Cash for Work, Behavioral Nudges, and Productive Activities

To make substantial improvements to their livelihoods—whether through material investment, accumulation of assets, or insurance against risk—Cash for Work beneficiaries must be able to put aside some of their transfer funds for future use. However, a number of contextual factors make this decision, and the action needed to carry it out, difficult. First, they must set some kind of goal for how much to set aside, and decide on a method (such as a safe place for storing funds) for doing so. The program gives them little basis for setting money aside, as they are paid weekly and never actually encounter the full amount of money they stand to earn over the full 80 day period of employment.

The impact evaluation tests two interventions:
i) Village Savings and Loan Associations:
ii) Behavioral Nudges:

2017-12-10

2020-08-10

consumption, diet, food security, nutrition, productive and household asset accumulation, labor force participation, and agricultural productivity.

Assets: We are interested in four types of assets: moveable assets, savings, housing, and land. Amongst moveable assets we will examine the total number of livestock and farming assets. We will examine the amount of current savings of the households. We will also examine the ownership and quality of housing stock (e.g. building materials of walls, roof and floor) as well as the availability of electricity and sanitation facilities.

Consumption: We will examine impacts on food consumption as well as food and non-food expenditures. Specifically, food consumption measured as self-reported market value of self-produced food, as well as food purchased from the market in the 7 days before the survey data are collected. Non-food expenditures include transport, leisure, communication, personal care goods, school fees, household expenses (e.g. clothing and utensils), and social expenditures (e.g. baptism, funerals, donations and festivals).

Nutrition and Food Security: Here we will analyze the self-reported number of meals consumed by adults and children in the household. We will also look at the incidence of non-availability of any food in the household, the necessity to borrow food from friends and relatives and the need to reduce the number of meals eaten in a day and a measure of dietary diversity.

Agricultural outcomes: We are interested in total production per hectare, total expenditures on inputs and seeds and the number of crops cultivated.

Household income: We will examine variables related to household income from self-employment, own business, livestock sale, agricultural products sale, and other sources. We want to explore the different income sources of the household and the amount received.

Coping with shocks: We will examine if the household experiences at least one type of shock in the last 12 months and which type of shock. Within the possible shocks we include: drought, flood, death of a family member, and asset theft, loss or damage. We will also assess resilience from the shock.

Women’s outcomes & family wellbeing: Within this section we are interested in studying the following categories: happiness, economic satisfaction, and decision-making power.
We will also analyze the education of children 5 years and above. For this, we measure the school attendance of any/all children in current school year.

Aspirations: We will examine parent’s aspiration for their daughters and sons and assess if there is any difference between boys and girls. For this, we see aspiration on final educational level, future occupation and the desired age of marriage.

A mixed methods approach was selected for the evaluation.

At the first level, 4 comparison regions were selected, chosen to be as similar as possible to the study regions. Within the 4 comparison regions, respondents were selected using the same community-based targeting technique as was used in the program regions.

At the second level, within the 4 study regions, the sample was randomized into three groups at the level of the payment site which is the level at which interventions were delivered. One group got just the cash for work program, the second group received cash for work plus behavioral nudges, the third group received cash for work and training and support on creating savings groups.


A difference-in-difference allows us to identify the overall effect of the ACT-P program on beneficiary households inside eligible villages since the treated households and non-treated households will be in different villages.

Since Training and Nudges were randomly assigned to working groups, this allows us to examine the impacts of ACT-P when it is complemented with training and behavioural nudges.

office by a computer

payment site

Yes

104 clusters

5000

1000 households : ACT-P
1000 households: ACT-P + nudges
1000 households ACT-P + VSLA
2000 control households