1. Collective Property Leads to Private (Household´s) Investments: Lessons from Land Titling in Afro-Colombian Communities María Alejandra Vélez Ximena Peña Juan Camilo Cárdenas Natalia Perdomo Universidad de Los Andes 2016 WORLD BANK CONFERENCE ON LAND AND POVERTY The World Bank - Washington DC, March 14-18, 2016

2. Larson and Soto (2008); Agrawal 2007; Larson, Barry and Dahal (2010) Collective Property in Developing countries 27% of forest in developing countries is common property 33% in Latin America More than 30% of Colombia's territory

3. Property Rights & Poverty The focus for poverty reduction programs has been on assigning individual (private) property rights Efficient use of resources: residual claimants can make long-term decisions Transform wealth into capital via access to credit (De Soto, 2000) Investment in physical and human capital (Galiani & Schargrodsky, 2007) Unproductive land would move to the hands of more productive labor (landless farmers interested in investing) The channels through which collective titling may decrease poverty are not the same

4. Collective Land Titling in Afro- Colombian Communities Afro-Colombian communities have been settled in the Pacific region since 17th century Until 1991 the Colombian government viewed the Pacific region as a massive forest reserve of unoccupied lands Article 55, 1991 Colombian Constitution Law 70 in 1993

5. Collective Land Titling in Afro- Colombian Communities Benefit the groups who historically occupied these territories and to preserve valuable ecosystems Community Council as a political and territorial organization The first title was issued on 1996 5,3 million hectares & 168 Community Councils Collective titles are inalienable and indefeasible, which, among other things, prevents the use of the land as collateral There are no systematic evaluations of the effects of this process on households’ wellbeing!

6. Channels for Poverty Reduction (i)Strengthening the sense of ownership may extend the time horizons in which households plan their individual investments (ii)Strengthen the consolidation of organizational structures and their role in the political landscape of the region might increase their access to productive and aid projects. (iii)Strengthening the sense of belonging to the community implies a better use of cultural values and social networks to provide informal social insurance to respond to different natural and economic shocks iv) The possibility of taking advantage of the economies of scale

7. Methodological Challenges Lack of information and no baseline data Complex economic, political, cultural and social dimensions encompassed in the process Several parallel external factors: the presence of illegal armed actors, the internal forced displacement, drug trafficking, gold rush and institutional changes. Finding an adequate control group is challenging

8. Methodological Challenges Land titles were not randomly assigned No exogenous scaling of the process To this date almost all community councils have been granted a title. Communities with and without collective titles may have had different levels of human or social capital before the titling, or different organizational structures

9. Empirical Strategy The titling process has been taking place over the last two decades We use information from the second (2003 and 2008 Baseline) and the third waves (2009 and 2013: Post- Treatment) of the SISBEN Treatment: Rural districts located within a community council that had received the collective title before the baseline data collection

10. Empirical Strategy Restrict our analysis to the Chocó region ( 56.87% of collective titled in Colombia) Our final sample: 458 rural districts with 25.000 households Treatment: 60% of titled rural districts Control group: 40% of rural districts where no collective titling occurred at baseline Outcome variables: Income per-capita Housing quality Children’s educational attendance

11. Empirical Strategy We use Analysis of Covariance (ANCOVA) We estimate the following model for rural district i, in municipality j, observed in the period t:

13. Model 1: includes only the rural district fixed effects and the number of years between the second and the third wave of the census. Model 2: extends Model 1 by adding all the pre-treatment covariates Empirical Strategy

15. Summary of Results Collective titling has significant effects on household per-capita income, housing investment and educational attendance of children between 5 and 15 years old We conclude that collective titling has improved the quality of the housing and living conditions for the black rural communities of the Chocó department We report similar results as those discussed as benefits of individual titles by Galiani and Schargrodsky (2010) These conclusions are robust to heterogeneous effects: coca, mining, displacement

16. Conclusions Access to collective land have created a more secure resource base and a longer time horizon for these households, which led to investing in their private physical and human capital The results contradict the claim that without private ownership of the land there are no incentives to invest in the productive assets at the household level Our results suggest important policy lessons for the continuing and improvement of these land titling programs in the developing world Contribute to the ongoing debate on rural poverty reduction, land reforms and development

20. Individuals Households Rural districts SISBEN II SISBEN III SISBEN II SISBEN III SISBEN II SISBEN III Rural districts in the department of Choco 149,475127,56133,48533,071961933... inhabited mainly by afro- colombian population 144,650127,03832,84432,714930 … with households consisting of at least one member aged 12 or older 143,091126,32831,89632,569926925 … without duplicated records in the database 121,43931,327809 … formed by no less the five households 142,313120,50231,69031,033773673 and surveyed in both waves of the census 113,74794,22025,28825,173474

21. (1) Household Per-capita Income (2) School attendance rate- Children between 5 and 10 years old (3) School attendance rate- Children between 11 and 16 years old Model 1Model 2 Model 1Model 2 Model 1Model 2 Treatment 13718***12182*** 3.4536.417** 2.1862.572 (3816)(4172) (2.566)(2.879) (2.385)(2.424) Municipalities with a larger area under coca cultivation -18515***-3835 -0.8440.996 0.5405.606 (5782)(7446) (3.896)(4.500) (3.387)(4.335) Interaction 14343*9822 -0.4880.302 2.752-1.691 (7772)(8422) (4.645)(4.905) (4.015)(4.492) Constant 30136***38685*** 62.50***55.68*** 50.14***48.39*** (9873)(11615) (5.487)(6.076) (6.287)(6.281) Observations 458 446 442 R-squared 0.1070.217 0.0240.076 0.1910.233 (4) % HH with low-quality walls (5) % HH with dirt or sand floors (6) % HH living under critical overcrowding Model 1Model 2 Model 1Model 2 Model 1Model 2 Treatment -3.756***-2.535*** -2.026*-1.535 -1.8780.227 (0.931)(0.852) (1.222)(1.378) (1.344)(1.385) Municipalities with a larger area under coca cultivation -3.422***-0.495 3.5423.513 1.2310.328 (1.089)(1.432) (2.757)(2.555) (2.392)(2.411) Interaction 3.250***1.866 -4.736-4.160 -7.219***-5.841** (1.229)(1.413) (2.885)(2.658) (2.760)(2.689) Constant 4.335**4.310**5.320*3.89320.29***21.52*** (1.882)(2.154)(2.772)(2.939)(2.942)(3.573) Observations 458 R-squared 0.1700.2130.7020.7320.1570.252 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1