House for Two: Empowering Women with Property Rights

To test the impact of strengthening women's property rights on their social and economic empowerment, we exploit the lack of awareness about a co-ownership program implemented by the local government, which added married women residing in rural areas of Pune as co-owners of their residential houses if the house was owned by their husbands before the program. We randomize an information intervention at the household level in 57 villages in Pune. Out of the sample of houses in each village that is owned by the couple, we randomly assign houses into the treatment arm, where the couple is given a copy of their ownership document and told about the co-ownership program. In the houses that are assigned to the control arm no such information is provided.

2023-05-17

2023-06-30

The first key outcome variable is the change in beliefs of men and women about their own house ownership rights. The second set of primary outcomes relate to the impact on investments made in the house by the woman, the probability of taking a loan against the house, and women's perception of land security (de facto land rights). The third set of outcomes will relate to studying the impact on women's bargaining power in the household. We will study this through three different ways. First, we will create an index of women's empowerment. Second, we will study the impact on the wife's and husband's consumption on goods consumed by each one of them. Lastly, we will include an incentivised task game along the lines of Mckelway (2019) in our endline survey.

References:
McKelway, Madeline. Experimental evidence on the effects of women’s employment. Working Paper, 2019.

A women’s bargaining power index comprising of five questions will be created. The five questions that we use are the following:
1. Who in your family decides whether to buy an expensive item like a bicycle or a cow in the house ?
2. Who in your family decides whether to buy land or property?
3. Who in your family decides what to do if a child falls sick ?
4. Are you permitted to visit the homes of friends/relatives in neighboring villages to talk with them ?
5. Do you have to ask the permission of other household members to buy clothing/footwear/nailpaint/fake jewellery/ or anything else for yourself ?

For constructing the index, each survey question will be coded as a continuous variable from least empowered to most empowered, made comparable by normalizing each to have a standard deviation of 1, and then averaged.

We will collect data on intimate partner violence (IPV) to see if strengthening women's rights results in lower incidence on IPV. Impact on woman's savings decisions and expenditure on children will also be studied. We will also try to collect data on daily chores of the woman to see if there is any impact on how the woman allocates time across different activities in a day.

The study will employ the administrative list of addresses in each village as the sampling frame. This address list contains the names of the owners of each residential house in the village. From this list, first houses where both the husband and wife are the owners will be selected. From this smaller list the investigator will randomly assign houses to treatment (T) and control (C). For the houses assigned to the treatment group, the couple will receive the information intervention, while those in the control group will not be provided with this information.

The treatment assignment will take place before the baseline survey. After the baseline survey, the houses in the treatment group will be provided with the information. The information will be given to both the wife and the husband together.

As a part of the baseline survey, all the outcome variables will be measured along with other demographical variables. Post the intervention, after a gap of 4-8 months, an endline survey will be conducted.

Not available

Randomization is done in the office by the PI on the computer.

Randomization is done at the household-level.

No

2000 households (4000 individuals). However, given the difficulty in recruiting houses due to geographical and time-constraints faced by the surveyors on the ground, we expect the sample size to be between 1600-1800 households.

2000 households (4000 individuals). However, given the difficulty in recruiting houses due geographical and time-constraints faced by the surveyors on the ground, we expect the sample size to be between 1600-1800 households.

1000 households in control and 1000 households in treatment. But realistically, we expect the number of households in each arm to be between 800 and 900.