Behavioral Designs for the Urban Productive Safety Net Program (UPSNP) - Ethiopia

The objective of ideas42’s partnership with the Government of Ethiopia and the World Bank is to uncover the behavioral barriers that beneficiaries face in taking the necessary preparatory steps to help start or expand their business, and to use these insights to design scalable solutions that address the identified barriers. This study aims to test the effectiveness of the behavioral designs at promoting desirable beneficiary socio-economic outcomes.

As part of the UPSNP, beneficiaries are provided financial support (business grant) and training to help them start their chosen business. However, key-informant interviews with beneficiaries, UPSNP staff, and representatives from the Government of Ethiopia revealed that while beneficiaries have the intention to start or expand a business, many did not appear to actively think about the preparatory steps needed to start/expand nor act on them in any way. These steps include conducting background research (Assessing strengths and weaknesses, market opportunities and threats), selecting business choice based on background research (if starting a new business), building relevant skills, creating a realistic business plan, and allocating adequate capital (business grant and savings) for business.

In order to understand why beneficiaries were not completing the preparatory steps, ideas42 collaborated with the Government of Ethiopia and the World Bank to conduct individual interviews and focus groups discussions with beneficiaries and program staff in two cities in Ethiopia (Addis Ababa and Adama) over the past couple of years. These activities helped us understand the various decisions and actions that beneficiaries face in taking the necessary steps to help start or expand their businesses. We further explored features of the program and the beneficiaries’ environments that created barriers to achieving their stated goals and aspirations. Data was collected using an interview guide with the help of local translators. The questions in the interview guide tried to understand the following areas:

• how the UPSNP’s public works program is framed to beneficiaries, program details, and beneficiaries’ on-boarding to the program and their current experience with the program
• beneficiaries’ savings and spending habits
• how beneficiaries choose a business goal and their mindset and motivation for that goal
• the business training provided by the government and business plan elements
• how beneficiaries think about executing on their business plan to start or expand a business

From the field visits, ideas42 identified five key behavioral barriers contributing to beneficiaries not taking preparatory steps (before getting their business grant) to start a sustainable business:
• Lack of agency: Some beneficiaries are heavily reliant on government support, and don't believe they can be successful in their business without additional support such as a free work space from the government
• Incorrect mental models: Beneficiaries think that they cannot start working on their business until they get the $500 business grant despite program’s promotion of starting the business between the second and third year.
• Inattention: Beneficiaries do not always consider all the relevant factors (e.g., right skills, experience, market conditions) when selecting their business. For example, they don’t think about the market demand of the business they select and whether it is already oversaturated in the market (i.e., injera making).
• Uncertainty/lack of clarity: Beneficiaries do not know when to start executing on their business plan. They are unclear of various steps (both monetary and non-monetary) they need to go through to start their business. (A monetary step could be that beneficiaries need to make sure their initial investment/capital requirements do not exceed the business grant and their 20% savings, while a non-monetary step could be beneficiaries outlining ways to reach their customers such as calling over the phone or foot marketing.)
• No social proof: Beneficiaries do not see others like them following the preparatory steps to start or expand their business.

By helping beneficiaries to overcome these barriers, we have the potential to make a significant impact on the behavior of beneficiaries to help them successfully start or expand their business in a low-cost and light-touch way.

The intervention will consist of a package of the behaviorally-informed designs aimed at helping beneficiaries to translate their intention of starting/expanding a business into a reality by completing the preparatory steps needed to start/expand a business in a timely manner. The package of interventions will include:
a. A poster which has been behaviorally designed to showcase beneficiaries who have started a business while in the program and how they started it, and another poster which provides testimonials from other beneficiaries on their experience launching/expanding their business. The purpose of this design is to make "role models" that were part of the program and successful in launching/expanding their business visible to beneficiaries in order to show them that there are people like them who have accomplished their business goals and motivate/inspire them to follow the necessary steps to launch/expand their own business.

b. An exercise on self affirmation aimed at reframin the safety net program to create the mindset that both beneficiaries and the government bring resources and have responsibilities to the program (and combining those resources results in success). The design also primes beneficiaries to think of their positive skills and values they can bring to their business.

c. An activity to help beneficiaries re-evaluate whether they have chosen the appropriate business and help them make an informed decision about the type of business to start, and to help them identify what gaps they need to fill to be able to launch/expand their business of choice and prompt them to take those actions.

d. A tool to help beneficiaries to think about the non-monetary steps and actions they can take to start the business (while still in the program and before they receive their business grant). This design also helps beneficiaries identify which steps they can take towards launching/expanding their business before receiving their $500 business grant and 20% savings.

2021-11-22

2022-05-31

1. Do beneficiaries feel a sense of agency

Self-affirmation & agency:
• % reporting confidence in ability to start/expand business within the duration of the UPSNP program;
• % reporting various expectations of government support during and after program (e.g. help finding workplace, training, etc.);
• % reporting that beneficiaries have various responsibilities (e.g. attend training, work hard, etc.)

Steps taken to start/expand business:
• % reporting having decided which business to invest business grant to;
• % reporting wanted to invest into different types of business (e.g. selling food, tailoring, livestock farming, etc.);
• % that report various processes for selecting business choices (e.g. Seen others do the same business, experience working in business, having required skills, observed demand for product/service, etc.);
• % reporting various factors that they believe are important to consider when choosing a business (e.g. experience, skills, advice of others, etc.);
• % reporting various ways to respond to gaps in business preparedness (e.g. choose a different business, learn the needed skills, do nothing);
• % reporting having attended the government business training;
• % who believe the best time to start preparing towards business is before receiving grant vs. % that believe the best time to start preparing is after receiving grant;
• % that believe that there are things they can do to start preparing for the business vs. % that believe that they can do nothing but wait to receive the grant;
• % that are taking steps to prepare for their businesses while they wait to receive their grant;
• Number of hours spent preparing towards starting business;
• % reporting various challenges faced in preparing towards starting business (e.g. not having enough time, insufficient funds, etc.);
• % reporting having secured a workplace for business;
• % reporting having identified potential customers for business;
• % reporting having developed a plan for reaching potential customers;
• % reporting having determined price of goods/services

The evaluation will be done through an individual randomized controlled test (RCT) in which the treatment group will receive the full intervention package while the control group will receive no intervention.
As part of the UPSNP public works program, beneficiaries participate in business training sessions in the second year of the program. These training sessions occur at the woreda or kebele level in each city. Before the business training begins, a local research firm called Laterite will conduct a baseline training of the treatment and control groups. After the baseline data collection, beneficiaries will be randomized into treatment and control groups, making sure the sample is balanced on key demographic and socio-economic criteria. After the business training, treatment group beneficiaries (800) will be invited to the TVET training center to receive the behavioral interventions. The interventions will be delivered in small groups of 25-30, in keeping with the government training protocols as well as to ensure that safety precautions (sanitizers, mask wearing, and social distancing) are taken against any possible risks of COVID-19 infections.
Data Collection: Data collection will occur at two stages.
• Baseline survey data will be collected before the intervention.
• Endline survey data will be completed four to six months after the delivery of the intervention.

Methodology: Data collection will be done by direct interview of beneficiaries on the basis of a standard and pre-coded survey. The surveyors will use tablet computers (electronic questionnaire) instead of a traditional paper questionnaire and data will be collected using Computer Assisted Personal Interviewing (CAPI) software, specifically SurveyCTO. Data collection will happen at the public works sites in Adama. Local woreda staff and beneficiaries will be informed ahead of time about the data collection that is to take place at the selected site through in person visits and phone calls. On average, the duration of the in-person survey collection is estimated at 20 minutes. A trained enumerator from the hired local survey firm, Laterite, will ask the questions on the survey, prompt oral responses from the participant, and collect the responses on the tablet. The modules in the survey will include:
• Individual & household demographic information
• Employment details
• Steps taken to start or expand business
• Response to intervention (for treatment subjects only at endline)
• Additional questions for gauging potential spillover effects

Location: The interventions will be conducted in Adama, Ethiopia. Adama is a representative city that is easy to access as well as has high implementation feasibility.

Beneficiaries who are already in the UPSNP will be randomly divided into treatment and control groups. Random selection will be conducted using a statistical software such as STATA, based on a list of beneficiaries obtained from the government. Only beneficiaries who are selected into the treatment group will be invited to participate in the behavioral intervention. The interventions in the study will be delivered in classrooms at the TVET center in Adama, which is where the government typically organizes training for beneficiaries. Both treatment and control groups, however, will be interviewed at the baseline and endline. Baseline and endline interviews will be conducted at the public work site where beneficiaries gather to work on their assigned community projects.

Randomization will be conducted using a statistical software program called STATA

Individual

No

The randomization will not be clustered. The sample size is 1,600 UPSNP beneficiaries

1,600 UPSNP beneficiaries

800 treatment, 800 control

Using the conservative base rate estimate of 0.5, we calculated the minimum detectable effect of .07 (power=.80, alpha=.05) with a sample size of 1,600.