CASE STUDY Sustaining the success of the systematic Land Tenure Registration in Rwanda DANIEL ALI, KLAUS DEININGER AND MARGUERITE DUPONCHEL INTRODUCTION Since 2010, the Government of Rwanda addressed land tenure through a series of actions including the establishment of institutional and administrative structures for land management and administration. At district, town, and municipality level, District Land Bureaus (DLBs), complemented by sector and cell level land committees, assumed responsibility for land administration and planning. After careful piloting in 2008/09, a process for systematic low cost demarcation and adjudication using aerial photography or high resolution satellite imagery was roll-out nationally. In less than 3 years, the Rwanda Natural Resource Authority (RNRA) demarcated 11.4 million out of an estimated 11.5 million land parcels in the country in a participatory way and at a unit cost of less than USD 6 per parcel (Nkurunziza 2015). Positive impacts at household levels were identified at various stages of the project Fig. 1. Parcels on a Rwandan hill implementation. First, the evaluation of the pilot agricultural investment in the medium- and long- 1.5% in commercial, and 0.3% in administrative exercise suggested that land titling had: (i) term (Ali et al. 2015). To sustain and expand these use (table 1). For some 80% of parcels, registered improved land access for legally married women impacts, especially given Rwanda’s active land claimants are natural persons, followed by 16% and better recordation of inheritance rights; (ii) markets, it is essential that records accurately with claimants not yet recorded (likely to be related significant investment impacts (in soil reflect the current status of any parcel of land, to the 14.2% of parcels the use of which is not conservation), particularly pronounced for implying that all subsequent transfers are categorized) and 4% with claimants being legal female-headed; and (iii) a marginal reduction in registered. persons. The data also support the gender sensitive land market activity rather than a wave of distress nature of LTR: although we lack data on sales (Ali et al. 2014). Second, data collected soon households’ gender composition, 86% of parcels after the national roll-out program completion in CURRENT LEVEL OF REGISTRATION AND owned by natural persons have a woman either as 2012 showed that (i) tenure security had increased TRANSACTION FREQUENCY – EVIDENCE sole (25%) or co-owner (61%) and only 14% of markedly for males and females equally, FROM THE ADMINISTRATIVE DATA these were registered exclusively to male including those who were not legally married; (ii) claimants. the program led to a marked improvement in land Official land administration information system rental market functioning and associated Although LAIS records information on many types (LAIS) data highlight Rwanda’s accomplishments. efficiency-enhancing land transfers (Ali et al. of changes in the registry, we focus on ownership, During LTR, information was collected on 11.42 2015); (iii) although investment had not increased i.e., sales, gifts, and inheritances. Table 2 provides million parcels. Of these, 64% are in agricultural, yet, it has provided a basis for higher levels of 12% in residential (34% in Kigali), 8% in forest, SUSTAINING THE SUCCESS OF THE SYSTEMATIC LAND TENURE REGISTRATION IN RWANDA figures on total registered transactions in 2014 maintain the high level of gender equality Table 2: No. of registered transactions leading and 2015, mean numbers of transactions for each attained by LTR, the bottom panel of table 2 to transfer of ownership in 2014/15 of the 2,148 cells, and the share of registered disaggregates registered land transaction by Total Kigali city parcels for which ownership changes were gender. Comparing to table 1 points towards a Panel A: Registered residential land registered, separately for residential and consistent shift in gender composition of Total number of transactions agricultural land. The share of parcels where claimants with the share of those involving All 45,293 16,710 changes in ownership were registered is higher women being 86% for first but only 76% for Sales 21,367 15,145 Others 24,120 1,665 for residential (1.6% for sales and 1.8% for other subsequent registrations, suggesting attention to Sale value ($/m2)a 30.07 39.38 ownership changes) than for agricultural (0.2% this issue be warranted. Mean no per cell per annum of sales and 0.12% other transfers) land. All 10.54 51.9 Differences are even more pronounced between Sales 4.98 47.04 MEASURES UNDERTAKEN BY THE Others 5.62 5.18 Kigali where, in the two years for which we have Percent of registered parcels per annum GOVERNMENT TO ENSURE data, 11.3 % of residential and 3.07% of All 1.69 6.23 SUSTAINABILITY Sales 0.80 5.65 agricultural parcels were transferred via Others 0.90 0.62 registered sale, compared to the rest of the Panel B: Registered agricultural land While the process of registering transfers is clear country where corresponding figures are 0.5% Total number of transactions in principle, unaffordable fees, lack of for residential and 0.12% for agricultural land. All 22,849 6,635 awareness, and difficult access may create Sales 14,497 6,087 Interestingly, a total of US$ 2.6 billion of obstacles. Aware of these, the Government has Others 8,509 609 mortgage lending based on 49,694 mortgages 2 a Sale value ($/m ) 4.45 6.65 started to tackle potential bottlenecks. (65% secured with residential, 30% with Mean no. parcels per annum agricultural, and 5% with commercial land) had First, registering a transfer incurs a flat fees of All 5.32 20.6 Sales 3.38 18.9 been entered in the registry as of Jan. 1, 2016 RwF 27,000 or about USD 40. As flat fees are Others 1.98 1.9 (almost two third of which in Kigali). regressive and may be unaffordable or difficult Percent of registered parcels per annum to justify for rural people with small plots, a All 0.16 1.68 Table 3 points towards differences in prices of Sales 0.10 1.54 review was conducted in 2015. It recommended land (with structures) between residential land Others 0.06 0.16 to waive stamp duty and possibly reduce Panel C: All registered land with a mean price of US$ 30 per m2, from US$ registration fees, for rural parcels below 5 ha and New registrations by gender 39.4/m2 in Kigali, where more than 70 % of Female only 0.18 0.16 to set urban fees at levels that would allow the sales are concentrated, to US$ 4.75/m2 in the Male only 0.23 0.23 registration system to sustain itself, in line with East (slightly more than 10% of registered sales). Joint 0.60 0.62 evidence that transfer taxes raise little revenue No. of cells 2148 161 Agricultural land has a mean sales price of US$ compared to land taxes but encourage Others 0.06 0.16 4.45/m2 (US$ 6.65/m2 in Kigali to US$ 1.47/m2 Source: RNRA, Land Administration Information informality and threaten sustainability of land in the East). To check if subsequent transactions System (LAIS). aTwo residential and two agricultural registries A revised fee structure is, however, yet land with values above 5,000 $/m2 from Kigali city are to be finalized and to be approved formally by dropped. Table 1: Number and size of registered the government. parcels by type Total Second, lack of awareness by concerned parties Total no. of parcels information and be in charge of all land-related issues may be a key reason for high variability in the Total no. of registered at local level (including the registration of transferred 11,420,885 extent to which subsequent transactions are parcels land), thereby helping to decentralize land services. All of which residential 1,341,467 registered. Indeed, many buyers thought taking of which commercial 169,709 SLMs received some basic training and were sworn so possession of the seller’s title without name of which agricultural 7,344,802 they could notarize transactions. SLMs were deployed of which forest 916,359 change would establish ownership or, unaware in three batches in the third quarter of 2014 and the of which admin./sc./social 29,749 of the documents needed and process to be of which not Categorized 1,618,799 second and third quarters of 2015. By the end of 2015, followed, had transfers authenticated by village Parcel size (ha) 0.18 367 of the 416 sectors in Rwanda had a SLM in place. officials (see below). To address this, the .. for residential land 0.13 .. for agric. land 0.19 Government conducted two national campaigns, Share by claimant type referred to as ‘land weeks’, in May/June 2014 EXPLORING POTENTIAL EFFECTS OF THESE Natural person 0.80 and February/April 2015. These involved MEASURES Non-natural person 0.05 Claimant not yet registered 0.16 extensive coverage on television and radio, and Share by gender of claimants (if face-to face events in 150 sectors. To explore if such measures to improve awareness and natural person) Female only 0.25 Third, as DLOs may be too distant from average services access affected registration of transfers, we Male only 0.14 citizens to act as the first point of contact for exploit quarterly data for 2014 and 2015 on the number Male and female jointly 0.61 of registered transactions across the country’s 2,148 delivery of land services (12-15km), sector land Source: RNRA, Land Administration Information System (LAIS). managers (SLMs) were recruited to provide cells. DANIEL ALI, KLAUS DEININGER AND MARGUERITE DUPONCHEL Table 3: Determinants of subsequent transaction registration using quarterly data at the cell level (2014 and 2015) Involve transfer of Sales only Inheritance, rights donation, etc. Panel A: Residential land Distance to DLO in km 0.018 0.008 0.010 (0.035) (0.015) (0.029) Sector land manager appointed 1.339*** 0.791*** 0.536** (0.385) (0.301) (0.224) Sector covered by land week events -0.072 -0.095 0.015 (0.337) (0.302) (0.111) No of residential parcels in the cell (‘000) 4.235*** 2.728*** 1.516*** (0.805) (0.624) (0.325) Ratio of urban population in the sector 6.335*** 4.704*** 1.671** (0.292) (0.255) (0.085) Constant 0.229 1.824* -1.558** (1.096) (0.956) (0.627) Number of observations 17,184 17,184 17,184 R2 0.120 0.373 0.071 Panel B: Agricultural land Distance to DLO in km -0.038*** -0.014* -0.024*** (0.012) (0.008) (0.006) Sector land manager appointed 1.101*** 0.985*** 0.110 (0.272) (0.235) (0.083) Sector covered by land week events -0.050 -0.073 0.022 (0.253) (0.212) (0.063) No of agri. parcels in the cell (‘000) 0.183*** 0.129*** 0.055*** (0.043) (0.032) (0.017) Ratio of urban population in the sector 1.641* 0.358 1.281*** (0.215) (0.160) (0.071) Constant 3.410*** 3.680*** -0.229 (1.088) (0.946) (0.263) Number of observations 17,184 17,184 17,184 R2 0.145 0.200 0.069 Note: Robust standard errors clustered at the sector level in parenthesis: *** significant at 1%; significant at 5%; * significant at 10%. Province, year-quarter fixed effects included but not reported. Fig. 2. Females signing their land lease certificate Table 3 presents results separately for residential and agricultural land. Although more relevant for sales than for non-monetary transfers, presence of a SLMs emerges as a key determinant of having transactions formalized. It is estimated to raise numbers of registered sales by about 0.79 for residential and by 0.96 for agricultural land per quarter per cell, corresponding to 64% and 117% increases for residential and agricultural land, respectively, per year. While structural characteristics such as the number of residential parcels in the cell and the share of urban population in the sector are strongly associated with registration of residential transfers, this is relatively less the case for agricultural land. Distance to the DLO, which serves as a proxy for transport costs, is significant and quantitatively large for agricultural land, in particular for non- monetary transactions; estimated coefficients suggest that for distances above 10 km from the DLO, the likelihood of registration of non-monetary transfers will be cut by half and that of sales by about a fifth. At the same time, while it is possible that the awareness campaigns made a significant contribution to the observed increases in the number of registered transactions over time, conduct of land week events in specific sectors, over and above national campaigns, does not seem to have had a significant impact on greater formality. Table 4: Key household characteristics Total Owns land 0.94 INCIDENCE OF RURAL INFORMALITY AND POSSIBLE DRIVERS Total land owned (ha) 0.68 Land market participation While the use of registry data allows us to assess changes in the level of formally registered Transacted in land 0.24 Share who purchased 0.62 transfers, assessing informality requires linking household- to registry data to see if transfers by … if yes, size purchased 0.12 households are recorded in the registry. We use the latest round of a three-round panel of some Share who inherited 0.46 3,600 households aimed originally to evaluate the impact the national roll-out of the LTR … if yes, size inherited 0.07 Transacted out land 0.28 program. It captures all transactions conducted between 2012 and early 2015. As some land … if yes, share of subdivision 0.29 transfers may have been formalized in the context of LTR rather than via subsequent … if yes, size transacted out 0.18 registration, the ability to use the survey to assess informality of subsequent transactions is Access to land services somewhat constrained and to do so properly, a new survey would be needed. At the same time, Distance to DLO, km 14.5 No. of households 3348 Source: Own computation from 2015 round of Rwanda Land Tenure Regularization Survey 4 SUSTAINING THE SUCCESS OF THE SYSTEMATIC LAND TENURE REGISTRATION IN RWANDA 4 Table 5 Key parcel characteristics Total Transfer Old New Characteristics Size, ha 0.18 0.19 0.12 Main use residential 0.25 0.27 0.13 Main use annuals 0.56 0.55 0.65 Main use perennials 0.16 0.15 0.18 Mode of acquisition Purchased 0.34 0.33 0.41 Inherited 0.45 0.48 0.29 Allocated by gov’t 0.03 0.04 0.02 Rented in 0.05 0.03 0.16 Other 0.10 0.10 0.05 Perceived land rights Can sell 0.90 0.90 0.89 Can give 0.92 0.92 0.91 Can use as collateral 0.90 0.90 0.90 Can rent out 0.91 0.91 0.90 Can divide 0.93 0.93 0.92 Any dispute 0.02 0.02 0.02 No future dispute 0.97 0.97 0.97 Won’t lose if left fallow 0.91 0.91 0.91 No expropriation 0.67 0.67 0.70 14,05 12,2 No. of parcels 4 53 1,801 Source: Own computation from 2015 round of Rwanda Fig.3. Meeting during Land Week Land Tenure Regularization Survey and having this caveat in mind, using survey were purchased while 29% were inherited and For only 2% of the total (23 parcels), official results provides interesting information that can about 16% were rented in for temporary use. Yet, registration had been started/completed. With help to inform decisions as long as results from a knowledge about procedural and legal provisions 62% vs. 31% for purchases and inheritances, proper survey are not available. to be followed upon property transfer, remains levels of informality of transactions are much low, with only 7% and 4% of male or female higher for the form (most of the inherited While 94% of households own agricultural land, respondents knowing the correct process, figures parcels were in fact registered during LTR). total holding sizes, with an overall mean of 0.68 that do not improve dramatically if only those There are also enormous differences across ha, are small (table 4) and the average parcel size involved in land transfers are considered (table regions, with levels of informality ranging of 0.18 ha (table 5) is substantially smaller than 6). from 37% and 42% in the North and South to the minimum 1 ha required for subdivision of 53% and 60% in the East, highlighting the agricultural parcels legally, a potential cause for Linking the household survey information to the importance of management and supervision the absence of registration if transferred in LAIS through the unique parcel identifier (UPI) that might differ across provinces. fragments. In fact, respondents perceive an assigned at the time of demarcation enables to unrestricted right to sub-divide for 93% of the check whether parcels that newly acquired are Second, with almost half (42%) of households parcels, pointing towards a potential clash with also registered in the name of the current owner having only registered these transfers locally the land law. With 24% and 28% of the sample, in official records (table 7). Key results point to (with village leaders), levels of awareness of respectively, reporting to have acquired or three main findings (reflecting the reality on the the need for officially registering subsequent transferred out land (62% having purchased land, ground at the time of survey). First, of parcels that transaction to ensure their legal validity 46 % having inherited), land markets are very had been transferred, 51% had been covered by remains low. The fact that 21% of the active. About 1,800 parcels with an average size first-time land registration under LTR, another households who failed to register their of 0.12 ha were acquired mainly for agricultural 47% had not been officially registered (not in the transactions quoted ‘lack of information’ as purposes (87%) since 2012. The majority (41%) name of the current owner in the LAIS records). main reason reinforces this. Third, though the number of registered parcels remains low, the fact that their mean value is Table 6: Knowledge of Legal procedures for Registration of Land much higher than that of unregistered ones Transactions Post LTR could imply that, beyond information, the Acquired land rather high flat fee of registering, which, as a Total No Yes share of parcel value, is lower for more Land markets - Male 0.07 0.06 0.10 Land markets - Female 0.04 0.04 0.04 valuable parcels, could be a relevant factor Source: Own computation from 2015 round of Rwanda Land Tenure (table 8). Computing the registration fee as a Regularization Survey DANIEL ALI, KLAUS DEININGER AND MARGUERITE DUPONCHEL 5 Table 7: Incidence of Informal Transactions Total Southern Western Northern Eastern Share of newly acquired land since last visit (i.e., 2012)a 0.12 0.11 0.11 0.12 0.14 All types of transactionsb Registered during LTR 0.512 0.577 0.369 0.597 0.469 Subsequent registration 0.020 0.006 0.035 0.039 0.000 Not officially registered 0.468 0.417 0.596 0.365 0.531 Of which: register only at the village 0.423 0.659 0.145 0.363 0.547 RwF willing to pay for reg. 4442 2937 3794 4208 7188 Number of parcels 1,147 324 255 310 258 Market transactions Registered during LTR 0.352 0.354 0.218 0.418 0.395 Subsequent registration 0.026 0.007 0.056 0.050 0.000 Not officially registered 0.622 0.639 0.726 0.532 0.605 Number of parcels 576 144 124 141 167 Inheritance transactions Registered during LTR 0.673 0.756 0.511 0.746 0.604 Subsequent registration 0.014 0.006 0.015 0.030 0.000 Not officially registered 0.313 0.239 0.473 0.225 0.396 Number of parcels 571 180 131 169 91 Reason for non-registration Lack of information 0.21 0.34 0.09 0.18 0.31 Lack of interest 0.13 0.10 0.10 0.24 0.09 Fees to high 0.13 0.16 0.06 0.13 0.19 Will register later 0.12 0.10 0.03 0.22 0.18 Seller didn’t have certificate 0.02 0.03 0.04 0.00 0.02 DLO is too far 0.01 0.01 0.01 0.01 0.01 Conflict on the parcel 0.01 0.01 0.01 0.01 0.01 Other reason 0.38 0.25 0.67 0.22 0.21 No. of owned parcels 12,798 3,574 3,501 3,234 2,489 Source: Own computation from 2015 round of Rwanda Land Tenure Regularization Survey a Includes parcels acquired outside the cell boundary of the sampled villages as well as those with missing detailed information on subsequent registration. b Restricted only to parcels acquired within the cell boundary of the sampled villages as UPI was not collected for these parcels. Using the UPI, the survey data was matched with administrative data from the Land Administration Information System (LAIS) to further cleanup the timing of registration (i.e., at the time of LTR or subsequent registration) of the newly acquired parcels. share of property value using data on the value of Table 8: Parcel characteristics of newly acquired land in rural areas (restricted to those within the cell boundary) transferred parcels from the survey or, for Kigali, Registered Not officially Subsequent administrative data from the land registry during LTR registered registration emphasizes the hindrance created by the high flat Year acquired 2012 2013 2013 fee in rural areas (table 9). For 10% of rural Land area in ha 0.13 0.11 0.09 Purchased 0.35 0.67 0.65 parcels, the self-reported parcel value is below the Registered land price in USD 290.12 365.01 698.98 registration fee, rendering registration difficult at Inherited 0.56 0.22 0.26 best. Even if these are dropped, the effective Government allocated 0.05 0.02 0.00 Subjective land value in USD 479.46 447.95 665.59 registration fee rate for rural parcels is close to Registration fee paid (US $) 2.14b 30.06 25% of the median property value, compared to Number of parcels 587 537 23 only 0.64% of the median property value for urban Source: World Bank 2015 third round follow up survey. b parcels. In light of this, changes to make the fee Note that 90% of the cases paid only the official rate of USD 1.42. structure more equitable could likely support the sustainability of Rwanda’s land registry. Market, buyers tend to have significantly higher levels of pay, this should provide a lower bound estimate to pay (WTP) for registering land transactions, farm abilities. of true willingness to pay. Hence, while 20% while the self-assessed quality of land or were not willing to pay, average WTP for those whether the plot is used for residence do not with positive amounts is RWF 4,500 (about seem to influence amounts households are WILLINGNESS TO PAY FOR 20% of the official fee) with a median value of willing to pay. WTP is lower for registering REGISTRATION AND ITS DETERMINANTS RWF 1,000, equivalent to the nominal government allocated land but increases with But what are individuals willing to pay for having participation fee charged for first-time land size: doubling parcel size from the mean their transferred land registered? To explore this, registration of a rural parcel. of 0.18 ha is estimated to increase WTP by we asked those who acquired new land but did not about 11%. Households located in sectors To investigate determinants of households’ yet register to name the amount they would be where an SLM had been posted at the time of willingness, we used a regression analysis willing to pay to have the acquired land registered. survey increases the self-estimated value of (Table 10). At parcel level, the size and mode of To the extent that lack of awareness would likely registering transactions are willing to pay more acquisition matter for households’ willingness reduce the amount households would be willing to for official registration (about 60%), 6 SUSTAINING THE SUCCESS OF THE SYSTEMATIC LAND TENURE REGISTRATION IN RWANDA supporting the notion that greater awareness Rwanda’s institutional infrastructure that other Table 9: Effective registration rate increases as the proximity of land services institutions in the public or private sector can use Effective reg. fee (%) reduces. Households further away from their to add value and derive benefits. Decile Rural Kigali 1 99.5 5.27 respective DLOs are willing to pay less for While an ongoing survey can provide a more up- 2 69.4 1.83 registering transactions, possibly compensating 3 42.2 1.15 to date picture, including a documentation of the for higher transport costs. 4 29.9 0.84 benefits from reform and their incidence among 5 22.6 0.64 different groups for Kigali, having up to date data 6 17.2 0.49 CONCLUSION on levels of informality in rural and peri-urban 7 13.3 0.36 8 9.8 0.26 areas beyond Kigali would be highly desirable. The evidence suggests that, some three years after 9 6.0 0.16 Conduct of a follow-up census in selected rural 10 2.3 0.06 LTR was formally completed, formal markets for Note: The price of Kigali is from official registry villages could provide an opportunity to gain a residential land in Rwanda have been well for the period 2013-2015 on about 25,000 better understanding of the true extent of parcels. The exchange rate was 647, 682 and entrenched and, with more than 5% of transfers 720 Rwanda Franc per USD for 2013, 2014 informality as well as ways to properly address it. per year, are very active by global standards. This and 2015, respectively. The data for the rural areas is from the 2015 World Bank survey of supports the notion that Rwanda made significant 3600 households. For this analysis, 10% of progress in addressing the challenges of rural parcels with reported prices less than the official registration fee were dropped. subsequent registration that have undermined the sustainability of many land registration programs globally. Table 10: Willingness to pay for registration - LN (UGX) Combining administrative and household survey Household survey data allows us to show that, while a significant cleaned with LAIS - Household survey share of sales transactions still remain informal, restricted to parcels cleaned with LAIS located in the cell of RNRA’s efforts at decentralization and provision residence of information, in particular placement of SLMs, Ln sale value of assets, USD 0.151** 0.134* have helped to reduce levels of informality. Yet, (2.285) (1.807) further efforts in this direction will be needed to Head/Spouse knows about registration process 0.377 0.425* (1.598) (1.706) replicate the high levels of registration of sales Legal knowledge on registration, computed ind. -0.047 -0.070 transactions achieved in urban residential areas for (-0.664) (-1.028) rural land. Being able to do so be a precondition to Parcel purchased 0.050 0.101 firmly establish the land registry as a key part of (0.365) (0.660) Parcel allocated by Gov. -1.379*** -1.728*** (-4.514) (-5.903) Ln size of parcel, Ha. 0.114** 0.110** (2.417) (2.298) Share of parcels certified in Sector 0.364 0.320 (0.918) (0.780) Ln. dist. to DLO, km -0.213** -0.357*** (-2.237) (-4.367) SLM in place at time of int. 0.385 0.584** (1.408) (2.105) Land Week event 0.170 0.086 (1.254) (0.636) Constant 9.068*** 9.677*** (7.797) (8.131) Number of observations 447 377 R2 0.177 0.209 Source: World Bank 2015 third round follow up survey. These regressions includes regional fixed effects as well household and parcel characteristics that are not reported. Note: *** p<0.01, ** p<0.05, * p<0.1 This case study was prepared by a team from the Development Economics Research of the World Bank in collaboration with the Rwanda Natural Resources Authority. The UK Department for International Photo credit: RNRA Development, the Bank Netherlands Partnership Program and the International Growth Centre provided generous financial support. The findings, interpretations, and conclusions expressed are entirely those of the authors. They do not necessarily represent the views of the Government of Rwanda, and of the World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.