Forecasting Demand for Seed: The Case of Ethiopia

Agrilinks contributor: S34D Feed the Future Global Supporting Seed Systems for Development

Agrilinks contributor: Catholic Relief Services

S34D Feed the Future Global Supporting Seed Systems for Development | BHRAMAR DEY | Catholic Relief Services

December 1, 2021

Close up of seed — December is Seed Systems Month.

This post is written by Bhramar Dey with Support Seed Systems for Development at Catholic Relief Services and Bezabih Emana.

Ethiopia is one of the few countries in sub-Saharan Africa to have a working process and system to forecast seed demand across all administrative levels in the country. The system is used every year to forecast demand for certified seed and early-generation seeds. However, despite the efforts made, the current system is often unable to match the demand and supply of seeds. Thus, the priority of the Ethiopian seed sector is to strengthen and modernize the current system and processes in place. To that end, Supporting Seed Systems for Development (S34D), led by Catholic Relief Services, undertook a study to assess the present system and methods employed to forecast demand. The goal was to provide a roadmap on what the next steps could be to modernize the approach such that the gap between effective and realized demand narrows over time.

Using the framework of a seed data value chain and bottom-up approach, S34D-Catholic Relief Services designed customized survey instruments for each of the different administrative levels, as well as for national and regional stakeholders who perform critical functions in the seed demand forecasting process. Thirty-one key informant interviews were conducted to collect information regarding the system and processes used at present. These informants spanned across more than two dozen institutions and administrative levels in Amhara and Oromia regions, which are the dominant seed producers and users in Ethiopia.

Recommendations for a system that works

Results indicate that Ethiopia has a system that works and there is a strong momentum among actors and stakeholders to improve the processes, to modernize the system, and to strengthen the methodology used. For example, currently there is no digitized data archives that could enable panel analyses and trends analyses using rigorous econometric approaches. Therefore, shifts in the demand curve cannot be estimated from the present system. Inability to do that leads to mismatches in demand and supply on the ground. Additionally, market information is rarely used (nor collected) in the methodology to estimate demand, leading to biased results. For example, survey results indicate stakeholders would appreciate the collection and utilization of important parameters such as seed-grain price ratio. Figure 1 shows the key variables for which data is collected.

Discussions with stakeholders indicate gender and age disaggregated information is neither uniformly collected nor systematically used throughout the process. Furthermore, data adjustments that are made at different administrative levels are not always transparent and well documented. The seed price-setting includes only the public seed enterprises and omits the private sector players. The actual production costs (and associated cost functions) are not transparent. The final price at which the seeds are sold varies based on distribution and handling costs. The process is furthermore complicated through various communication channels – often transmitting similar information bits – without the use of a single platform, leading to a time lapse in receipt of the data.

Recommendations from the assessment are provided in Box 1.

Figure 1. Mapping of Variables for Which Certified Seed Demand Data is Collected at Different Administrative and Stakeholder Levels

Box 1. Recommendations

Move from paper to digital data collection at all administrative levels.
Build technical capacity, especially at grassroots levels.
Collect age- and sex-disaggregated data to assess preferences of women and youth.
Collect and share good quality near-real-time information on the seed-grain price ratio to increase market transparency.
Improve the methodology of forecasting (using econometrics to do predictive modeling; big data and micro-level information, near-real-time forecasts, etc.). Capture shifts in demand.
Initiate data coordination nodes to exchange both micro- and macro-level data necessary to support econometric models (in collaboration with the CSA, ATA, MoA, the World Bank, and groups that deal with statistics from space data, weather data, etc.).
Put centralized databases in place to increase transparency; strengthen the documentation system, which also enables online tracking.
Establish a digital library with data archives across space and time.
Revisit the price-setting strategy if the country wishes to liberalize the seed sector. As a first step, it can factor in seed quality while setting prices. Seed-pricing mechanism should also be checked with grain price.
Create awareness and develop marketing strategies (for new and existing varieties which have a huge market pull) such as farmer segmentation models; willingness-to-pay approaches; constructive feedback mechanisms.
Increase transparency in seed-demand forecasting, price-setting and seed distribution
Provide feedback to farmers regarding the response (if they get the demanded seed) until the seed distribution time. Lack of feedback to the farmers on the status is discouraging them to register their demand in the future.
There should be standardized and well-documented adjustments of data at all levels.
Pilot approaches and case studies with unions who are showing interest in seed-demand forecasting.
The early-generation-seed forecasting process should be standardized along regions and within regions. Applying uniform approaches and transparency in the process needed.
Finalize early-generation-seed demand forecasting and the decision for its production period not later than the end of February each year.

Concluding remarks

Seed demand forecasting is a function of many variables, including infrastructure and institutional factors. For that reason, all major stakeholders have a role to play in it. Avoiding or excluding some is not beneficial for the long run. Ethiopia has a good system and process in place (unlike many countries in sub-Saharan Africa). That is already a promising start to augment the current system through modernization using digital approaches and rigorous forecasting methodologies that incorporate both micro-level and big-data structures.

Currently there exists positive momentum for strengthening and enhancing the seed-demand forecasting process. There is abundant talent and capacity within the country to match the requirements. However, often that is not uniformly spread across the country and administrative levels. Mobilizing technical capacity and resources to initiate this change is a necessary.

Finally, it is possible to modernize the system only through partnerships – this is not a “one-project” or “one-stakeholder” task. This encompasses multiple groups, stakeholders, administrative levels, geographies, crop varieties, etc. Going forward, Ethiopia could initiate a representative working group to pilot the recommendations laid out in this assessment.