7. Limitations and lessons learned

STAKEHOLDER CONSULTATION

• Eliciting key constraints systematically from a wide range of banana stakeholders through a global online survey was received very positively and ensured that the most pressing constraints to smallholder banana production were shortlisted as candidates for the assessment.
• The process of identifying potential respondents through the regional banana networks and sending out personalized email invitations was time consuming but resulted in a very high response rate, good representation of all regions, and a sense of ownership of this effort in the banana community.
• The planned inclusion of more and more immediate stakeholder feedback (such as the e- Forum conducted in parallel to the expert workshop) proved challenging mainly for logistical reasons. Real-time translation of content into several languages, time zone differences and alternative social media channels are among the issues to be thought through.
• Most stakeholders consulted appear to have been (national) researchers. Other groups such as producers, extension staff, and private sector players were underrepresented and not sufficiently heard. More diversified contacts among the country representatives of the regional banana networks would partially address this challenge. However, the language for the survey in each region presents greater challenges in certain regions. While Spanish proficiency is common throughout nearly all spheres in Latin America, English proficiency in Asia is primarily found among leading researchers, but much less frequent among producers, (small-scale) processors and development field staff.
• We hope to close the stakeholder consultation loop with our feedback survey. Participation in this online survey which will be available in English, Spanish and French in January 2017 will give the global banana community a chance to comment on results and process of the RTB banana priority assessment and provide input to parameter estimates as well as process and methodology for any subsequent similar efforts [survey].

SELECTION AND FORMULATION OF RESEARCH OPTIONS

• Selecting and formulating a limited set of “research options” based on the expert survey results with a group of stakeholders (mainly researchers) proved challenging since participants were biased towards their own area of expertise/interest.
• Packaging or bundling of research “projects” into broader themes or research options was more an art than a science. A greater range of participatory methods needs to be marshaled to manage this task in a large group setting.
• A closer alignment with existing or planned RTB research themes and activities would have made this stage easier and would also have led to results more readily applicable for RTB resource allocation considerations. However, the RTB research themes have also been in flux since that time. The formulation of viable research options addresses the heart of research planning. More prior effort may be useful in horizon scanning, the formulation of briefs on consumer demand and technology trends and background data on regional differences as they feed into a global agenda.

DATA AND INFORMATION

• Banana production, area and yield information by cultivar group and focused on small- holder production are not readily available.
• Banana prices are seasonal, location specific and depend on quality characteristics and marketing channels. The information we extracted from the FAOSTAT database is in many instances not very up-to- date and likely only very approximate.
• For the estimation of parameters used in the assessment (current and future spread of constraint, yield and cost changes resulting from adoption of research outputs, and level and pace of adoption) we relied on a small number of resource persons. While experts in their field, they likely have individual biases or knowledge gaps. One way to address this issue is the stakeholder feedback survey which will help validate assumptions and/or result in modified parameter estimates to be included in a subsequent sensitivity analysis.
• The calculation of avoided losses when the spread of emergent pests and pathogens is contained through both improved understanding of underlying factors and more effective implementation was a particular challenge. The estimation of the losses themselves builds on accurate estimates of current banana production and distribution which is not well mapped and the current understanding of how and why a particular problem is spreading. 3 This often involves very local issues, although estimates of losses need to be calculated at a more global level and over a longer time span. Our effort to develop a framework for this process highlights important research gaps.

<

METHODOLOGY

• When interpreting the results of the priority assessment, we should keep in mind that research options at very different points in their “life cycle” were included. This influences the expected benefits (e.g. ongoing/more advanced research options will look more favorable because of (more) disregarded sunk costs, higher probability of success, and shorter time spans before onset of benefits).
• The current assessment followed the assumption that one single market exists for all “bananas,” disregarding differences in price and elasticity for different types of banana and assuming that all production will be traded fresh within the country and not processed or exported.
• Initially, the priority assessment team aimed at including additional costs and benefits which required different methods in order quantify effects. One example are impacts on gender equity e.g. through changes in labor demand or intra-household resource allocation. While the RTB gender research group conducted initial case studies, a methodology which would allow the quantification of effects and thus inclusion in the assessment did not emerge.
• We did not quantify and human health or nutrition impacts which may result from changes in the application level (or type of) pesticides, or changes in nutritional content of the produced bananas. Methodologies to address these issues exist and thus collaboration with nutrition experts would benefit similar future effort.
• Finally, impacts on the environment or (on-farm) natural resources are currently not included in the assessment. Some research investments will lead to increased sustainability of banana production and/or a reduction in yield or profit variability over time. Disregarding these effects results in underestimation of benefits for these research options.
• More and more systematic sensitivity testing, especially to address the importance of such factors as probability of success which is highly influenced by researcher biases should be used to stimulate discussions within the research planning community.