Can graduation programs be greener? Evidence from the BOMA program

The Ultra Poor Graduation approach is a 24-month program that provides beneficiaries with a holistic set of services including livelihood trainings, productive asset transfers, consumption support, savings services, basic health services, and life skills coaching. The BOMA Project has adapted the graduation approach for the arid and semi-arid lands (ASALs) of Africa with their Rural Entrepreneur Access Project (REAP). As in other graduation programs, communities nominate their poorest, most vulnerable members to participate in the program. Participants receive financial and life skills training and mentoring to manage and grow their business. They also contribute to a savings pool, which they can access as needed to cope with shocks or to invest in expanding their business. REAP is delivered in groups both for savings and business: REAP participants form small business groups, draft a business plan, and receive a cash grant to start a small business.

The “Green” REAP program was created as a way for communities to minimize the negative environmental impact of business activities, spur positive engagement with the environment, adapt to the climate crisis and mitigate the impact of climate change on the community. Green REAP establishes environment-neutral or environment-supporting businesses, “green” businesses, as a pathway out of extreme poverty. Further, community resources management groups (community forest associations, natural resource management groups, etc.) in Green REAP villages will receive additional training in how to manage forest and pasture resources for climate resiliency. With Green REAP, the BOMA Project seeks to simultaneously move households out of extreme poverty and create positive environmental impacts.

2023-09-01

2025-03-01

The primary research questions for the project are:
1. Is Green REAP able to maintain (or increase) the poverty and welfare impacts of REAP while achieving environmental benefits?
• Key outcomes: measures of poverty and welfare such as assets, household income, food security, and school enrollment, among others
2. Does Green REAP change environmental knowledge?
• Key outcomes: a module that tests knowledge of participants, especially those that they learned in the forest schools, e.g. knowledge about climate change, the interaction between humans and the environment, or how harmful resource extraction can be for the forest
3. Does Green REAP change environmental attitudes?
• Key outcomes: In contrast to knowledge, attitudes are beliefs formed about a topic, questions will include for instance: Do you think people should cut down trees to produce charcoal or recycle waste (“normative” questions)
4. Does Green REAP change environmental (business) practices? Does participation in Green REAP encourage households to engage in environmentally friendly businesses such as beekeeping or gum and resin harvesting? Does it shift existing businesses to more sustainable practices? And are environmentally harmful businesses abandoned?
• Key outcomes: indicator and share of income from beekeeping, tree nurseries, gum or resin harvesting, or other lower-environmental impact alternatives. For existing business practices, adoption of climate-smart agricultural technologies, and use of methods to reduce extensive grazing pressure. Indicator and share of income from businesses harmful for the environment such as charcoal production and amount and type of resources extracted from the forest and share of income from these resources.
5. Does Green REAP change environmental (non-business) practices?
• Key outcomes: individual actions to mitigate or adapt to climate change (e.g., tree planting, water storage), participation in community actions to protect forest (e.g., community resource management group, forest management activities), intensity of rules governing forest, observable measures of resource extraction from forest (number of charcoal kilns, trees damaged, number of animals grazing, amount of land cleared for agriculture). Indicator of waste recycling type.
6. In the long term, is Green REAP effective at increasing biomass and reducing deforestation?
• Since we agreed that we are under-powered for a reasonable tree cover and biomass analysis, we will not collect direct measures along these dimensions. However, there might be the possibility in the future with improved satellite coverage to examine long-term effects
7. How does Green REAP compare to other environmental interventions in environmental impacts per dollar spent on the program?

This is a multi-arm clustered randomized evaluation in 105 clusters with the following arms:
1. Standard REAP: 35 clusters
2. Green REAP: 35 clusters
3. Control: 35 clusters
Our evaluation will be randomized at the community level. Both treatment and control households will be screened through the same selection process to ensure the groups are comparable. The evaluation includes the following steps.
1. BOMA Project screens households in treatment and control clusters and provides a list of eligible households to IPA.
2. IPA conducts baseline survey with sample households in all study communities, plus additional data collection as needed
3. IPA randomly assigns each geographic cluster to one of the selected treatments, or to control, using a computer
4. BOMA Project launches implementation
5. IPA conducts a midline survey to capture short-term changes in participant activities such as business launches and incorporation of environmental components
6. BOMA Project concludes intervention
7. IPA conducts endline socioeconomic surveys with households and businesses

We will sample a small number of ineligible households in clusters in order to identify spillover effects.

Not available

On a computer

Cluster

Yes

105 clusters

3,465 women over two survey rounds for a total of 6,930 observations. Most surveyed individuals will be women who are direct beneficiaries.Environmental spillovers will be measured at endline on ineligible individuals within green and non-green treatment groups.

35 control, 35 standard REAP, 35 Green REAP

Using an intracluster correlation (ICC) of 0.05, the 80%-power minimum detectable effect (MDE) is approximately 0.20 standard deviations for continuous variables or .13 percentage points for binary outcomes (at a control mean of 50%).