A report titled Digitization and Coordination of Kenya’s Agricultural Sector Data was completed in July 2019. It is a guide for implementing the data and innovation flagship number 8 of the country’s Agricultural Sector Transformation and Growth Strategy (ASTGS). I got my hands on a copy which was recently publicly available on the Kilimo Open Data portal (download full report (7.9mb pdf)). In this article, I refer to the report as a digital agriculture strategy for Kenya’s Ministry of Agriculture, Livestock, Fisheries and Irrigation (MoALFI). It is the closest I see to Kenya’s national strategy for harnessing the transformative power of digitalizing agriculture. The report, authored by global consulting giant McKinsey and funded by AGRA on behalf of MoALFI prioritizes seven (7) use cases to be implemented. It is a four year plan covering the fiscal year 2019/20 to 2022/23. I find no in-depth coverage of the seven prioritized use cases, so I feel it’s “better late than never” to write about them, especially considering typical delays in Government strategy implementation. In this post I provide my nuanced summary of what I see as the rationale, and design of the use cases. This is more so as the use cases (at least the second one) typify instances of what my earlier published research conceptualised as an aggregator platform for digital services in agriculture.
Use case 1: Improving farmer inputs subsidization
This use case is designed to improve input subsidization programs supported by the Government and other funders. The problems to be solved under this use case are that government subsidised inputs are not necessarily aligned to soil health management principles, and the unintended creation of a dual-pricing system aggravated through diversion of subsidized inputs by cartels. Other problems to be solved across the stages of the existing subsidy system include rent-seeking behavior hurting farmers, and fertilizer delivery being delayed. The key problem solving elements of this use case are, (a) A digital farmer registry managed by MoALF, (b) eVouchers sent by the national treasury directly to farmers, and (c) Payments to registered agro-dealers being made in a timely fashion while securing traceability of inputs to fight counterfeiting. The spending decision on the eVoucher is to be made by the farmer and is to be guided by the farmer’s uptake of registered agronomic advisors. Funding for the e-incentives is expected to come from a “basket fund”, pooling GoK and development partner finances, and with checks and balances for any release of funds. This use case aims to impact 1.4 million farming households and 2,300 agro-dealers by 2023. Its implementation is estimated to cost KES 1.2 Billion (Approximately USD 11 Million) by 2023. The digital platform implementing this use case is to be hosted by the Kenya Agricultural & Livestock Research Organization (KALRO). The prospects of this use case can be enhanced by harmonizing efforts with concurrent input subsidization programs such as those at the 47 counties and by national programs such as KCEP-CRAL.
Use case 2: Improving farming practices through customised e-extension
This use case is designed to improve practices among farmers to enhance their productivity through enhanced agronomic advisory services. This is more so considering a grower-to-extension-officer ratio as poor as 1:5,000 whereas the FAO recommended ratio is 1:600. Data analytics on weather, pest and disease trends, production yields and pricing are envisaged to help amplify improvements in farming practices. The use case envisages farmers accessing expanded resources on farming practices through extension service providers vetted by KALRO. These extension service providers are in turn expected to tap into a searchable portal of digital agriculture solutions to support their consultations with farmers. The extension service providers may include the traditional extension officers, village-based advisors, private sector field officers, model farmers and members of the 4-H foundation. Under this use case, the consent of growers is required for a digital agriculture service to incorporate a grower to its userbase through the government-run digital platform. Growers and other users of the digital platform can be expected to rate and review the constituent digital services. This can boost platform-wide efficiency according to the research on sources of value creation in such digital platforms. Under this use case, the platform reserves the right to remove digital agriculture services whose ratings remain low in what could contribute to loyalty-centredness as a source of value creation. The extension service providers are also to be paid an incentive to log details of their consultations with farmers in the digital platform. This is so as to help generate rich new insights for KALRO and MoALFI. Such details logged may include geolocation, type of digital service used and type of problem solved. This use case targets the registration of 2,300 extension service providers by 2023. It also targets 500,000 farmers per year accessing services to improve their farming practices. The estimated annual cost of this use case is KES 50 million (USD 460 million).
Use case 3: A food balance sheet monitoring national emergency reserves
This use case aims to maintain a reliably robust Food Balance Sheet (FBS) to help MoALFI and the Strategic Food Reserve Trust Fund (SFRTF). This is so as to reduce food shortages during emergencies. The use case is about digital inventory monitoring of national food reserves with accurate and real time data. It includes the use of satellite imagery to gather production data and to predict future stock needs. It also includes gathering intelligence from customs declarations and other trade records as proxy signals for fraudulent or anomalous market trends. Other features of this use case include gathering intelligence from consumption data and projecting future consumption patterns. With a robus FBS, the government agencies will be expected to make informed decisions that will also reduce the cost of procuring stocks under duress. This use case targets reducing volatility in stocks purchased for the SFRTF by 50%. It also targets boosted food resistance among close to 4 million high risk households during emergencies. The annual cost of this use case is estimated at KES 200 million (USD 1.84 million) by year three of the plan.
Use case 4: Early Warning System (EWS) for food price inflation
This use case seeks to facilitate dynamic trade and price stability decisions using an Early Warning System (EWS) for food price inflation. The problems solved through this use case include the unreliably varying levels of accuracy and frequency of data collection on production which is manually done by enumerators and with limited validation. This use case integrates data with early warning elements to indicate likely changes in the prices of food commodities. Data on production, soil quality, pest and disease trends is expected to have early warning components. So is data on commodity trade from sources such as the Regional Agriculture Trade Intelligence Network (RATIN). The resulting early warning system is expected to be a single-source of information to forecast food price inflation. With such reliable information, MoALFI, the SFRTF and Cabinet can be expected to make timely, cost effective and targeted interventions with minimal distortion to market mechanisms. This use case targets to reduce volatility in food prices by 50% to match regional averages by 2023. The estimated annual cost of maintaining the system for this use case is KES 8 million by year 3 of the plan. I’d argue for additional consideration in this use case to incorporate market data from proven private sector driven price stabilization platforms such as Twiga Foods in the fruits and fresh vegetables space. This is especially with the demonstrated potential to use big data to power inclusive food markets.
Use case 5: Optimal agricultural value chain selection
This use case envisages selection and prioritization of focus value chains (crops, livestock, fishing etc) based on land use optimization models. The existing processes for selecting focus value chains by national and county level organs are faulted for having contradictory objectives and not being adequately evidence driven. Data to be fed into the optimization models may include aspects such as soil health and economic data including export markets. The land optimization models may be tuned with parameters such as environment protection and yield optimization. Integrating the large datasets onto a single platform facilitating modeling under this use case anticipates support from enabling policies for data sharing, security and privacy potentially addressed under use case 7. This use case on value chain selection targets doubling the yields among small-scale farmers by 2023. It also targets boosting household food resilience for 1.3 million farming, pastoralist, and fishing households during drought in arid and semi-arid regions of Kenya. The use case estimates an annual cost of KES 120 million (USD 1.1 million) by year three of the plan’s implementation. It is notable that the implementation of this use case is envisaged to be county-based, allowing more local-level prioritizations that are more likely to resonate with growers on the ground. Grower-level solutions for value chain selection such as Waterwatch, CropIn are also acknowledged in the plan and are considered complementary to the national and county level prioritizations. This begs the question of whether value chain selection should not be considered a grower-level decision made with the support of agronomic advisory under use case #2.
Use Case 6: M&E dashboard for data collection, verification and visualization
This use case is expected to streamline the data collection, verification and visualization of the outcomes targeted in use cases 1-5 (above). It builds on the observation that there are more than 10 visualization efforts capturing Kenya’s agricultural sector data from 200+ data sets. Some of these efforts are observed to be inactive or outdated, ostensibly due to their scope-related cost constraints. Examples of such visualization efforts include the upcoming KCSAP big data platform funded by the World Bank, the National Crop bulleting, the AfDB’s Africa Information highway, and the defunct Kenya National Bureau of Statistics visualization portal. For simplicity, the focus of the dashboard in this use case is limited to approximately 10 KPIs considered most relevant to the senior leadership at MoALFI. Upon success with the simplified, focused dashboard, the cascading of the same is envisaged within MoALFI and among the 47 counties. More granular project tracking tools are then expected to emerge, automating the build-up of insights for the macro-level dashboard. The estimated annual cost of this module is KES 20 million (USD 184,000)
Use case 7: Standards and protocols for a national data sharing platform
This use case seeks to establish standards and protocols for a national shared-access platform for agriculture data. The idea is to start with data handled by GoK and to progressively incorporate data handled by private sector players. Under this use case, platform users are expected to build on the high volume, highly interoperable data to create new insights for their interventions in agriculture. The problem to be solved by this use case is typified by the existence of multiple siloed farmer registries profiling more than two million farmers. These registries include those maintained by MOA-Info, Digifarm and OneAcre Fund. Under this use case, The Zambia Agriculture Management Information System – ZIAMIS, a FAO supported platform, is indicated as a candidate for adapting and scaling out in Kenya. The estimated annual cost of this use case is KES 20 million which is indicated as already funded by the World Bank under the Kenya Climate Smart Agriculture Project (KCSAP).
Caveat: Government prioritized value chains and intervention areas
It is worth noting the intentional prioritization of digital interventions that the government is well placed to champion and implement among the seven use cases. As such, digital agriculture services deemed best implemented by the private sector and players other than Government agencies were not prioritized in the strategy. For instance, my favourite topics of overall market efficiency and demand driven production did not feature among the digital agriculture use cases. This is ostensibly because such issues are arguably best addressed by the private sector and are arguably related to market linkage concerns being addressed in the warehouse receipt system law enacted in 2019. Furthermore, the use cases have inherent prioritization of selected value chains, especially the grains sub sector including beans, maize, rice, wheat. In any case, I encourage readers keen on the full context and finer details about the planned implementation of these use cases to read the full report. Moreover, the full report has additional insights about implications for data ecosystem requirements in the use cases that should be of interest to the digital agriculture ecosystem actors, especially the providers of digital services.
Appreciation: For this post, I am grateful to Nixon Gecheo, AGRA’s Senior Program Officer for Digital Systems and Solutions for Agriculture. He brought this report to my attention, I have been otherwise clueless.
John Kieti 