The demand for and the impact of formal property rights in Dar es Salaam

There were two levels to the intervention, a block-level intervention and a parce-level intervention.

In two communities, parcels were divided up into contiguous `blocks’ of 40-50 parcels each. Parcels in treatment blocks and their owners were:

Provided with a free cadastral survey, demarking the boundaries with cement beacons.
Invited to local meetings run by a local NGO (the Women’s Advancement Trust - WAT) to discuss their involvement in the project and the benefits of obtaining a land title.
Invited to apply for a Certificate of Right of Occupancy (CRO) for their land, with support from WAT, for a heavily-subsidized price of 100,000 Tanzanian Shillings (TSh), to be paid for over a period of 5-6 months.

Parcels in control blocks were still able to obtain a CRO, but did not receive a cadastral survey, any support or subsidization. In addition to this block level randomization, parcels within treatment blocks were allocated vouchers which further discounted the cost they would pay for a CRO. These vouchers could take on values of 20, 40, 60 or 80k TSh. There were two voucher types: one could be applied to the cost of a CRO without condition, the other could only be applied if, during the CRO application process, the household included at least one woman as a co-owner of the property. A parcel could be allocated both types of vouchers at the same time. Names of households that chose to apply a conditional voucher were checked at the time of application to ensure their compliance.

Following these two interventions, owners of parcels in treated blocks were allowed to sign up and begin repaying at will.

Note: due to several adverse circumstances, there was no actual distribution of CROs following the application process. The Tanzanian government decided to raise the price of CROs being charged to the project, drastically increasing the cost of providing them. Second, following the initial sign up phase, the Government of Tanzania declared a large proportion of the study area to be ineligible for land titling

2010-10-15

2012-07-01

Primary outcomes of interest for the block-level randomization include:
Household investment in the parcel
Perceptions of the parcel’s risk of expropriation, land disputes and actual expropriation events
Access to credit
Perceived sale price
Household level economic outcomes (income, assets, employment, enterprise activity)

These outcomes were measured at baseline, but no follow up was conducted due to the suspension of the block-level intervention (see above).

Primary outcomes of interest for the voucher-level randomization included:
Whether the owner of the parcel fully paid for a CRO through the program.
Whether the household included at least one woman as an owner of the parcel.
Several proxies for women’s empowerment (self-reported measures of control over money, assets and decision making; expenditure on children)

The first level of the intervention was block-level randomization (within each of the two communities). Parcels were divided into contiguous blocks of 40-50 and randomly allocated to treatment and control groups. Re-randomization was used to achieve balance on several pre-intervention characteristics.

The second level of the intervention was, within treatment blocks, a parcel level randomization. Parcels were allocated vouchers which further discounted the cost they would pay for a CRO. These vouchers could take on values of 20, 40, 60 or 80k TSh and could either be unconditional or conditional (see above). The randomization method is described below.

First level: done in office by computer

Second level: a combination of computer and public lottery using the following process:

For each block, 100 draws of general/conditional voucher distributions were drawn
We tested balance using several parcel-level pre-treatment characteristics (using average t-stat values) and kept the three most balanced draws.
Draws were presented as three folders to attendees of block-level community meetings. One of the attendees was nominated by the community to choose one of the three folders (draws) and the result was the final allocation.

First level: block
Second level: parcel

Yes

Planned Number of clusters
Block level randomization:
Community 1 - 20 blocks
Community 2 - 30 blocks

Planned Number of Observations Block level randomization: Community 1: 919 Community 2: 1,362 Parcel level randomization: Community 1: 463 Community 2: 683

Sample Size (or number of clusters) by treatment arms
Block level randomization:
Community 1 - 10 treatment, 10 control
Community 2 - 15 treatment, 15 control

Parcel level randomization:

* Community 1 *
Unconditional |
voucher | Conditional voucher value
value | 0 20 40 60 80 | Total
------------------+-------------------------------------------------------+----------
0 | 24 40 42 36 53 | 195
20 | 30 20 41 37 0 | 128
40 | 27 29 22 0 0 | 78
60 | 14 21 0 0 0 | 35
80 | 24 0 0 0 0 | 24
------------------+---------------------------------------------------+----------
Total | 119 110 105 73 53 | 460

* Community 2*
Unconditional |
voucher | Conditional voucher value
value | 0 20 40 60 80 | Total
------------------+-------------------------------------------------------+----------
0 | 46 52 32 53 56 | 239
20 | 55 42 47 34 0 | 178
40 | 42 40 53 0 0 | 135
60 | 37 36 0 0 0 | 73
80 | 58 0 0 0 0 | 58
------------------+-------------------------------------------------------+----------
Total | 238 170 132 87 56 | 683