCLUSION General voucher 0.00109** 0.00158** 20,000 to obtain a CRO. Although this falls Conditional short of the price the Ministry of Lands 0.00228*** 0.00288*** voucher estimates must be charged for immediate If the benefits from formalization are indeed Constant 0.326*** 0.324*** full cost recovery (over TSh. 150,000 per as large as claimed in the literature, the high Baseline controls Yes Yes parcel to cover cadastral survey costs, levels of informality found in most developing Obs 1037 603 etc.), it is much higher than what is country cities, as well as the gender-biased Voucher Effects on Female Co-titling* currently charged for RLs. The mean per way in the few titles in existence tend to be Pooled capita monthly income in Dar es Salaam allocated, are puzzling in a number of ways. 1 2 was approximately TSh. 80,000 in 2006. On the one hand, they suggest that, by General voucher -0.000828 The fact that even poor households prefer failing to formalize informal settlements and Conditional CROs to RLs is demonstrated in their empower women by giving them a stake in 0.00372*** voucher willingness to commit a significant portion of property, society may forgo a large amount Conditional = 20 0.288*** their resources to obtain a CRO; this of social benefits. Also, if benefits are as Conditional = 40 0.297*** suggests that there is unsatisfied latent high as claimed, individuals should demand Conditional = 60 0.324*** demand for formal recognition of land formalization and a market for providing it Conditional = 80 0.304*** ownership. should eventually emerge. Constant 0.733*** 0.645*** Baseline controls Yes Yes Second, our results show that households By providing informal residents in poor Obs 264 264 respond positively to price incentives by informal neighborhoods the opportunity to *Conditional on CRO Application co-titling. General vouchers have a positive acquire formal property rights but still Notes: Linear probability model. Dependent effect on co-titling because they induce charging a household-specific price, this variable = 1 if household has fully paid for a households to purchase CROs, many of study allows us to advance towards CRO and included a woman on their CRO application. Sample is restricted to house- whom go on to co-title even without addressing these puzzles in two areas. First, holds with application data. Column 1 uses conditional incentives. Households which we find that, although willingness (and linear measures of voucher values. Column receive any conditional voucher are 29-30 ability) to pay for formal documentation of 2 introduces dummy for each voucher value; percentage points more likely to co-title property rights remains well below the costs general voucher figures deleted. than those that receive no voucher. This involved in current sporadic efforts of *p < 0.10; **p < 0.05; ***p < 0.01 effect is persistent and statistically indistinguishable across all voucher values, indicating that households are effectively nudged into co-titling by conditional vouchers. Furthermore, in our experiment, dissemination activities helped to raise the share of households who indicated that they would co-title from 24% to 89%. Even if some of these intentions may not be followed through in practice, this is a large increase from the 5% of RLs issued carrying a female name. The results are consistent with the hypothesis that the collaborating local NGO responsible for Cartoon used as part of the intervention to inform residents on the importance of having land marketing land titles exerted a strong titles to prove rightful ownership, allowing for secure land transactions. influence on all residents, at all levels of 4 ARE POOR SLUM-DWELLERS WILLING TO PAY FOR FORMAL LAND TITLE? Project Area associated with RLs could go a long way in only the commonly held belief that poor Deininger, K., H. Selod and A. Burns. 2011. dealing with the city’s problems of people are not be able to recognize or pay Improving governance of land and informality. The cost reductions achievable for more secure land title but also the way in associated natural resources: The Land by a systematic approach, combined with which most land titling interventions allocate Governance Assessment Framework residents’ willingness to pay for real resources- between efforts aimed at Washington, D.C.: World Bank. improvements in property rights, moreover surveying and those aimed at suggest that even in cases where full cost sensitization and awareness. Kironde, J. M. L. 2006. "The regulatory recovery is unlikely, large scale property framework, unplanned development and rights reforms could generate a sustained * * * urban poverty: Findings from Dar es Salaam, flow of revenues (eg. through land taxes), Tanzania." Land Use Policy 23 (4): 460-72. 1 the present value of which may well exceed As Tanzanian legislation requires that the cost of initial establishment of such a anybody who has been in possession of a Kyessi, S. A. and A. G. Kyessi. 2007. system. plot for three years becomes eligible for "Regularization and formalization of informal compensation in case of expropriation for settlements in Tanzania: Opportunities and Second, our data confirm that traditional public purpose, the two-year limit was chalenges in the case of Dar es Salaam interventions in Tanzania have indeed motivated by the desire to not create any City." . Hong Kong: Paper presented at the been characterized by a strong anti-female rights that could be eligible for FIG Working Week, Hong Kong SAR, China, bias. A large part of such bias may be compensation. In response to widespread 13-17 May, 2007. attributed to a lack of sensitization. The fact complaints and in order to revive flagging that even households who at the time of interest in RLs in general, this limit has World Bank. 2009. "The urban transition in the baseline survey had no intention of recently been increased to 5 years. Tanzania." . Washington, DC: The World putting a female co-owner on a Bank, Tanzania Country Management Unit. 2 hypothetical CRO did eventually sign up for While male headed households purchased joint title implies that the awareness nearly 80% of their parcels, it is only a bit campaign had a significant impact on the more than 60% in the case of female headed * * * decision of who to declare as a land owner. households. On the other hand, the share of It will be of interest to explore the extent to inherited land is relatively higher for female This case study was prepared by Daniel which these intentions are followed through headed households as compared to male Ayalew Ali and Klaus Deininger from the and their eventual impact on economic headed households. However, as data on World Bank Development Economics outcomes and intra-household bargaining. the source of the inheritance is not available, Research Group (DECRG), Matthew Collin it is difficult to examine the potential link with and Stefan Dercon from the University of Even without the evidence that later studies the inheritance law. Oxford, Justin Sandefur from the Center for may provide, our evidence is strong Global Development, and Andrew Zeitlin enough to conclude that, in light of the REFERENCES from Georgetown University. The Knowledge rather modest resource requirements, more for Change Program (KCP), UN Habitat’s investment in gender sensitization efforts Alden-Wily, L. 2003. "Community-based land Global Land Tool Network, DFID’s could have large social and economic tenure management: Questions and answers International Growth Center, and the World benefits. From a policy perspective, our about Tanzania's new Village Land Act, Bank’s Gender Action Plan Trust Fund findings imply a need to re-evaluate not 1999." IIED Issues Paper 120. London: IIED. provided generous financial support. The findings, interpretations, and conclusions expressed are entirely those of the authors. They do not necessarily reflect those of their respective institutions.