Abstract
Overview
This study explores the effects of guaranteed output markets and access to timely tractor services to understand how these factors affect farmers' investment, cultivation patterns, technology adoption, and output through a randomized impact evaluation of the Savanna Farmers Marketing Corporation’s (SFMC) program in Northern Ghana.
SFMC is a private company that has operated in Northern Ghana since 2004. SFMC started its activities with the idea of creating a marketing company, managed by local experts, that profitably links small-holder farmers in remote rural areas to output markets. Most of farmers in Northern Ghana are subsistence oriented, mainly producing for their own consumption. With SFMC’s intervention, the company gives the opportunity to smallholder farmers to shift from subsistence to market-oriented farming, creating conditions to allow for both higher profits and better living conditions. In the past 12 years, SFMC has expanded its services across three regions in Ghana. SFMC’s main goal is to improve accessibility to output markets by linking smallholder farmers to agricultural value chains.
At the beginning of each farming season, SFMC visits soya-producing communities in Northern Ghana and indicate that they plan to return to purchase crops. Later in the season, they declare a minimum price, and return following harvest to complete the transactions. In addition to providing marketing services for farmers, SFMC offers some communities a tractor services contract that allows farmers to access timely tractor services with payment in-kind after harvest. In some communities, SFMC provides additional services to farmers who agree to sell to the company, including agricultural extension and a program that provides tractor services to farmers on credit. We randomly assign communities to three experimental groups in which communities respectively receive purchasing offers, purchasing offers + tractor services, and no services.
Additionally, the study includes an extension overlay, in which half of the communities within each of the experimental group are randomly assigned to receipt of SMS forecast updates and agricultural videos on best farming practices. The study’s design and focus on soya—a crop that is commonly grown by women in Ghana—further allow us to disaggregate effects by gender and explore gender-related outcomes.
For this study, the intervention is led by the implementing partner, Savanna Farmers Marketing Company (SFMC).
SFMC completes the pre-intervention activities for this study in the following stages.
Pre-intervention activities:
- SFMC selects new communities they would be willing to provide services in for the upcoming 2019 farming season.
- SFMC makes this selection based on the soya producing potential of the community.
- After community selection, SFMC conducts initial community entry in the following 3 stages:
(1) Introductory meeting with community chiefs and elders
(2) Community-level “sensitization” event in which SFMC describes its services and informs community members that the program may expand to this community for the upcoming farmer season.
- SFMC prompts interested farmers to sign up on a list.
- During this community wide meeting, a lead farmer is elected by the community members.
- SFMC tasks the lead farmer with collecting the list and mobilizing individuals to sign up.
(3) SFMC returns to the community and collects the list of interested farmers for the upcoming farming season. Before SFMC retrieves the list, the lead farmer and elders within the community, are tasked with vetting the list and are instructed to include members who have a good history of repayment and are trustworthy.
Main Intervention Activities
This study tests two interventions that SFMC will implement, Treatment 1 (T1) and Treatment 2 (T2):
T1– Offer of marketing MOU for soya. SFMC will make an offer to the community with a guarantee to farmers to purchase a set quantity of soya at harvest at a particular minimum price.
T2 – Offer of marketing MoU for soya and offer of tractor contract. SFMC makes an offer to purchase soya at harvest to the SFMC-formed FBO in the community. Additionally, SFMC makes an offer for a tractor contract to the SFMC-formed FBO. SFMC then collects orders for tractor services from farmers within this FBO, identifies local tractor owners willing and able to provide tractor services in the community, coordinates the timely delivery of tractor services and confirms that the tractor services were completed. The purpose of the tractor contract is to efficiently match tractor supply and demand, reduce disagreements over plot size, and loosen credit constraints on tractor-related investments. The tractor contract includes the following terms:
o Farmers agree to make payment for the tractor service in-kind at harvest
o SFMC pays the tractor owner for services rendered in the community at the time of land preparation and verifies that the services were completed
o SFMC collects in-kind payment, the debt incurred by the farmers who signed the tractor service, at the time of harvest
Additionally, some communities will receive a treatment consisting of weather forecasts delivered by SMS and agricultural extension videos.
Primary Outcomes:
o Yields
o Productivity/profitability
o Income/consumption
Secondary outcomes
o Changes in investment patterns
Investment in soya
Share of soya in crop portfolio
o Tractor plowing during land preparation
Use of tractors
Timing of tractor plowing
Cost of tractor plowing (price/supervision/search costs)
o Knowledge/beliefs
Agricultural practices – decision-making and self-report behaviors
Final outcomes of interest
Experimental Design
In the 2019 farming season, SFMC will enter approximately 200 new communities across 6 districts in the Northern Region of Ghana.
Following SFMC’s completion of community entry procedures, researchers will implement an in-person baseline survey and randomly assign communities to three experimental groups. The researchers will randomize the roughly 200 communities into three experimental groups corresponding to the two treatments explained in Section 4 and a control group receiving no intervention: Treatment 1 (T1, roughly 65 communities), Treatment 2 (T2, roughly 65 communities), and Control (C, roughly 70 communities). Following randomization, SFMC will implement its two treatments in the T1 and T2 communities respectively. Control communities will be notified that they will not receive SFMC services for the 2019 farming season, but will be under consideration for future seasons.
SFMC will implement T1 and T2 in the assigned communities for two consecutive agricultural years. A brief phone midline survey will be conducted in the middle of this period. Following the end of treatment, a detailed in-person endline survey will be conducted.
• Index 1: size of community (number of compounds )
• Index 2: market access (time to nearest market, transport cost of soya, transport cost of maize)
• Index 3: tractor access (tractors per compound, price of tractor ploughing per acre multiplied by -1)
• Index 4: wages (cost of one person-day of weeding)
• Index 5: number of farmers in pre-vetted FBO