How can we keep getting vital farmer input to index insurance design while also expanding the insurance products to reach many more people? It’s been a lingering puzzle, but IRI staff have developed a new tool to improve efficiency in data collection — a key step to finding a solution.
Subsistence farming of maize, groundnut and millet dominates most livelihoods in Tambacounda, a region in eastern Senegal about the size of Denmark. The farmers are mostly from the multi-ethnic groups of Wolof, Pular and Serer, who have historically settled in the region and have adopted farming as their main occupation. Most farmers in Tambacounda don’t have irrigation, so their crops are dependent on rainfall, known as rain-fed agriculture.
If you visit during the dry season, you are most likely to see farmers gathered under the shade of a big eucalyptus tree, taking a well-deserved break after the harvest. They are shucking peanuts and drinking tea (called ataya) while chatting about this year’s yields of corn and millet. Gathered in bunches are cobs of maize that cover their low hut roofs, where they leave them to dry once the rainy season is over.
Tambacounda is one of the regions involved in the R4 Rural Resilience Initiative (R4), established in 2013 by the World Food Programme and Oxfam America to help build local farmers’ resilience to climate-related shocks. The initiative uses a range of risk management strategies. The International Research Institute for Climate and Society is a technical partner in R4’s micro index insurance component, developing indices that use rainfall measurements to determine when farmers receive payouts.
In index insurance projects such as the one in R4, data collection — the purpose and process of which are further explained below — is key in designing robust indexes that reflect the reality of farmers’ experiences of drought.
Designing robust indexes has always been our priority as members of IRI’s Financial Instruments Sector team. We revel in putting pieces of the puzzle together, turning advanced, complex science computations and models into simple, transparent and user-friendly weather-based indices. The CGIAR Research Program on Climate Change, Agriculture and Food Security, or CCAFS, shares our passion in this matter and is well aware of the importance of this process, which is why it decided to fund IRI to kick-start a new phase of index insurance data capture. We designed a new tool and tested it in late 2016 with our team in Tambacounda. Here’s why, and then how, this transformational step in data collection happened.
The problem: Data collection is critical, but also resource intensive
The strength of index insurance products over traditional indemnity-based contracts (see definition below) depends on (1) the effectiveness of the design to capture the reality on the ground and (2) efficient payment calculation and processing. These two differences give index insurance the advantage of significantly reducing costs and making scaling to many more farmers possible
[Indemity-based contract: A product that insures against one or multiple perils, and for which financial compensation of losses is processed after the damage occurs and is physically assessed, usually conducted by an insurance representative in the subject field]
However, an index insurance product is bound to face some challenges for it to reach the numbers of farmers needed to meaningfully impact poverty, and become financially viable. And in this scenario, people’s livelihoods are always at stake. These are the challenges that we lose sleep over.
One major obstacle lies in the absence of effective, systematic tools that can help streamline the capture and flow of information. This includes things like methods of data collection, sign-up and payment, and how to communicate with farmers about insured periods and payouts.
The IRI index insurance team conducts traditional data capture and collection through a series of field visits at different stages of the index design and validation. We are often joined on these visits by experts in climate science, insurance and agro-meteorology from key local organizations. As a team of index design and insurance representatives, we conduct the visits and gather the data, usually on clipboards and flipcharts. Once back to our respective offices, design team members analyze, compile and then share data with other partners.
The data isn’t immediately accessible for analysis with computers, however. In the picture below, our colleague Bristol Powell examines a color-coded farmers’ “bad year” table during a village visit to Chula, Malawi. The color-coded sticky notes are then transcribed into a dataset of bad years by phase (i.e. the period in which the drought happened during the season), and compiled with other datasets from previous visits to the same village. This ranking of farmers’ bad years is then compared with the satellite-based rainfall data for the corresponding pixel in Malawi.
The primary objective behind this process is to gather information in each village on farmers’ practices, their agricultural calendar and the drought years they’ve experienced. We repeat this every year in the continuous effort of refining the indexes and scaling to new regions.
But the process gets slowed down when index insurance projects expand and the number of villages multiplies. We wish we could afford having hundreds of Bristols in the field, gathering data all-year-round, analyzing databases for all villages of interest and generating diagnostics and reports for index design and refinement, all in a timely matter. But we can’t. These visits are time- and money-consuming, which can be a hindrance when data is still needed at a micro (village) level.
To overcome this challenge, CCAFS took initiative and financed our team to put in place and test a new prototype tool in the Tambacounda region — a vision finally crystallized.
This tool constitutes a package of automated forms that could be used by local implementers to facilitate the data collection and information flow between the different stakeholders (including farmers), a task that would otherwise be hindered by paperwork processing, slowing down the monitoring and validation processes on the ground.
CCAFS saw the groundbreaking value of such a tool, if it could be implemented in a systematic manner across index insurance efforts and beyond. Our team developed the prototype to suit the type of data we gather in the field and comply with the participatory activities we conduct with farmers. The automated forms contain a series of questions and a data entry system through which the user can submit the information captured. The submitted data are then automatically saved in a database at the IRI.
The test run
In November, Mélody Braun and I, both members of the Financial Instruments Sector Team, headed to Tambacounda to conduct a series of “End of Season Assessment” field visits in eight of the region’s villages. These end-of-season assessments include an interactive exercise that captures the historical bad years experienced by local farmers. We follow this exercise with a focus group discussion to gather information on the progress of the rainy season and crop performance in each village. On this trip, for the first time, the activities would be facilitated through the use of the new automated form tool.
Upon arrival to Senegal, we hopped in a 4x4 along with the R4 Design team to make the 6-hour journey to Tambacounda. This team includes local experts from key organizations collaborating within the R4 project, and they were eager to join us and lead the test run.
Will this tool be valued enough? Will it be well received by our partners, and could it eventually help solve the problem of data collection quality at scale?
Multiple scenarios of how this prototype will be received raced through my head as I sat in the Jeep, preparing for our meeting with the farmers.
As we proceeded from village to village, my initial anxieties began to dissolve. Overall, Mélody and I thought the test-run went better than we’d hoped. We successfully captured the information farmers shared during these visits and were able to submit them through the automated forms.
The success was especially encouraging knowing the multiple glitches that we had to deal with getting the forms ready for the field. This opinion was shared with our systems analyst, Igor Khomyakov, who requested live reporting from the villages on the prototype’s performance.
To help with refining the tool in the future, both teams documented the extent to which this tool can play a role in facilitating information capture of the participatory processes and activities conducted on the ground. We also recorded how the prototype can be improved to successfully replace traditional data capture and streamline data collection.
Financial Instruments Sector Team:
- The tool allowed for more uniform capture of data across villages (reducing the risk of missing data since implementers are going through the automated forms list of questions as they engage in the participatory processes)
- The data collection/submission was more time-efficient
- It reinforced the need to incorporate offline-accessible data collection forms to accommodate for the weak internet reception in the region.
R4 Design Team:
- The tool allowed for easier, more guided data collection and reduced processing time
- There is a need to adjust some details in the forms to reflect the local conditions on the ground
- There is a need to have the option to include anecdotal, qualitative data.
Interest by CCAFS in this initiative stems from the promising potential of this prototype to help launch an integrated, automated and quantitative seasonal assessment approach to pre-existing projects in the area and beyond.
This kind of tool can be developed to bring users through the process of generating seasonal reports, using a range of both intuitive and quantitative seasonal risk assessment, including participatory processes in communities, local expertise and remote sensing products. The tool will connect all of the above mentioned activities in an automated fashion starting from data capture to report generation.
At a more developed stage, the tool will then be used to connect the dots between CCAFS initiatives such as research on crop modeling, work around sub-seasonal to seasonal forecasts and participatory process methodologies (similar to that of the Financial Instruments Sector Team). It will also provide a platform to integrate CCAFS key outputs into a user-friendly interface.
When integrated in a larger streamlined online tool, it further can be used by CGIAR practitioners and partners for index design and validation, disaster risk preparedness and community-based adaptation strategies, particularly as projects go to scale.
My team and I are finally starting to catch glimpses of the possibilities this tool has laid the ground for. We now see a tangible way through which local teams, such as the R4 design team, can work with farmers at the massive scales needed to make a meaningful difference.