Sharing alternative data in agricultural value chains to unlock agri DFS for smallholder farmers

This blog explores how data generated by the rollout of digital payment and procurement solutions in the agricultural last mile can be shared to unlock agri Digital Financial Services (agri DFS) for smallholder farmers. It also identifies considerations that come into play in the sharing of data.

In low and middle-income countries (LMICs), commercial banks provide just three per cent of financial services currently available to smallholders. Other formal financial service providers (FSPs) like microfinance institutions (MFIs), rural banks and savings and credit cooperative organisations (SACCOs) are typically more socially minded and likely to be located in farming communities. They play a critical role in the financial lives of smallholders. Yet, a lack of data on farmer activities, incomes and credit histories limits their ability to predict future cash flows and extend financial products to farmers like loans and insurance. That is changing. Alternative data coming from the digitisation of agricultural value chains present opportunities for FSPs to bridge data gaps, speed up credit decisions and reduce the cost of serving farmers, who represent a significant untapped customer segment. This blog highlights the need for partnerships between agribusinesses or cooperatives deploying digital payment and procurement solutions in agriculture value chains, FSPs and digital payment providers, including Mobile Money Providers (MMPs), to support farmers’ financial inclusion.

The digitisation of agricultural value chains generates a range of alternative data points

Agribusinesses and cooperatives are increasingly deploying digital procurement solutions to address value chain inefficiencies that affect systems and processes in the trade of crops, gain greater control over their operations, and enable more transparency in their transactions. Often, digital procurement solutions integrate digital payments, for example via mobile money, to address the high cost and risk of cash payments and traceability systems that enable more visibility into food chains. Such digital procurement solutions capture a wealth of data that offer significant opportunities to support the creation of digital footprints for farmers and build economic identities that can unlock agri DFS. Data captured by digital procurement solutions include farmer profile, demographic, farm profile, production, financial, agronomic and capacity building data (figure 1). For example, the GSMA AgriTech Rural Loan proof of concept pilot that is enabling access to credit for vanilla farmers in Papua New Guinea assesses creditworthiness against seven criteria based on data captured on a digital procurement app by agribusiness field agents at the point of farmer registration. This includes farmer profile and farm location data, number of vanilla vines on a farm, procurement history with the agribusiness partner, alternative sources of income, loan training information (provided as part of the pilot), mobile phone and bank account ownership data.

Figure 1 Type of data and example data points

Data-sharing partnerships enable agri DFS

Following the generation of alternative data, the next step in enabling agri DFS for farmers is for agribusinesses and cooperatives to store this data in a secure location before sharing it with FSPs for credit scoring. Most of the data collected through the digitisation of agriculture value chains are saved in basic spreadsheets such as Microsoft Excel. This data can be stored on a physical hard disk or USB drive or virtually in the cloud. A growing number of digital agriculture implementers are using cloud services to store data instead of hosting in-house servers, which significantly reduces investment costs. For example, Koltiva, an agritech company based in Indonesia, is integrating global supply chains and business processes into cloud-based collaboration platforms connected through a suite of mobile and web applications on behalf of agribusiness clients. Storing data in the cloud enables farmers to access all the data collected and held on them by Koltiva via an app.

The decision of where and how to store data takes into account three key considerations:

  1. The business value of the data. Of all the data that agribusinesses and cooperatives collect using digital procurement solutions, some types of data are more critical for FSPs for credit scoring and need to be accessed and shared more frequently than others. For example, farmer production and financial data need to be updated regularly for credit scoring purposes. Storing this data in a more accessible location and in a standardised format agreed upon between an agribusiness and partner FSP to ensure it is fit-for-purpose makes accessing and sharing data easier and more efficient.
  2. Security is paramount, especially when the data includes personal details such as those captured during farm and farmer profiling. To ensure data security, it is vital to know where the data is stored and who has access to it, and to deploy data monitoring mechanisms that detect any unusual activity.
  3. Backup and recovery are essential in the event of data loss. Planning ahead by setting up secure systems that allow data recovery will minimise downtime and losses to the business.

Effective data sharing requires an agreement between project partners like agribusinesses, MMPs and FSPs to ensure adherence to the principles of data governance. These include maintaining integrity and transparency in the collection and use of data; defining accountability for data-related decisions and processes; and standardising the capture and sharing of data. A data-sharing agreement is a formal agreement that sets out the terms for sharing and/or accessing data. The contract may outline implements that empower farmers to authorise the capture, storage and sharing of their data (see next section). There is no set format – it can take various forms, depending on the scale and complexity of the data-sharing partnership. The complexities that influence data-sharing partnerships include the number of partners, national data privacy and protection policies, regulations affecting cross-border data sharing and the leadership and data culture of companies involved in the partnership.

At this nascent stage, data sharing agreements to enable agri DFS in agricultural value chains are often simple bilateral agreements to primarily share farmer profile, demographic, farm profile, production and financial data, with agronomic and capacity-building data being less common. Agribusinesses and cooperatives typically share data via static reports such as in Microsoft Excel. Participation and effective data-sharing partnerships between stakeholders depend on the organisational capabilities they have. Most data-sharing partnerships include agritech companies as the providers of tailored Software-as-a-service (SaaS) digital procurement solutions to agribusinesses and cooperatives. For example, the DATAGREEN platform by SourceTrace provides multiple modules to streamline procurement activities, increase operational efficiency and bring transparency and traceability to the agricultural value chain. However, there are a few examples of agribusinesses such as Olam, who have built in-house solutions like OFIS and Olam Direct. Digital payment providers, including MMPs, also have a significant role to play by sharing transactional records from the sale of crops, which can be used as evidence of business activity and to verify transactions captured by digital procurement solutions.

Considerations when using alternative data to enable access to agri DFS for farmers

A number of considerations come into play when using alternative data generated by digital procurement solutions to enable farmers access to agri DFS.

  1. Ensuring the quality of data at the data generation stage is particularly important when assessing the creditworthiness of farmers, as inaccurate and incomplete data may lead to incorrect credit scoring. This can threaten financial inclusion by denying creditworthy farmers access to credit or making others seem more creditworthy than they are, leading to high default rates as they access credit amounts they are unable to sustain. Digital procurement tools must capture data that is complete, unique, valid, accurate, timely and consistent within and across datasets (figure 2). Data quality problems result from inconsistent data formatting, redundant or missing entries within databases and a lack of data structure. During the Rural Loan pilot in Papua New Guinea, a number of data quality issues were encountered with the data inputted by field agents in the Kamapin-Field Buzz digital procurement app. For example, the same farmer was assigned to two different villages due to a misspelling of the village name; the same individual was entered once as a female and once as a male; and the same individual was entered with different mobile phone or bank account numbers. Project partners also found significant omissions, potentially due to a field agent not inputting the data in a timely manner. To resolve these issues, further training was provided to field agents to improve all dimensions of data quality when inputting data, and a thorough data cleaning and validation exercise was undertaken.

Figure 2 Data quality dimensions. Source: Adapted from Lean data. (2018).

2. Data ownership in digitised agriculture value chains can be ambiguous and needs to be clarified by members of a partnership before generating, storing and sharing data. While farmer profile, demographic and farm profile information are less contentious, agribusinesses and cooperatives often claim ownership of production, financial, agronomic and capacity building data captured via their digital procurement solutions, although this data is directly attributable to individual farmers. For example, the location of a farm is digitally linked to a farmer’s name and financial information. It is not always clear where the distinction between personal and non-personal farm and farmer data lies

3. Data protection legislation is key to protecting farmers’ data privacy. More LMICs are enacting such legislation. Kenya and Uganda are among the latest to do so in Africa, in 2019 and 2020, respectively. One of the aims of data protection legislation is to regulate the generation of personal data, preserve the rights of data owners – in this context, the farmer and/or agribusiness or cooperative – and set out the obligations of data controllers such as the agribusinesses and cooperatives, who are responsible for determining the purpose and means of the processing of personal data. Data protection rights are grounded on the principle of customer consent; therefore, data controllers require farmer consent to share and use their data for any purpose. In our Rural Loan pilot, a farmer’s consent to use and share their data is sought orally by a field agent during the registration process and it is recorded by ticking a box on the registration form on the Kamapim-Field Buzz app. Elsewhere, HARA, an Indonesian agritech that uses a suite of apps powered by blockchain to connect smallholder farmers with financial institutions, off-takers and input producers acquires farmer consent to share their data via an SMS-based electronic authentication system. However, consent is not enough to protect farmers who often have low levels of literacy, no exposure to formal financial services, poor understanding of how their data is being used or could be used and limited choice, as consenting to share their data means getting access to products and services or not consenting means losing access.

4. Not all data is equal for farmer credit scoring. Different types of data do not have the same predictive power for credit-risk assessments.  For FSPs, the best predictor of a borrower’s ability and willingness to repay future credit is past credit repayment history. In the absence of a credit history, alternative data generated by digital procurement solutions, such as transactional records from crop sales and data on past farm yields and input expenses rank high on the hierarchy of data for credit-risk assessments. These data types are directly relevant to a farmer’s ability to repay a loan in the future and provide a good estimate of their income and cash flows, unlike social media data, for example, which for now are less relevant to smallholder farmers and therefore rank low.  

Figure 3. Ranking of data sets for farmer credit scoring. Source: Adapted from Grow Asia and SAFIRA. (2018). Digital credit scoring in agriculture: best practices of assessing credit risks in value chains.

The role of value chain actors in farmer financing

To unlock the agri DFS opportunity in agricultural value chains and support the creation of farmer economic identities, value chain actors such as agribusinesses and cooperatives along with FSPs and digital payment providers each have significant roles to play:

  • Agribusinesses and cooperatives should ensure that the digital procurement solutions they use generate the relevant alternative data points required by FSPs while considering the issues around data quality, data ownership and data protection. For example, agribusinesses and cooperatives must build their internal capacity through training to ensure that field agents collect quality data and that office clerks know how to store and share the data.
  • FSPs require a fundamental mindset change regarding the opportunity to use alternative data to offer agri DFS and they should develop appropriate strategies to serve farmers, who remain the largest untapped customer segment. For example, given that the financing needs of farmers are determined by their expenses and cash outflows at distinct stages of the growing season, FSPs must take into account the life cycle and stages of a loan and reflect farmers’ unique circumstances in their offering. In addition, FSPs should ensure that loans cover a range of crops, enable access to capital intensive investments such as farming assets and machinery and take a gender-inclusive approach.
  • Digital payment providers, including MMPs, should leverage their existing assets to enable the integration of digital procurement solutions with their payment systems and encourage the development of a digital rural ecosystem. Digital payment providers can take one of two approaches, either: through a partnership with agribusinesses and cooperatives or by developing their own digital procurement solutions. For example, Vodacom Tanzania has taken the approach of the latter by developing an enterprise solution named Mkulima.

Alternative data is widely available in digitised agricultural value chains as agritech companies, agribusinesses, cooperatives and digital payment providers increasingly deploy digital procurement solutions in the last mile. New data sharing partnerships with FSPs evoke a range of considerations that become relevant when using alternative data to enable agri DFS for farmers. They outline the roles of value chain actors and enable the identification of innovative business models that recognise the business opportunity and unlock financing for smallholder farmers.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.