Draft Report 29 Oct 2019docx
Draft Report 29 Oct 2019docx
Page 1 of 230
Karl Hughes shared this file. Want to do more with it?
  1. 1Testing the GLDC Scaling Framework: Design, Performance, and GapsDraft ReportKarl HughesKai MauschAlastair OrrAnne Mureithi29October 2019
  2. 2AbstractAccelerating the adoption of new technology –improved varieties and management practices –by smallholders remains a challenge for agricultural research and extension systems, particularly in sub-Saharan Africa. The objective of the CGIAR Grain Legumes and Dryland Cereals (GLDC) research program is to achieve adoption of these technologies at scale in the semi-arid environment. The program has developed a scaling framework which integrates ninecomponents required for successful scaling of these technologies.We tested the utility of thisframework using case studies of four large scaling projects. The framework was useful because it provided a systematic way to review the design of the projects and their scaling methods. This highlighted potential design flaws as well as opportunities for testing alternative scaling methods. The framework was less useful for evaluating project performance. Although poor performance may be the result of poor design it may also be the result of factors beyond the project’s control. Rather than use the framework to adjudicate ‘success’ or ‘failure’ the framework is more useful as a springboard for systematic learning from project experience and ensuring that these lessons are incorporated in the design of future scaling projects. The casestudies exposed some gaps in the framework. One is the need to situate the framework in its wider context, as the product of a theory of change based on the transition from subsistence to commercial agriculture. Another gap is insufficient attention to process, specifically partnerships and gender, which are both emphasised in the case study projects. The framework is a useful visualisation of the scaling process. To realise its full potential, however, the framework needs to be developed into a scaling toolkit. This toolkit would set the framework in context, explain the individual components in more detail, suggestingquestions to ask about the content of each component, include cross-cutting processes like partnershipsand gender, and give concrete examples of how the framework might be applied in practice to scaling projects.
  3. 3Acknowledgements: We are grateful to Chris Ojiewo, Moses Siambi, Dominik Klauser, Cathy Mwema, and Arega Alene who provided information and documents without which this article could not have been written. The views expressed in this report are those of the authors and should not be attributed to the organisations to which theyare affiliated.
  4. 41.IntroductionThe low adoption of new technology by smallholders is a challenge to agricultural research in Sub-Saharan Africa. The search for explanations has a long history (eg. Anderson, 1992). More recently, attention has focused on ‘the science of delivery’ or to removing the structural barriers that impede adoption at scale. The emphasis has shifted from the product to the customer. Research is no longer just about ‘getting the product right’ but about ‘getting the delivery right’. Delivery embraceseverything that is critical for this process, including institutions and policies. Thus, the old and widely-debated problem of low adoptionis now framed as a problem of social marketing, where adoption at scale is the result getting the right product tothe right customer. The rise of ‘philanthro-capitalism’ (Brooks, 2013) or of new donors with a background in the private sectorhas been influential here, bringing not just funding but expertisein commercial R & D where the ‘scaling’of innovation is an integral part of the research process.At long last, has ‘the science of delivery’ found the key to Africa’s elusive Green Revolution?While there is no universally agreed definition of scaling, the definition proposed by the International Institute of Rural Reconstruction is widely cited: “Scaling up brings more quality benefits to more people over a wider geographical area, more quickly, more equitably, and more lastingly.” (IIRR, 2000). Two pathways for scaling agricultural innovations have been identified (Gundel et. al., 2001): Scaling-out, or horizontal scaling, by expanding the numbers (and diversity) of people sustainably benefiting from (and/or area covered by) an innovation, policy, or program; and Scaling-up, or vertical scaling, by influencingor strengthening policies, political processes, and/or institutions, with the aim of creating a more enabling environment for horizontal scaling. This definition of scaling therefore takes a holistic view of the problem of low adoption that embraces constraints at the level of both the farm household and the wider enabling environment.This framing of low adoption has recently been institutionalised by theConsortium for International Agricultural Research –a consortium of 15 research institutes –in its new research programs. Between 2012-2017, seven CGIAR research programs were funded, of which six have now entered a second phase. These programs align research outputs directly with strategic development goals like poverty reduction, improved health, and sustainability. These research programs are expected to deliver new technology to millions of smallholder households. To reach these ambitious targets, the programs must rely on ‘scaling’ projects. Some projects are hybrids that combine both research and scaling components in a single project. Othersare designed specifically as scaling projectsto delivernew technology developed by earlier ‘research’ projects.The design and performance of these projects may offer insights into how to overcome the legacy of low adoptionand achieve adoption at scale.The research program on Grain Legumes and Dryland Cereals (GLDC) focuses on the cereal crops sorghum (Sorghum bicolor) and millets (Pennisetum glaucumand Eleusine coracana), and on six major grain legume crops –chickpea (Cicer arietinum), common bean (Phaseolus vulgaris), cowpea (Vigna unguiculata), groundnut (Arachis hypogaea), pigeonpea (Cajanus cajan) and soybean (Glycine max). (GLDC, 2017). Scaling new technology for these crops faces a specific set of challenges.One is on the supply side. They are grown in the semi-arid environment, where thin populations and poor infrastructure make smallholders harder to reach. Two challenges lie on the demand side. Many crops in this environment are open-pollinated and can therefore be recycled for several years without significant loss in yield,
  5. 5which reduces demand for new seed. This limits the incentive for private seed companies to market the seed of improved varieties. Finally, commercialisation is limited. Smallholders may live far from markets, and the crops they sell may also be staple food crops that are needed for household food security. Market prices may be low, reducing the demand for new technology. Urbanisation and higher incomes reduce consumer demand for sorghum and millets compared to maize. Alternative uses that can increase demand may be limited or problematic in other ways. In combination, these supply and demand-side constraints make it particularly challenging to achieve adoption at scale. The GLDC program has developed a conceptual framework for the scaling of GLDC technologies. The general objective of this report is to assess the utility of this framework by applying it to four of the GLDC program’s largest scaling projects. The specific objectives are to:1.Review the framework against the design and implementation of these projects;2.Identify gaps in the scaling framework; and 3.Recommend changes to the framework to improve the design of scaling projects and the effectiveness of scaling in the GLDC program.The primary focus of this report is not on the four projects or their success or failure in reaching the targets that they set for scaling new technology. Rather, the focus is on the scaling framework and whether it is a useful tool foridentifying what has to happenfor these targets to be met.Our main concern is not the relevance and utility of the projects but of the framework itself.The report is organised as follows. Sections 2 and 3 present the scaling framework and describe data and methods. Sections4 and 5 discuss the value of the framework for project design and performance. Section 6 identifies gaps. The final section summarises our conclusions and recommendations.2.Conceptual FrameworkThe literature dedicated to understanding the scaling of agricultural innovations indeveloping countries is vast (Feder et al., 1985; Sunding and Zilberman, 2001; de Janvry, Macours, and Sadoulet, 2016; Stevenson and Vlek 2018). Most of this literature focuses on the microeconomics of adoption, exploring decision-making, time horizons, and risk preferences at the household level (Foster and Rosenzweig, 2010; Suri, 2011; Magruder, 2017). Yet adoptionor non-adoption may be due to factors that are beyond farmers’ control. They are embedded in complex value chains that may deliver technologies that are unsuitable, or unaffordable, or without the necessary training, knowledge, and market linkages (Trienekens, 2011; Wiggins and Keats, 2013; Vroegindewey and Hodbod, 2018). In consequence, the scaling of agricultural innovations must address weaknesses along the entire value chain (Wigboldus et al., 2016). The CGIAR’s scaling framework is structured around the value chain(CGIAR, XXXX). In this framework, plant breeders develop improved varieties that are appropriate and demanded by farmers. Seed systems deliver quality seeds at accessible prices, while training and extension services equipfarmers with the knowledge and skills they need to successfully cultivate these improved crop varieties anduse improved management practices. techniques. To spur adoption, market opportunities exist that allow farmers to recoup upfront investmentsand bring additional benefits.Failure in any single link in the value chain may prevent successful adoption at scale.Thus, even if varieties with desirable traits exist and quality seeds are accessible, farmers will fail to adopt them if they are unaware of the benefits or lack the skills to cultivate them. Scaling istherefore viewed as a series of ‘hurdles’ that must be cleared
  6. 6before farmers can even begin to weigh the on-farm benefits and feasibility of improved varieties (Shiferaw et. al., 2008).Agricultural Research for Development (AR4D) provides an alternative conceptual framework where scaling is the final phase in a continuum of innovation, starting with ‘discovery’, then ‘testing’, before ‘piloting’ new technology under real-world conditions(Bernhardt 2016). Yet successful scaling is rarely this straightforward(Wigboldus et al.,2016). Substantive modifications often take place during the scaling phase, as new technology is modified to fit a specific environment (Cartwright XX)or farmers’ circumstances (Coe et al. 2014). Moreover, impacts may be reduced if scaling saturates local markets or result in negative externalitiessuch as the depletion of groundwater. Finally, treatment effects documented in more tightly controlled pilots often diminish during the scaling phase, because of less intensive support by extension services or less strict adherence to innovation protocols(Singal, Higgins, & Waljee, 2014). Dissatisfaction with the linear treatment of scaling has led to calls for a more iterative and adaptive approach (e.g. Linn 2012; Fatubi et al. 2015; Hughes et al.,2018), as well as the better integration of context-based research into the scaling process itself (Coe et al.,2014). Many also advocate the incorporation of a monitoring, evaluation, and learning (MEL) component to support adaptive management and the refinement of scaling strategies and even of the innovations themselves.Figure 1[need to re-label Figure from 4 to 1]presents a modified Innovation Scaling Framework, which combines both the value chain and non-linear approaches. Thisframework comprises nine elements, each of which is likely to influence the success of any scaling effortprogram, albeit to differing degrees depending on the context.
  7. 7Supportive policies and institutionsare critical for successful scaling. Farmers may be deterred from adopting agricultural technologies as a result of price regulation, cumbersome regulatory requirements, unequitable subsidy regimes, and monopolies in processing and trading(Tunde et al, 2018). Policy analysis and engagement with policy actors—coupled with capacity development—is therefore often critical for successful scaling(Westermann et al.,2018). Similarly, effective coordination among value chain actorsis important, particularly when challenges are complex and require cooperation among stakeholder groups (Orr,2018). Indeed, there is a burgeoning interest in the role of multi-stakeholder collaborationin strengthening agricultural value chains, e.g. via innovation platforms(Devaux, Torero, & Horton,2018). Although Wigboldus et al. (2016) correctly highlight the limitations of focusing exclusively on the farm level, nevertheless farmer preferencesremain critical. Innovationsmust have a decisive advantage over existing technology. They must also fit within the overall farming system (Birner et al., 2009)and be consistent with the farmer’s risk preferences (Koundouri, Nauges, & Tzouvelekas, 2006). Farmer preference are shaped by several factors, including awarenessof the innovation and its potential benefits and the abilityto apply the innovation effectively (Campenhout, 2019; de Janvry, Sadoulet, & Rao, 2016). These in turn are correlated with accessibilityto extension and information delivery services, participation in relevant social networks (Meijer et al., 2015), as well as access to the material inputs such as seed or inorganic fertilizer associated with the innovation (Melesse, 2018). Successful scaling is often determined by the marketabilityof a crop, whichis linked to consumer demand. Farmers are generally unwilling to invest in an innovation unless they foresee a financial return (Wiggins and Keats, 2013; Verkaart et al. 2019) orat least some value for domestic consumption. Finally, successful scaling ismore likely to result from adaptive management supported by a timely feedback system, which includesupstream researchtoenhance the scaling process.3. Data and methodsOur evaluation of this conceptual framework is based on four projects in the GLDC research program(Table 1).All four were ‘flagship’ scaling projects with large budgetsthat were expected to reach between 400,000 –4 million households. Twoprojects(HOPE and TL III) were research projects with sizeable scaling components, whereas AVDC and MISST focused on scaling existing technology. Between them the four projects coveredseven major crops commonly grown in semi-arid environments, including two dryland cereals (sorghum and millets) and five grain legumes (groundnuts, chickpea, cowpea, pigeonpea, soybean and common bean) as well as root crops (Irish potato, sweet potato) and the value chains for livestock and milk.Two projects (HOPE and TL III) wereregional in scope and coveredboth East and West Africa, while two coveredjust one country(Kenya and Malawi). Their duration ranged from three (AVCD) to five years (HOPE). Funding came from both private philanthropy (BMGF) and the public-sector (USAID).The common objective of all four scaling projects was to improvethe delivery of certified seed.HOPE 2 and MISST targeted a specific volume of seed per country or per crop (ICRISAT, 2014; FTF-Malawi, 2019). TL III sought to supply “at least 20% of the seed required on a national basis for each crop” (ICRISAT, 2013).The AVCD project focused on commercialisation (“increase the value and volume of the products in each value chain by 15%”) but included seed targets for all its crops(ILRI, 2015).
  8. 8Table 1. Case study projects for evaluation of scaling frameworkNameTarget populationSeedscaling targetsDurationGeographiesCropsLead AgencyBudget (USD)DonorHarnessing Opportunities for Productivity Enhancement (HOPE) of Sorghum and Millets in Sub-Saharan Africa (Phase 2)4,000,000 householdsIncrease supply certified seed by 12,000t/yr;300,000 Small Seed Packs;30,000 ISSFM minikits2016-2020East Africa: Ethiopia, TanzaniaWest Africa: Niger, Nigeria, Niger, Mali, Burkina Faso SorghumPearl milletFinger milletICRISAT15 m.Bill & Melinda Gates FoundationTropical Legumes III (TLIII): Improving Livelihoods for Smallholder Farmers: enhanced Grain Legume Productivity and Production in Sub-Saharan Africa and south Asia500,000 householdsSupply 20% of certified seed required for target crops2015-2019West Africa: Burkina Faso, Ghana, Mali, Nigeria, East Africa: Ethiopia, Tanzania and UgandaIndia: Uttar Pradesh (chickpea)Common bean, cowpea, chickpea, groundnutICRISAT25 m.Bill & Melinda Gates FoundationAccelerated Value Chain Development (AVCD) Program for Kenya 444,000 householdsIncrease volume and value of market sales by 15%2016-2019Kenya (17 counties)Value chains for livestock/fodder, dairy, sweet potato, Irish potato, sorghum, groundnuts, pigeonpea ILRI, ICRISAT, CIP25 m.Feed the Future-United States Agency for International DevelopmentMalawi Improved Seed Systems and Technologies Project (MISST)220,000 households14,500 tons of seed2014-2019Malawi (7 districts, later increased to 10) Maize, sweet potato, pigeonpea, groundnuts, soybean, biological controlICRISAT, CIP18.6 m.Feed the Future-United States Agency for International DevelopmentSources: ICRISAT (2013), ICRISAT (2014), ILRI (2015), Akinwale et. al. (2016).Information on project design and performance was obtained from relevant project documents.For TL III, information was obtained from the project proposal (ICRISAT, 2013) and publications (Varshney, 2019; Rubyogo et. al., 2019). Information for HOPE 2 was obtained from the project proposal(ICRISAT, 2014). Since both TL III and HOPE 2 have not yet ended, project completion reports werenot yet available. For the AVCD project, both the project proposal (ILRI, 2016) and the last annual report were available (FTF-Kenya, 2018). For MISST, information was available from the project completion report (FTF-Malawi, 2019) andfromreporting on soybean (Akinwale et. al., 2016). In addition, information on lessons learned was gleaned from skype interviewswith selected project personnel,including Principal Investigators, project managers, and research scientists.
  9. 94.The Scaling Framework and Project DesignTable 2 maps the scaling framework against the design and scaling methods of the four projects. The contents are not intended to be exhaustive but to summarise the main features. The results show that the scaling framework captures the approach to scaling used by the four projects. Most scaling activities can be fitted into one of the ninecomponents of the framework. The advantage of the scaling framework is that, by combining separate components on one page, it gives planners a clearer picture of the overall project and the methods that will be used to address each component of the scaling process. This is not to claim that the framework includes everything required for scaling, but that it is sufficiently comprehensive to be a useful planning tool for scaling projects.Table 2. GLDC scaling framework and design of case study projects Scaling Framework Case Study ProjectsHOPE TL IIIAVCDMISSTFarmer preferencesFarmerParticipatory Varietal Selection (FPVS)National crop strategiesFarmer Participatory Varietal Selection (FPVS)National crop strategiesFarmer Participatory Varietal Selection (FPVS) for potato and sweet potatoFarmer evaluation at field daysAwarenessand AbilitySeed FairsField DaysVideoPrinted information in local languagesSeed FairsField Days/demonstration plotsInformation packsPrinted information in local languagesPrivate extension agentsField Days/demonstration plotsDemonstration plotsField DaysAccess to technologyQuality Declared Seed (QDS)Small Seed Packs (SSP)ISSFM MinikitsSeed roadmapsSmall Seed Packs (SSP)‘Seed delivery platforms’Bilateral seed projects (eg. AGRA-PASS)Seed roadmapsSmall Seed Packs (SSP)Smallholder seed producers and private seed companiesCommunity seed banksSmall seed packs (SSP)Seed producer groups, small seed companiesPolicies and institutionsQuality Declared Seed (QDS)Capacity building of NARS breeding programsPrivate seed companies Capacity building of NARS breeding programsEvidence for policyBarriers to livestock trade and market developmentRevolving fundfor early generation (breeder & basic) seedMarketabilityTraining courses and manuals for seed production and marketingInnovation platformsMarketing associations/organisationsDairy business hubsConsumer demandIncorporate ‘consumer trait preferences’in breeding programIncorporate short cooking time into breeding programTV programs for consumers of SmartFoodsPhotovoiceValue Chain Coordination Partnerships with NARS, NGOs and seed companiesResearch for development partnershipsJoint planning of seed needsBusiness support to producer organisations and private seed companiesMarket information through local radioCollective marketingBrokering bulk salesFacilitated sale of seed to farmers’ associations and cooperatives
  10. 10Upstream researchAdvanced genomics to shorten the breeding cycleGenome Wide Assessment Studies (GWAS) Advanced genomics to shorten the breeding cycleFeedback systemsMLE Plan and MLE specialistBreeding Program Assessment Tool (BPAT)MLE Plan and MLE specialistBreeding Program Assessment Tool (BPAT)Results Based MonitoringSources: ICRISAT (2013), ICRISAT (2014), ILRI (2015), Akinwale et. al. (2016).The framework identifies several possible gaps in the design of these four scaling projects. In some cases, components in the scaling framework are missing. Three examples of missing components are Marketability,Policies and Institutions, and Upstream research.The framework defines Marketabilityas the“ready access to markets with profitable and predictable returns”(Figure 1). This component is fundamental for scaling new technology designed for commercial markets. Yet it was deliberately omitted from the TL III projecton the grounds that “research on markets, policy and seed regulation ... are being adequately covered by other initiatives or that sufficient information is now available” (ICRISAT, 2013, p. 13). Similarly, the framework defines Policies and Institutionsas “an enabling environment with supporting institutions and appropriate incentives/absence of bottlenecks” (Figure 1). Yet only one project (AVCD) explicitly addresses policies that threaten adoption at scale. In this case, these policies are internal barriers totrade in livestock, including taxes, movement certificates, and veterinary certificates that limit the movement of livestock between different counties in Kenya. The project targets the removal of six key trade barriers to scaling (ILRI, 2016, pp.43-44). Policy engagement around seed systems is complicated by the fluid nature of government policy –making, with some Seed Acts not yet implemented (IFPRI, 2018).Influencing policy requires specialist expertise and time. This may explain the reluctance of these projects to invest resources in this component of the scaling framework.Upstream research does not feature prominently in thedesign of all four projects. The framework defines this component as “insights used to inform discovery & proof of concept research” (Figure 1). This component is associated with research projects. The discoveries of the TL III breeding program fill an entire issue of Plant Breeding(Varshney, 2019). Upstream research might seem irrelevant for scaling projectsthat promote existing technology, but there is much still to discover about the effectiveness of different scaling methods. We return to this subject laterin this section.Of course, thesegaps identified by the scaling framework may not be genuine gaps if they are addressed by other programs or projects. Individual projects may not have to address all nine components of the scaling framework in relation to market development. Both HOPE 2 and TL III were part of a continuum of older projects. TL III succeeded two earlier phases lasting seven years (2007-2014), while HOPE 2 followed one earlier phase of five years (2010-2016). This meant they could build on a legacy of previous research. Similarly, scaling in the AVDC project was complemented by other USAID-funded projects on policy, livestock marketing, and value chain development (ILRI, 2016, pp. 2-3). However,the scaling framework can be used to identify these complementarities at the design stageto avoid gaps appearing during implementation.The scaling framework can also help identify scaling components that deserve greater attention. Among these we may single out Feedback Systems. All four projects included ‘learning’ as part of their M & E systems. In practice, however, for large projects burdened with multiple milestones, the tyranny of the logical framework means that only lip-service is paid to
  11. 11learning, with the result that learning becomes the ‘invisible output’.1Some case-study projects did make a conscious effort to learn. The proposal documents for HOPE and TL III include sections andappendices that list learning; there are also examples of project publications that synthesise lessons (Monyo and Varshney, 2016). But learning is often left to the discretion of the project manager.2Learning needs to be made more systematic. Table 2 shows no evidence that specific learning outputs were identified at the design stage of these scaling projects, or how these lessons will be used to change the design of the projectduring implementation.Another use of the scaling framework at the design stage is to direct attention on the methods themselves. Table 2 shows an impressive range of scaling methods. However, a closer look suggests room for improvement. The Awareness and Abilitycomponent showsthe use of new communication methods. Alongside traditional standbys like demonstration plots there are TV programs, video and DVD and the use of local radio for market information. Upstream Researchemploys new tools to shorten breeding cycles. But the methods in Consumer Demand(“catering to and/or stimulating affordability and availability”) focus on consumer trait preferences. Agricultural economists are skilled in providing this information through baseline surveys or choice experiments. But national programs require other kinds of informationto set breeding priorities, such as the size of the market, the various market segments, and the trait preferences of these segments.Yes, economists can map value chains and estimate elasticities of demand. But they are not market researchers. And it is the tools of market research, like the segmenting-targeting-positioning approach, that breeding programs need to develop “customer profiles” that match their products with the needsof their customers (Orr et al., 2018). Breeding programs need methods like these to understand Consumer Demand.Scaling projects are fertile ground for testing scaling methods. There is a large literature on this subject. Examples that spring to mind include the post-mortem on the Training and Visit (T & V) extension system(Feder et. al., 2006), the continuing debate over the effectiveness of Farmer Field Schools (van der Berg and Jiggins, 2007; Feder et al., 2008; Friis-Hansen and Duveskog, 2012)and a recent study on the effectiveness of Farmer Business Schools (Chilemba and Ragasa, 2019). Less common are studies that directly compare the effectiveness of two scaling methods. However, some methods are more effective than others, particularly where new technology is knowledge-based (Bentley et. al., 2007; Riker-Gilbert et. al., 2007).A recent review of scaling projects in Malawi found little assessment of scaling up methods, and recommended “qualitative assessments of the methods used for scaling up and share the experiences through a formalized process, rethinking projectsto be more amenable to longer term targets, thus enabling more analysis of methods within projects”(IFPRI, 2018). The scaling framework can help projectsat the design stage toidentify opportunities to assess the effectiveness of scaling methods.Using the framework to map different scaling methods can also identify opportunities for learning from other projects. Some scaling methods in Table 2 are novel.Reality TV shows like Shamba Shape-Up already attract a wide audience across Swahili-speaking East Africa (Clarkson et. al., 2018). The AVCD project developed TV programs with celebrity chefs to showcase the nutritional benefits of sorghum and milletsto viewers in Kenya. The shows 1Phase 1 of the HOPE project had more than 1,000 milestones, which meant that annual planning meetings were dominated by monitoring progress against milestones rather than reflecting on what we were learning and the implications of these lessons for the project’s Theory of Change. Even this was tame compared with the AVCD project in Kenya, funded by USAID’s Feed the Future Program, which required monitoring reports every two weeks (M. Siambi, pers. comm.)2In Phase 1 of the HOPE project, for example, one of the authors submitted a list of lessons learned from market development which challenged some of the project’s assumptions. The response from the Project Manager was that it was “too late”.
  12. 12reached a weekly audience of 800,000 viewers (ICRISAT, 2019, p. 43). Other innovative tools included Photovoice, in which village women are taught to photograph all the food thehousehold eatsover 24 hours, and the results used for training them in improved nutrition (ICRISAT, 2018).These methods have a wide range of applications.5.The Scaling Framework and Project PerformanceThe scaling framework can also be applied retrospectivelyto explain project performance. However, this is trickier. Scaling projects may fail because of poor design: a key component of the scaling framework was missing. But they may also fail because the design was not properly implemented or because of events that could not be anticipated at the design stage (for example,a sudden reversal in government policy).These confounding factors limit the value of the scaling framework as arear-view mirror to explain why some projects succeeded while others failed.Success is usually measured against the original targets set out in the project design. Over three phaseslasting 12 years (2007-2019) the TL III project delivered 500,000 t of certified seeds (Varshey, 2019).However, because of rapid growth in the area planted to legumes, the original target of meeting 20% of seed requirements was not met (C. Ojiewo, pers. comm.). No final figures were availablefor HOPE phase 2, which ends in 2020. The AVCD project exceeded its target to increase incremental sales by 17 % (USAID, 2018). The MISST project met about 80 % of its production targets for early generation and certified seed (FTF-Malawi, 2019).In general, these projects were successful in meeting their targets for seed production.Performance did not therefore reveal major design flaws. However, implementation did reveal weaknesses in some assumptions. The scaling framework includes a component on VC Actor Coordination, defined as a “holistic approach with multi-stakeholder planning and trouble-shooting” (Figure 1). Although this component was not missing from the TL III project, it proved difficult ti implement. The project used a public-private partnership model of seed deliveryin which national breeding programs and statutory bodies produced foundation seedfor private seed companies and farmer groups.The initial funding for foundation seed came from project funds but, to make production sustainable, the income from the sale of foundation seed was supposed to be re-invested in a revolving fund that financed production in subsequent years. In practice, the income went into a government account in the central bank and was used for other purposes, so the fund failed to revolve(Chris Ojiewo, pers. comm.).This was outside the project’s control.Similarly, the project design envisaged production and delivery of certified seed through private seed companies, which were start-ups funded by grants from the Program for Africa’s Seed Systems (PASS) project(ICRISAT, 2013, p. 15). However, these start-ups proved unable or unwilling to become fully-fledged seed companies. They complained of unfair competition from national breeding programs, some of which set up their own companies to produce certified seed. They also faced regulatory hurdles. In theory, seed regulations have been harmonised to allow the free movement of registered varieties and certified seed within African regions (Rohrbach et. al., 2003).In practice, this is frustrating and time-consuming for small seed companies. Private companies also want proprietary branding of the products they sell, but because improved varieties are public goods this requires licensing agreements and royalty payments to national breeding programs.Smallseed companies have needed specialist help from Syngenta’sSeeds2B platform to navigate these bureaucratic hurdles (Syngenta Foundation, 2019).Finally, the expectation of further grants made some companies reluctant to invest their own resources. As a result, most of the certified seed in the TL III project has been produced and delivered not by private seed companies but by the informal sector (C. Ojiewo, pers. comm.). Tanzania saw a dramatic increase, from just 10 farmer organisations and community seed producers in 2007 to 470 in 2017 (Rubyogo et. al., 2019).
We use cookies to provide, improve, protect and promote our services. Visit our Privacy Policy and Privacy Policy FAQs to learn more. You can manage your personal preferences, including your ‘Do not sell or share my personal data to third parties’ setting using the “Customize cookies” button below.