WPS4573 Policy ReseaRch WoRking PaPeR 4573 Linking African Smallholders to High-Value Markets: Practitioner Perspectives on Benefits, Constraints, and Interventions Spencer Henson Steven Jaffee John Cranfield Jose Blandon Paul Siegel The World Bank Agriculture and Rural Development Department March 2008 Policy ReseaRch WoRking PaPeR 4573 Abstract This paper provides the results of an international smallholder gains in this area. The results confirm a survey of practitioners with experience in facilitating growing `consensus' about institutional roles, yet suggest the participation of African smallholder farmers in some ambiguity regarding the impacts of smallholder supply chains for higher-value and/or differentiated participation in higher-value supply chains and the agricultural products. It explores their perceptions about appropriateness of the indicators most commonly used the constraints inhibiting and the impacts associated to gauge such impacts. The results also suggest a need with this supply chain participation. It also examines to strengthen knowledge about both the `old' and `new' their perceptions about the factors affecting the success of sets of constraints (and solutions) related to remunerative project and policy interventions in this area, about how smallholder inclusion, in the form of the rising role of this success is and should be measured, and about the standards alongside more long-standing concerns about appropriate roles for national governments, the private infrastructure and logistical links to markets. sector, and development assistance entities in facilitating This paper-- Agriculture and Rural Development Department--is part of a larger effort in the department to promote broad-based agricultural growth and market development. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at sjaffee@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Linking African Smallholders to HighValue Markets: Practitioner Perspectives on Benefits, Constraints, and Interventions Spencer Henson*, Steven Jaffee**, John Cranfield*, Jose Blandon* and Paul Siegel*** *International Food Economy Research Group, University of Guelph **Lead Economist, Agriculture and Rural Development Department, World Bank ***Consultant, Agriculture and Rural Development Department, World Bank Linking African Smallholders to HighValue Markets: Practitioner Perspectives on Benefits, Constraints, and Interventions Introduction As part of the wider discourse on propoor agricultural growth and development, there are ongoing debates about the viability of smallholder production and the opportunities created by highvalue markets (Weatherspoon et al., 2001; Hazell, 2005; Pelupessy and Van Kempen, 2005; Humphrey, 2006). Proponents of smallscale farms highlight their potential competitive advantages over largescale commercial production, predominantly through lower costs in accessing and managing family labor and superior local knowledge (Dorward, 1999; Lipton, 2005 Poulton et al., 2005; Pingali et al., 2005). They also emphasize how highvalue markets offer lucrative opportunities for enhancing smallholder income as domestic markets for livestock and horticultural products exhibit particular dynamism (World Bank, 2007), and as nontraditional, highervalue foods have come to account for the majority of developing country agri food exports (Jaffee and Sewadeh, 2006). Conversely, skeptics argue that the highervalue market participation of smallholder farmers will be severely constrained by the requirements of more discerning consumers and supply chain `gatekeepers' (Maxwell, 2004). As buyer demand for timely and reliable delivery of products that conform to exacting quality standards increases, small farmers are exposed to transaction costs that are high and for which economies of scale are significant. Thus, while smallholder participation in high value markets is not a new phenomenon (Williams and Karen, 1985; Glover and Kusterer, 1990), changes in these markets have brought about a need to rethink the scope for such participation in the future (Reardon et al., 2004; Humphrey, 2006; Henson and Jaffee, 2006). Over the past decade, many developing country governments and international development agencies have concluded that increased smallholder participation in highervalue agrifood markets is essential for meeting economic development and poverty reduction objectives. Consequently, a growing range of interventions have been designed and implemented to link smallholders both to highervalue domestic markets and nontraditional export markets (see for example Shepherd, 2007; Temu and Temu, 2005). These have involved a range of market intermediaries, from producer organizations and cooperatives to private sector exporters, with both `top down' and `bottom up' approaches proliferating. While attempts have been made to characterize the plethora of interventions aimed at facilitating the participation of smallholders in highvalue markets (see for example Page 1 and Slater, 2003; Shepherd, 2007) there is a lack of systematic comparisons of experiences with respect to when and where participation in highvalue markets is beneficial to smallholders, the constraints that impede their participation and the effectiveness of alternative means for overcoming these constraints. There is also a lack of consensus on measures to be used in judging the `success' of interventions such that it is possible to contrast between those that `work' and those that do not. This paper attempts to throw some light on these issues. Specifically, it reports on the results of a survey of practitioners involved in schemes to better link African smallholders with highervalue markets. Before highlighting those results, we provide an overview of the existing literature. Benefits and challenges of smallholder participation in highervalue markets The existing literature on impacts of smallholder participation in highvalue markets is diverse in terms of mode of investigation and primary focus. The contract farming literature sheds some light on the impact of smallholder participation in highvalue markets (see for example Glover and Kusterer, 1990; Key and Runsten, 1999; MCCulloch and Ota, 2002; Humphrey et al., 2004; Minten et al., 2005; Maerterns and Swinnen, 2007), often comparing participant and nonparticipant farmers across a range of measures. Positive impacts of participation are generally associated with income generation, employment and improving access to credit and technical assistance (Glover and Kusterer, 1990; Grosh, 1994; Humphrey et al., 2004). Other studies extend the analysis to include farmlevel and/or regional spillovers related to productivity of food crop production, food security and/or development of business service markets (see for example Kennedy and Cogill, 1987; Von Braun, et al., 1989; Govereh and Jayne, 2003). Less tangible impacts, including social status and capacity development, are also reported (Reynolds et al., 2004; Masakure and Henson, 2005). The potential negative aspects of smallholder participation in highvalue markets generally relate to the institutional characteristics of supply chains and wider socio economic and environmental impacts. Thus, concerns are raised that patterns of specialization can lock smallholders into relationships that are more favorable to downstream buyers than to the farmers themselves (see for example Little and Watts, 1994; Singh, 2002). There are also concerns that the requirements of discerning buyers can act to exclude less endowed producers, with consequences for income disparities and more general social differentiation at the community level (Singh, 2002; Simmons, et al., 2005; van der Meer, 2006). Such exclusion of poorer smallholders would limit the achievement of broader economic development objectives (Reardon and Barrett, 2000; Humphrey, 2006). There are also concerns that such arrangements do not present a viable longterm approach to poverty alleviation, as market leaders will only commit to 2 sourcing from a particular base of smallholders while it is profitable to do so, yet will quickly `jump ship' to new suppliers if problems arise and/or the opportunity arises to reduce their procurement costs (Goss et al., 2000; Kolk and van Tulder, 2006). The scope for smallholders to participate in highvalue markets reflects both their own ability to compete against alternative suppliers and the willingness of buyers to procure from them in the context of their particular regulatory and commercial requirements. One strand in the literature (exemplified by McCulloch and Ota, 2002; Simmons, et al., 2005; WinterNelson and Temu, 2005; and Hernandez et al., 2007) emphasizes that smallholder farms' limited access to key resources (including cultivable land, irrigation and financial resources) inhibits investment and farm productivity. A second strand emphasizes the role of transaction costs in inhibiting small farmer participation in alternative marketing channels, including highvalue markets (Jaffee, 1995; Staal et al., 1997; Key et al., 2000; Pingali et al., 2005; Winters et al., 2005), suggesting that the stricter (and increasing) requirements of highvalue markets place smallholders at a competitive disadvantage relative to larger farmers. These assessments suggest that the lack of infrastructure, key production assets, information and/or collective action act to constrain initial smallholder entry to highvalue markets and threaten the sustainability of such participation over time. It is reasonable to assume that downstream buyers in highvalue markets will choose to procure from sources that meet their requirements at lowest cost, at manageable levels of risk. The lack of infrastructure and key assets that typify smallholder production can make it costly to meet exacting requirements and/or require that buyers have to take on certain transaction functions (Pingali et al., 2005; Poulton et al., 2006). With the restructuring of global value chains and more prominent governance role of agrifood standards (Henson and Jaffee, 2006; Humphrey, 2006), transactions are becoming more complex and there is a need for closer coordination of activity and traceability of products and raw materials within supply chains (Dolan and Humprey, 2004; Altenburg, 2006; Ruben et al., 2006). As a result, downstream buyers are looking to consolidate their supply base in favor of `preferred suppliers'. These may be largerscale producers (where they exists) unless there are offsetting reasons to continue procuring from smallholders (for example where a more decentralized supply base acts to mitigate climatic or pest/disease risks). Assessing the `success' of interventions aimed at the participation of smallholders in highvalue markets 3 Attempts to assess the efficacy of alternative interventions aimed at facilitating smallholder participation in highvalue markets are impeded by the lack of a common vocabulary with which to assess `success' versus `failure'. Key dichotomies influencing choice of indicators are, for instance, "deliverables or impacts," "technical indicators or poverty reduction indicators," "shortterm gains or longterm change" and "shortterm accountability or learning to do better" (Mebrahtu et al., 2007). Further, while we may be able to agree on the measures that `should' be used to assess the impact of interventions, often institutional constraints, especially related to project accountability (Mosse, 2004a; 2004b), limit the scope to apply these in practice. Shortterm indicators are easier to define and measure, since they are generally associated with activities and deliverables, for example in the form of services provided to beneficiaries (GDPRD, 2007). Mediumterm indicators, in terms of outcomes (for example improvements in income, enhancement of productivity and value of exports) and longterm indicators in terms of impacts (for example reductions in the incidence of poverty, environmental sustainability and gender equality) are more difficult to define. Further, there can be significant attribution problems where there are multiple interventions (often by diverse agencies) and/or turbulent economic and/or social conditions. At the same time, however, it is presumably the ultimate impacts that are most relevant to any definition of `success', while permitting measurement against country development objectives and/or poverty reduction targets (Reardon and Barrett, 2000; Morse, 2004; World Bank, 2006). De Janvry and Sadoulet (2005) contend that successful agricultural and rural development interventions require multidimensional assessment of the wellbeing of the rural poor. This suggests the need to focus on improving access to assets and understanding the context in which these are used (for example participation in domestic and international markets) and, in turn, developing enhanced livelihoods strategies. The sustainable livelihoods framework has been employed by some donors and development agencies in this vain (Ellis, 2000). Indicators of wellbeing might include increases in income, reduction of poverty and vulnerability, environmental stress, and promotion of security and equality. This implies that interventions aimed at enhancing the participation of smallholders in highvalue markets should, ideally, focus not only on indicators associated with targeted supply chains but also eventual impacts in terms of sustainable development. Putting aside the lack of a common metric of `success', it is important to recognize the range of factors influencing the efficacy of a particular intervention. While enhancing access to basic public and private assets is essential for participation in highvalue markets (Poulton et al., 2006), other factors that may strongly influence outcomes 4 include sectoral policies, entrepreneurial competence, coordination among stakeholders, and prior institutional development and technical upgrading (Wilk and Fensterseifer, 2003; GTZ, 2003; Poulton et al., 2005; Ruben et al., 2006). Further, close coordination of interventions with the private sector is becoming more critical, especially in the context of contemporary global value chains (Henson and Jaffee, 2006; Humphrey, 2006), where accurate and timely flows of information on market requirements are paramount. Who should be involved in linking smallholders to high value markets? Shifts in thinking about rural development over recent decades (Ellis and Biggs, 2001; De Janvry and Sadoulet, 2005; Reardon and Timmer, 2005), alongside the restructuring of global value chains (Dolan and Humphrey, 2004; Humphrey, 2006) have influenced the real and perceived role of particular stakeholders in bringing about the participation of smallholders in highvalue markets. For example, structural adjustment programs over the mid1980s to mid1990 focused attention on the liberalization of agricultural markets and limiting interference by government. Subsequently, however, the critical role of institutions in development processes and, in this context, the establishment and operation of effective market linkages was recognized. This is resulting in a redefinition of legitimate roles for both the public and private sectors in linking smallholders to markets. The current paradigm posits governments as `enablers', creating the conditions that facilitate and encourage the private sector to structure its supply chains to involve smallholders. Thus, the focus is on the development of policies and institutional frameworks as well as providing infrastructure and support services to enable market development (GTZ, 2003). The private sector's role is to invest in (soft as well as hard) supply chain infrastructure, develop service markets and/or transfer technical and market information to smallholders (Humphrey, 2006; Ruben et al., 2006). In this context, external development agencies and project implementers contribute by improving the planning and accountability of interventions, focusing not only on technical implementation but also on enhancing the longerterm benefits for smallholders and their livelihoods (De Janvry and Sadoulet, 2005; Wenar, 2006). Methods The aim of this study was to elicit the experiences and perspectives of practitioners and researchers on interventions aims at facilitating African smallholder participation in 5 supply chains for highervalue (for example horticultural products, spices, nuts, dairy products, etc.) and/or differentiated (for example organic or fair trade) agricultural and food product markets. Hereafter, these are collectively referred to as `highervalue supply chains'. The study focused on practitioner perceptions of: (i) the anticipated benefits/impacts of smallholder participation; (ii) the constraints that limit such participation; (iii) the indicators that actually are and those that should be used to assess project/policy impacts; (iv) key factors underpinning intervention success; and (v) the appropriate roles for various stakeholder groups. The sampling frame consisted of 391 individuals working in bilateral and multilateral donor institutions, organizations providing technical assistance, agribusiness organizations, private agribusinesses, research institutions, and nongovernmental organizations (NGOs). Defining the sample of practitioners was problematic given that no published list exists of people working in this area, including bilateral and multilateral donors, NGOs, research organizations etc. Thus, the research team compiled an initial list on the basis of an Internet search, review of the academic and grey literatures and consultation with colleagues. Respondentdriven sampling was then employed to widen the sample, whereby The identified individuals were then contacted and asked to suggest further practitioners, through respondentdriving sampling (Salganik and Heckathorn, 2004), whereby the identified respondents were asked to identify other practitioners for inclusion in the survey. While this technique could introduce biases into the sample, we would posit that this is offset by the fact that the initial listing of respondents included individuals from a wide range of organizations, cutting across the donor, academic/research, NGO and private sectors. It also served to offset any potential biases within the research team that might have influenced our own initial selection of respondents. The survey was implemented over the period October to November 2006. All potential respondents were contacted by email, inviting participation and providing a link to an internetbased questionnaire. After a period of three weeks, a reminder was sent to all potential respondents by email. Out of the initial sampling frame of 391 individuals, 49 email addresses were nonoperational, giving a valid sampling frame of 342. Of these individuals, 160 completed the questionnaire, with a response rate of 46.8 percent. The breakdown of these respondents by affiliation and geographical base is reported in Table 1.1 1 Overall, the survey provided a good crosssection of individuals, although with a lower representation from the private sector than would have been ideal. It is not possible, however, to judge the 6 Table 1. Number of respondents by affiliation: Category Frequency Based in Sub Not based in Sub Total Saharan Africa Saharan Africa Researcher 8 32 40 (5.0%) (20.0%) (25.0%) NGO 17 2 19 (10.6%) (1.3%) (11.9%) Donor 6 14 20 (3.8%) (8.8%) (12.5%) Technical assistant provider 12 30 42 (7.5%) (18.8%) (26.2%) Private agribusiness operator 7 0 7 (4.4%) (0.0%) (4.4%) Agribusiness organization 16 2 18 (10.0%) (1.3%) (11.2%) Public sector 10 2 12 (6.3%) (1.3%) (7.5%) Other 2 0 2 (1.3%) (0.0%) (1.2%) TOTAL 78 82 160 (48.8%) (51.3%) (100.0%) The questionnaire predominantly employed fivepoint Likert scales to gather attitudinal data from respondents. Thus, for example, respondents were presented with a series of items and asked the degree to which they considered each to be important on a scale ranging from "very important" (1) to "very unimportant" (5). Given the large number of items presented to respondents, a fivepoint scale was employed to minimize fatigue. While a sevenpoint scale may have produced a greater degree of discrimination between respondents, there is evidence that five and sevenpoint scales have quite comparable characteristics in terms of the mean, variance and skewness of responses (Dawes, 2008). The resulting scale scores were treated as interval data and analyzed using principle components analysis (PCA) to identify whether, within any one question, responses provided multiitem measures of underlying constructs; that is that multiple items were closely correlated with one another and really represented a broader category that had not been directly observed (McDonald, 1985; Jackson, 1991). Having identified the underlying latent variables in each case, multiitem scales were derived from the items loading heavily on to each of the derived principle components and used representativeness of the sample; we have no a priori information on the total number and institutional affiliation of practitioners engaged in linking smallscale producers to highvalue markets. 7 to derive measures that could be compared across the derived constructs and, for any one construct, across respondents. The reliability and unidimensionality of each of the multiitem scales was tested and the scales pruned as appropriate. Results This section reports the survey results according to the five key research issues in turn. Impacts of participation in highervalue supply chains To elicit the judgments of respondents regarding the potential impacts of smallholder participation in highervalue supply chains, a list of 21 items was provided that had been complied through a review of the literature and discussions among the research team. Respondents were asked to score each of these items on a fivepoint Likert scale from `very positive' (1) to `very negative' (5).2 Table 2 reports the potential impacts by ascending mean score. The most positive impacts were judged to be `access to improved technology', `understanding of market requirements', `level of technical/agronomic learning', `access to technical assistance' and `returns per unit of household labor'. Although, none of the presented impacts were rated on the negative side of the scaling (that is with a mean score above 3), the items with the highest mean scores were `level of financial risk', `health risks to farm workers', `environmental sustainability of the wider community' and `environmental sustainability of the farm enterprise', suggesting an ambiguity (`neither positive nor negative') in the perceptions of respondents as to whether smallholder participation in highervalue supply chains exposes them (and their communities) to higher levels of risk. As we will see below, indicators of risk exposure and risk management are rarely among the measures used to gauge project/policy impact. Table 2. Ordered mean scores of the potential impacts on African smallholders of participating in highervalue supply chains: Potential Impact Mean Standard Deviation Access to improved technologies 1.736a 0.319 Understanding of market requirements 1.763a 0.637 Level of technical/agronomic learning 1.794a 0.315 2 As with all of the scaling questions in the survey, respondents were given the opportunity to provide additional potential factors/impacts through an openended question. 8 Access to technical assistance 1.819a,b 0.372 Returns per unit of household labor 1.819a,b 0.333 Access to inputs 1.894b 0.356 Longterm agricultural income 1.894b 0.418 Access to finance 1.906b,c 0.352 Business management skills 1.906c 0.352 Economic impacts on wider community 2.025c.d 0.404 Ability to adapt to changes in market conditions 2.034d 0.480 Shortterm agricultural income 2.063d 0.406 Ability to be involved in viable farmerbased organizations 2.106d 0.306 Social standing in the community 2.238e 0.389 Stability/predictability of income 2.275e 0.483 Participation of women 2.301e 0.467 Food security 2.319e 0.421 Environmental sustainability of farming enterprise 2.469f 0.475 Environmental sustainability of the wider community 2.547f 0.509 Health risks to farm workers 2.699 0.360 Level of financial risk 2.965 0.401 Note: Mean scores denoted by the same letter are not significantly different from each other at the 5% level on the basis of a Wilcoxon signed rank test. In order to make further sense of the individual item scores, the Likert scale data were subject to principle components analysis (PCA). A total of three underlying factors were identified that collectively accounted for 67.9 percent of the variation in scores (See Annex Table 1). The KaiserMeyerOlkin measure of sampling adequacy was 0.752 and Bartlett's test of sphericity was significant (2 =1763.25 (66 df); p=.000), suggesting that the individual item scores factor well and that the factor solution is rigorous (Hair et al., 1998; McDonald, 1985; Jackson, 1991). Loadings on the first principal component (PC1) posit an association between participation in highervalue supply chains and enhanced access to resources and knowledge. This involves a cluster of impacts related to: `understanding of market requirements', `level of technical/agronomic learning', `access to improved technologies', `business managements skills', `access to technical skills', `enhanced ability to adapt to the changes in the markets' and `access to finance and inputs'. This factor accounted for 36.4 percent of the variation of the sample. Loadings for the second principal component (PC2) suggest an association between participation in highervalue supply chains and economic and social gains. This involves the cluster of impacts that includes: `shortterm agriculture income', `longterm agricultural income', `ability to be involved in viable farmerbased organizations', `economic impacts on wider community', `food security' and `stability/predictability of income'. Some 21.5 percent of the variation of the sample is accounted for by PC2. The 9 third principal component (PC3) can best be described as being related to the environmental and health impacts, involving such impacts as: `environmental sustainability of the wider community', `environmental sustainability of farming enterprise', `health risks of farm workers' and `participation of women'. Looking across these three principle components, it would appear that respondents judged the impact of participation in terms of the alleviation of constraints (PC1), on the one hand, and the direct (PC2) and indirect (PC3) impacts of participation on the other.3 Constraints inhibiting smallholder participation in highervalue supply chains The questionnaire explored constraints to smallholder participation from two perspectives: (i) the ability of smallholders to participate in such supply chains; and (ii) the willingness of buyers to procure from smallholders given prevailing conditions and alternatives. To capture the supply side constraints, respondents were first presented with a series of 21 potential factors constraining smallholder abilities to participate and asked to score each item on a fivepoint Likert scale from `very important' (1) to `very unimportant' (5). Following the scaling question, respondents were able to indicate further potential impacts through an openended question. The constraints considered most important (Table 3) were all related to infrastructure and logistics (for example `weak/inadequate rural transport infrastructure', `non competitive/high cost transport services', weak/inadequate rural marketing infrastructure' and `lack of scale') followed closely by a set of factors related to (weak) other support services (for example `weak/limited availability of technical advisory services', `lack of access to competitively priced finance' and `lack of pertinent information on prevailing standards'). In contrast, respondents downplayed the constraints typically associated with smallholder households (and assetholdings) themselves, including `insufficient availability of family labour', `insecure land tenure', `fragmentation of household land holdings', `risk management/food security' concerns, 3 The results of the PCA suggest that there are three broad potential impacts underlying the scaling data presented in Table 1. In order to derive measures of these three impacts, multiitem scales were defined on the basis of the critical items loading on to the relevant principle component. All three of the scales were reliable, with the Cronbach exceeding 0.81 in all cases (Spector, 1992; Aiken, 1996). According to the mean scores, access to resources and markets (1.853) was considered the most positive potential impact of participation in highervalue. Participation in such supply chains was also judged to have a positive economic and social impact (2.110). The environmental and health impacts, however, were judged to be ambiguous; the mean multiitem scale score was 2.506, corresponding to `neither positive nor negative'. 10 and `weak/absent farmer organizations'. Given that respondents suggested very few additional constraints, the list of items was considered to include all substantive supply side constraints. Table 3. Ordered mean scores for importance of constraints limiting African smallholder participation in highervalue supply chains: Constraint Mean Standard Deviation Weak/inadequate rural transport infrastructure 1.451 0.371 Noncompetitive/high cost transport services 1.622 0.425 Weak/inadequate rural marketing infrastructure 1.713a,b 0.401 Lack of scale 1.797b 0.358 Weak/limited availability of technical advisory services 1.831b 0.366 Lack of access to competitively priced finance 1.835b 0.347 Lack of pertinent information on prevailing standards 1.872b 0.353 Lack of timely access to improved inputs 1.881c 0.386 Lack of farmer `quality' consciousness 1.884c 0.408 Lack of pertinent market information 1.894c 0.369 Weak or high cost services of market intermediaries 1.919c 0.369 Limited technical knowledge of farmers 1.937c 0.307 High costs of standards certification 1.939c 0.412 Limited commercial skills of farmers 1.985c,d 0.332 Lack of business culture 1.989c,d 0.311 Lack of/limited capacity for irrigation 2.044d,e 0.414 Distrust/weak bargaining power vs commercial agribusiness entities 2.138e 0.357 Weak/absent farmer organizations 2.152e,f 0.358 Risk management concerns 2.247f 0.403 Fragmentation of household land holdings 2.312f 0.348 Insecure land tenure 2.361f 0.412 Insufficient availability of household labor 2.858 0.399 Note: Mean scores denoted by the same letter are not significantly different from each other at the 5% level on the basis of a Wilcoxon Signed rank test. Principle component analysis revealed four underlying constructs that collectively accounted for 72 percent of variation across the scaling items (Annex Table 2). The KaiserMeyerOlkin measure of sampling adequacy was 0.743 and Bartlett's test of sphericity was significant (2 =580.9 (78 df); p=.000). The first factor accounted for 31.8 percent of the variation and can be characterised as weakness of support services, with heavy loadings for `weak/absent farmers' organizations', `high costs of standards certification', `weak/limited availability of technical advisor services', `weak bargaining powers of farmers versus commercial agribusiness entities', `weak/inadequate rural transport infrastructure' and `lack of pertinent market information'. The second principal component (PC2), accounting for 21.7 percent of the sample variation can be 11 called prevailing farming structure. Pertinent constraints included `fragmentation of household land holding', `lack of scale' and `insecure land tenure' The third principle component (PC3) can be called access to finance and risk management and included the constraints of `lack of access to competitive priced finance' and `risk management concerns'. The fourth and final principle component (PC4) was related to producer knowledge and had heavy loadings for `limited technical knowledge of farmers' and `lack of farmer quality consciousness'.4 To capture demandside constraints, respondents were presented with a list of 15 possible factors that might inhibit the willingness and/or ability of processing, export trading and/or modern retail distribution companies from sourcing highervalue products or raw materials from African smallholders. Again, respondents were asked to score each factor on a fivepoint Likert scale from `very important' (1) to `very unimportant' (5). The constraints that are considered most important all related to logistics and supply chain management, including the specific constraints of: `weak/inadequate rural transport infrastructure', `inability/high costs of product traceability', `lack of scale among smallscale farmers', `inability to enforce commercial contracts' and `high transport costs for raw materials' (Table 4). The next set of relatively important constraints also impact on the transaction costs potentially faced by buyers, associated with `weak communications', `lack of reliable/timely information on production', `inability to recover loans from farmers', and `lack of confidence in regulatory enforcement related to pesticides and veterinary drugs'. Table 4. Ordered mean scores for constraints inhibiting African processing, export trading and modern retail distribution companies from sourcing from smallholders: Constraint Mean Standard Deviation Weak/inadequate rural transport infrastructure 1.611 0.304 4 To obtain measures of the importance of each of these four underlying constraints to smaller participation, multi item scales were derived from the individual item scores. Three of the four scales had high levels of reliability, all with a Cronbach exceeding 0.75. The scale for prevailing farming structure had a value of 0.67, indicating lower levels of reliability. Attempts to prune this scale did not produce a higher value. The constraints to the participation of smallscale producers judged to be most important related to lack of producer knowledge (1.812) and weakness of support services (1.819), the mean scores of which were not significantly different (p <.005). Access to finance and risk management (2.337) and prevailing farming structure (2.791) were considered to be less important constraints. 12 Inability/high costs of product traceability 1.720 0.319 Lack of scale among smallscale farmers 1.836a 0.328 Inability to enforce commercial contracts 1.855a 0.362 High transport costs for raw materials 1.861a 0.312 Weak communications infrastructure/facilities 1.971b 0.323 Lack of reliable/timely information on production 2.024b 0.331 Inability to recover loans from farmers 2.036b 0.405 Lack of confidence in enforcement of regulations related to 2.045b 0.414 pesticides and veterinary drugs High cost of grading and testing delivered products 2.165c 0.336 Limited technical knowledge of farmers 2.179c,d 0.371 Weak farmer organizations 2.180c,d 0.329 Lack of trust in market intermediaries 2.280d 0.365 Lack of/limited farmer irrigation capacity 2.442 0.354 Regulatory restrictions on direct sourcing from smallholders 2.853 0.432 Note: Mean scores denoted by the same letter are not significantly different from each other at the 5% level on the basis of a Wilcoxon signed rank test. Three underlying factors were identified in the scaling responses through the application of principle component analysis. Collectively, these accounted for 64.1 percent of the variation across the items (Annex Table 3). The KaiserMeyerOlkin measure of sampling adequacy was 0.735 and Bartlett's test of Sphericity was significant ((2 =439.9 (55 df); p=.000. The first principle component (PC1) accounted for 34.8 percent of the variation and was associated with access and cost of infrastructure. The second principle component (PC2), accounting for 19.3 percent of the sample variation can be called supply chain management. The third and final principle component was associated with food safety and quality control.5 Indicators of the success of project or policy interventions Respondents were presented with a list of 28 potential indicators of the `success' of project or policy interventions aimed at facilitating smallholder participation in higher value supply chains. They were asked to score each item according to frequency of use on a fivepoint Likert scale from `very frequently' (1) to `very infrequently' (5). 5 Measures of the importance of each of the three underlying constraints were derived from the associated multi item scales. The Cronbach exceeded 0.74 for two of the scales. The scale for supply chain management had a value of 0.65, indicating lower levels of reliability. The constraints considered of greatest importance were food safety and quality controls (1.990) and access and cost of infrastructure (1.993), which were statistically insignificant (p <.005). Supply chain management (2.439) was considered to be of secondary importance. 13 The responses provided point to the relatively shortterm time horizon of many development assistance efforts and to the difficulties of gauging deeper and longer term structural and capacity impacts (Table 5). For example, among the indicators most commonly used to measure project/policy success are the numbers of: (i) targeted farmers supplying the focal products; (ii) farmer groups formed; and (iii) farmers/groups that have been certified to a particular standard. Other commonly used indicators relate to changes in the quantities/value of farmer sales and country exports. In contrast, indicators related to risk management (for example `farm worker health', `interannual variability of farmer incomes' and `improved natural resource management') are much less commonly considered. It is therefore not surprising that respondents remain ambiguous on matters of farmer/community risk exposure. Also much less commonly measured are indicators of local capacitybuilding, whether among public sector organizations or (private) suppliers of technical and business advisory services. Given that the respondents also indicated that weak support services are among the leading constraints on the ability of smallholders to meet the requirements of highervalue supply chains, the relative lack of attention to the impact of interventions on service capacities is somewhat surprising. As with the other themes, principle component analysis was employed to identify any underlying constructs in the scaling data. A total of four factors accounted for 79.5 percent of the sample variation (Annex Table 4). The KaiserMeyerOlkin measure of sampling adequacy was 0.687 and Bartlett's test of sphericity was significant (2 =987.28(136df); p=.000), suggesting a rigorous factor solution. The four principle components are associated with managerial and marketing capacity (accounting for 27.6 percent of sample variation), productivity gains (23.1 percent), technical assistance support (20.8 percent), and number of participants (8.8 percent) (See Annex Table 4).6 6 To derive measures of the frequency of use for the four broader indicators of success, multiitem scales were derived on the basis of the item loadings for each principle component. All of these scales were reliable; the Cronbach exceeded 0.75 for all four of the scales. Respondents judged that the most frequently used measure of `success' related to the number of recipients of the intervention (1.712). This was followed by measures of the technical assistance and support provided (2.291). Measures of productivity gains (2.899) and managerial and marketing capacity (3.013) were ranked third in terms of frequency of use and the scale values were statistically insignificant (p <.005) from one another. As highlighted earlier, the measures judged to be used least frequently was impacts on external (organizational) capacity (3.612). 14 Table 5. Ordered mean scores for indicators used to measure success of project or policy interventions linking smallholders to highervalue supply chains: Indicator Mean Standard Deviation Number of targeted farmers supplying the highervalue crop/livestock product 1.841a 0.314 Quantities of technical assistance provided (for example number of people/groups 1.844a 0.341 trained, etc.) Gross value of sales of smallholders from focal commodities 1.968b 0.527 Number of farmer groups formed 1.973b,c 0.331 Volume/value of exports or domestic sales 2.101c 0.327 Measures of technical farm productivity (for example crop yields, animal 2.103c 0.503 productivity, etc.) Number of farmers/groups certified for a particular standard 2.323d 0.335 Trends in producer prices, share of value addition, etc. 2.354d 0.310 Distribution of participation based on gender 2.467d,e 0.315 Share of targeted commodity sold at higher grades or in targeted segment 2.497d,e,f 0.364 Evidence of improved access to technical or business advisory services 2.583f,g 0.299 Changes in measured rate of poverty among target farmers 2.606f.g 0.310 Crop/livestock budgets showing change in income compared to `traditional' 2.670g,h 0.328 crop/activity Measures of agribusiness profitability 2.673g,h 0.330 Evidence of increased viability and capacity of farmer organizations 2.697i 0.463 Measures of technical performance related to compliance with standards (for 2.780i,j 0.312 example rejection rates, incidence of price discounts, etc.) Returns to land and labour invested in focal commodities 2.987j,k 0.466 Indicators of reduced transaction costs faced by agribusiness firms 3.029k,l 0.463 Indicators of reduced transaction costs faced by farmers 3.051l 0.495 Indicators of ability of farmers to continue supplying target markets in the medium 3.111l,m 0.517 and long term Evidence of increased capacity/sustainability of local providers of technical/business 3.161l,m 0.589 advisory services Indicators of intraannual cash flow for farmers 3.190m 0.415 Evidence of increased ability to respond to new market opportunities 3.213m,n 0.404 Measures of improved natural resource management by farmers 3.219m,n 0.459 Impacts on wider community (for example access to input, flows of knowledge, etc.) 3.278n 0.331 Data on interannual variability of farmer incomes 3.320o 0.355 Measures of improved farm worker health 3.603o 0.410 Measure of public sector administrative/technical capacities 3.674o 0.532 Note: Mean scores denoted by the same letter are not significantly different from each other at the 5% level on the basis of a Wilcoxon signed rank test Respondents were then asked what indicators of success `should' be tracked and measured in order to truly gauge the impact of project and policy interventions in this 15 area.7 The indicators most frequently mentioned were the `number of farmers able to supply/certified to required standards on a consistent basis', `change in level/stability of farm income', `change in household welfare/standards of living/livelihood/poverty' and `longterm ability of smallholders to supply target markets' (Table 6 below). Many of these indicators were judged by respondents to be used relatively infrequently in practice, and indeed did not even figure in the four broad indicators derived through principle components analysis. While just over onethird of respondents included the viability/strengthening of farmer organizations among the indicators that `should' be gauged, very few included indicators associated with service provider or public sector capacities. Factor influencing success of interventions What factors most influence the performance or effectiveness of interventions in this area? Respondents were presented with a list of 30 potential factors, derived from research team discussions following a review of the literature. Respondents were asked to score each of these factors on a fivepoint Likert scale from `very important' (1) to `very unimportant' (5). Subsequently, they were provided with the opportunity to provide additional factors through an openended question. The factors judged to have the greatest impact on intervention success all are closely related to project design and management, including `close collaboration with the private sector', , `flexible project design enabling modifications and redirections over time', `technical expertise of project advisory service provider', and `management of financial support to farmers' (Table 7). In contrast, the collaborative involvement of government (either local or central agricultural ministry) was deemed to be least important for intervention success. This is not surprising given the current paradigm of governmentasenabler, but not direct market participant. It is somewhat surprising that several factors associated with prior experience (for example farmer experience with focus commodity or with producer organizations) and/or prior institutional capacities (for example strong industry organizations) were not rated among the leading ten to fifteen factors impacting intervention performance. 7 Each was asked to provide a ranked list of up to six indicators. The verbatim responses were coded by two members of the research team independently, with the aim of achieving a minimum level of intercoder reliability of 80 percent. Where there was discrepancy in the coding of any item, the two research team members consulted and agreed on an appropriate categorization. The defined categories were assigned numerical values according to the ranking from first (1) to last (6). The elicited items were then ordered by their total rank score in order to obtain some ranking in terms of importance. 16 Table 6. Nonelicited primary indicators that should be used to judge success of interventions linking smallholders to highervalue supply chains: Indicator Frequency Number of farmers able to supply/certified to required standards on consistent basis 142 Change in level/stability of farm income 122 Change in household welfare/standards of living/livelihood/poverty 118 Longterm ability of smallholders to supply target markets 106 Change in volume/value of sales/valueaddition, etc. 80 Change in farm profitability/returns to land/labor 78 Level/stability of prices attained 63 Viability/strengthening of farmer organizations 56 Change in number of farmers supplying target market/linked to supply chain 53 Change in ability to understand and/or respond to market opportunities 46 Change in supply chain transaction costs 43 Degree to which commercial/technical skills enhanced 33 Views/attitudes of buyers regarding smallholders/buyer performance 30 Infrastructural/supply chain development 26 Change in smallholder competitiveness/bargaining power 26 Level of buyer/consumer demand for products of smallholders 23 Number/area of smallholders 23 Degree to which use of technology enhanced 20 Proportion of market supply/exports supplied by smallholders 17 Inclusiveness in terms of gender, poverty level, etc. 17 Impact on wider community 17 Development of support services/access to support services 17 Regularity of payments to smallholders 13 Donor return on investment 13 Replicability/ability to scaleup 10 Other 36 As previously, in order to identify underlying constructs across the 30 scale items, principle component analysis was employed. A total of six factors were identified that collectively accounted for 74.3 percent of the variation. The factor solution was judged to be rigorous on the basis of the KaiserMeyerOlkin measure of sampling adequacy (0.749) and Bartlett's test of sphericity, which was significant (2 =931 (171 df); p=.000). The six principle components are associated with prevailing external conditions (accounting for 34.1 percent of sample variation), project management (11.3 percent), effective institutions (9.2 percent), role of the private sector (7.0 percent) level of farmer experience (6.5 percent) and ability to adapt to local conditions (6.2 percent).8 8 Multiitem scales were derived for each of the six underlying factors influencing intervention success. Five of the scales had a Cronbach exceeding 0.74. However, the scale corresponding to PC2 (`project management') required pruning to increase reliability. Namely, the item `appropriate timing of intervention in the `lifecycle' of an industry' was excluded. The pruned scale had a value of 0.87. The factors judged to be most important influences on the success of interventions related to ability to adapt to local conditions (1.722), project management (1.756) and role of 17 Table 7. Ordered mean scores for factors contributing to the success of interventions linking smallholders to highervalue supply chains: Factor Mean Standard Score Deviation Project interventions made in close collaboration with the private sector 1.428 0.349 Flexible project design enabling modifications and redirections over time 1.556a 0.303 Technical expertise of project advisory service provider 1.624a,b 0.465 Appropriate management of financial support to farmers 1.673b 0.390 Commercial expertise of project advisory service provider 1.717c 0.402 Security situation in targeted production areas 1.780c 0.468 National sociopolitical stability 1.827c 0.336 Clear `exit' and transition strategy beyond project interventions 1.832c 0.401 Condition of rural roads during rainy season 1.839c 0.376 Effective accountability and controls on corruption 1.858d 0.329 Sharing of project intervention costs with benefiting firms 1.931d 0.408 Agroecological conditions in targeted producer areas 1.940d 0.385 Targeting farmers prior experience in producing focal /similar commodity 1.950d 0.473 Absence of adverse weather shock (i.e. drought) 1.964d 0.297 Coherent ownership structure for farmer organizations 1.971d 0.326 Project interventions initiated by the private sector 1.991d 0.375 Positive trends in pertinent commodity prices 2.060e 0.319 Appropriate timing of intervention in the `lifecycle' of an industry 2.061e 0.484 Targeting farmers with prior experience producing any commodity for a discerning, 2.067e 0.507 qualityfocused market Sharing of project intervention costs with benefiting farmers 2.069e 0.402 Extended length of time for project implementation 2.161e 0.343 Existence of strong industry organization in the targeted sectors 2.162e 0.353 Prevailing currency exchange rates 2.168e 0.376 Welldefined property rights 2.238e,f 0.319 Requirement for financial contribution from farmers 2.252e,f,g 0.353 Strong enforcement institutions 2.287f,g 0.331 Prevailing local interest rates 2.342f,g 0.364 Targeting farmers with prior positive experience of producer organizations 2.379g 0.374 Collaborative involvement by local government authorities 2.408g 0.316 Collaborative involvement by national ministry of agriculture 2.715 0.426 Note: Mean scores denoted by the same letter are not significantly different from each other at the 5% level using Wilcoxon signed rank test the private sector (1.781). The mean values of these three scales were statistically insignificant (p <.005). Prevailing external conditions (2.151) and effective institutions (2.166) were considered of secondary importance and statistically insignificant from one another (p <.005). Contrary to a priori expectations, farmer experience (2.651) was considered to be the least important factor influencing the success of these interventions. 18 Institutional roles in facilitating smallholder participation Finally, the survey asked respondents to consider the most important roles of a series of stakeholders in facilitating smallholder participation in highervalue supply chains. These stakeholders are: 1) national governments; 2) private agribusiness firms or organizations; 3) external development assistance agencies; and 4) project implementation teams or organizations. In each case, respondents were asked to indicate the three most important roles and their verbatim responses were coded as described above. Overall, the results suggest a clear division of labor between national governments and the private sector, although more ambiguity in the division of roles between the latter and external agencies and project implementing teams (Table 8). Thus, government should be providing the required logistics and infrastructure (including road networks, air and sea ports, utilities and communications services), an enabling environment, and a coherent legal framework (especially relating to contract enforcement). According to respondents, private agribusinesses are not only the `target market' for smallscale producers, but also central players in facilitating the ability of smallholders to meet their requirements, via technical support, providing access to finance and managing supply chain logistics. Less than one in four of the respondents listed training and extension, research, or providing access to finance, among the roles that governments should play in facilitating smallholder participation in highervalue supply chains. These are traditionally areas deemed to have public good properties, experiencing market failure in an African context. The provision of training and access to finance are more widely reported by respondents as appropriate functions for private agribusinesses, external development assistance agencies, and project implementing teams. The roles deemed appropriate for external development organizations and project implementation teams are partly to support governments and private agribusinesses in pursuing their own functions and partly in the realm of `honest' broker between and among supply chain participants. These entities are also seen to have important roles in helping to identify and assess market opportunities (for farmers and firms) and sharing/promoting international `good practice' in pertinent areas. Quite a few respondents signaled a role for donors and project implementers in capacitybuilding, although, as noted above, achievements in capacitybuilding are rarely monitored and measured. 19 Table 8. Most important roles for stakeholders in enabling smallholders to participate in highervalue supply chains: Role Frequency National Government Development of logistics/infrastructure 76 Promote enabling environment 63 Legal framework/conditions for contract enforcement 52 Policy framework 48 Training/extension 39 Access to capital/finance 32 Macroeconomic management 31 Farm organization/group development 21 Law & order 19 Research 14 Other 42 Private Agribusiness Firms or Organizations Technical support/capacitybuilding 82 Creating market linkages 77 Access to fixed/working capital 58 Identification & dissemination of market opportunities 49 Compliance with food safety & quality standards 34 Input supply 33 Providing stable market demand & revenue 27 Relations with government 19 Other 32 External Development Assistance Agencies Linking smallholders and agribusinesses 69 Support to government/promoting policy reform 59 Training/capacitybuilding 57 Provision of fixed/working capital 55 Experiencesharing/promoting best practice 48 Linking government and private sector 36 Business development service creation/capacitybuilding 34 Market identification/creation 32 Advocacy for smallholders 20 Supporting infrastructural development 19 Supporting standardssetting, certification, etc. 17 Other 33 Project Implementation Teams or Organizations Linking/brokering smallholders and private sector 78 Training/capacitybuilding 61 Market identification, assessment, communication, etc. 59 Farmer mobilization, organization, etc. 46 Technical support 41 Promoting good practice 26 Project management, organization, planning, monitoring, etc. 23 Information source 21 Other 29 20 Conclusions The potential and pitfalls of linking smallholder farmers to higher value markets are growing themes within economic research and policy planning associated with developing country agriculture. For example, the World Development Report (WDR) 2008 notes the potential of the "emerging new agriculture". This is "led by private entrepreneurs in extensive value chains linking producers to consumers and including many entrepreneurial smallholders supported by their organizations". Dynamic opportunities exist in domestic, regional, and international markets. Yet, the WDR also identifies frequent constraints on smallholder participation in these more dynamic markets, emphasizing weaknesses in infrastructure, mismatches in delivery scale or scope when supplying modern procurement systems (that is of domestic supermarkets) and the challenges associated with compliance with emerging food safety and other standards (especially for exports). For the present study, respondent practitioners similarly downplayed production level constraints and emphasized the challenges of efficient transactionmaking between (African) smallholder farmers and the downstream gatekeepers of higher value supply chains. Responses from these practitioners did, however, highlight one frequently constraining factor of production; that is, farmer knowledge about market opportunities and requirements and about effective and marketacceptable production techniques. The existing literature provides numerous examples of where smallholder participation in higher value markets has been associated with higher levels of income and/or employment. However, without detailed and longer term studies examining patterns of participation and nonparticipation/exit, variability of outcomes over time, and a broader range of impacts, it is difficult to draw definitive conclusions. The respondents to our practitioner survey were most optimistic about the `learning' benefits (e.g. acquisition of technical and business skills) from smallholder participation in higher value supply chains. They were also positively inclined regarding income gains and other economic impacts. With regard to environmental and health impacts there was greater ambivalence, with the sample as a whole having a neutral position, while a considerable minority noting potential adverse impacts. The practitioner survey brings out an interesting contrast between the sets of indicators typically used to monitor and evaluate the efficacy of interventions to better link smallholders with higher value markets, and those indicators which respondents indicate `should' be used in order to best gauge impacts. The most frequently used measures relate to `inputs' (for example number of people trained; number of groups 21 formed) and selected shortterm outcomes (for example volume and value of sales), rather than to longer term impacts or indicators of effective capacity. The potential sustainability of results, relationships, services, capacities, etc. does not seem to be commonly gauged. Pragmatic considerations of measurement ease and timeliness are seemingly determinate. Survey respondents asserted that project design and management features are the lead determinants of the success or otherwise of interventions in this field. They downplayed other factors, including prevailing commodity market and macroeconomic conditions, the prior experience of farmers, and the preexisting strength of pertinent industry organizations. The sample as a whole substantially discounted the importance of a collaborative and effective direct role for local or national government agencies. Whether these perspectives are borne out in practice or simply reflect the hopes and aspirations of the development community is an empirical issue that is in need of testing. Finally, the survey respondents outlined a division of public and private sector responsibilities consistent with the current `linking farmers to markets' paradigm. Thus, the roles of government are cast as providing the economic, political, and infrastructural conditions necessary for private investment. The private sector, in turn, is tasked with the responsibility for driving the integration of producers into higher value markets via commercial relationships and associated provision of market information, technical advice, and logistical and other services. External development agencies and project implementers are tasked with supporting the efforts of governments and the private sector, while playing the potentially multi faceted role of `honest broker', between and among supply chain participants and government entities. In the "emerging new agriculture", intrasupply chain trust is critical, certain suppliers are deemed `preferred', public/private partnerships are sought, and economizing on transaction costs is paramount. In such an environment, an effective `honest broker' can be a valuable player. 22 References Aiken, L.R. (1996). Rating Scales and Checklists: Evaluating Behaviour, Personality and Attitudes. New York: John Wiley & Sons Ltd. Altenburg, T. (2006). Governance patters in value chains and their development impact. The European Journal of Development Research, 18(4), 498521. Dawes, J. (2008), Do Data Characteristics Change According to the Number of Scale Points Used? An experiment using 5point, 7point and 10point scales. International Journal of Market Research, 50 (1), 6177. De Janvry, A. and Sadoulet, E. (2005). Achieving success in rural development: Toward implementation of an integral approach. Agricultural Economics, 32(1s), 7589. Dolan, C. and Humprey, J. (2004). Changing governance patters in the trade in fresh vegetables between Africa and the United Kingdom. Environmental and Planning, 36, 491509. Ellis, F. (2000). Rural Livelihoods and Diversity in Developing Countries. Oxford, UK: Oxford University Press. Ellis, F. and Biggs, S. (2001). Evolving themes in rural development 1950s2000s. Development Policy Review, 19(4), 437448. GDPRD. (2007). Core Indicators for Agriculture and Rural Development Study. Global Donor Platform for Rural Development. Retrieved 10 October 2007, from http://www.donorplatform.org/ Glover, D. and Kusterer, K. (1990). Small Farmers, Big Business. New York: St. Martin's Press. Gomiero, T., Giampietro, M. and Mayuni, K. (2006). Facing complexity on agro ecosystems: A new approach to farming system analysis. International Journal of Agricultural Resources, Governance and Ecology, 5(2/3), 116144. 23 Goss, J., Burch, D. and Rickson, R. (2000). Agrifood restructuring and third world transnationals: Thailand, the CP group and the global shrimp industry. World Development, 28(3), 513530. Govereh, J. and Jayne, T. S. (2003). Cash cropping and food crop productivity: synergies or tradeoffs? Agricultural Economics, 28, 3950. Grosh, B. (1994). Contract farming in Africa: An application of the new institutional economics. Journal of African Economies, 3(2), 231261. GTZ (2003). Guide to Rural Economic and Enterprise Development. Working Paper No. 10. Eschborn, Germany. Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C. (1998). Multivariate Data Analysis. Upper Saddle River: Prentice Hall. Hazell, P. (2005). Is there a future for small farmers? Agricultural Economics, 32(1s), 93 101. Henson, S. and Jaffee, S. (2006). Food safety standards and trade: Enhancing competitiveness and avoiding exclusion of developing countries. European Journal of Development Research, 18(4), 593621. Hernandez, R., Reardon, T. and Berdegué, J. (2007). Supermarkets, wholesalers, and tomato growers in Guatemala. Agricultural Economics, 36, 281290. Humphrey, J., McCulloch, N. and Ota, M. (2004). The impact of European market changes on employment in the Kenyan horticultural sector. Journal of International Development, 16, 6380. Humphrey, J. (2006). Policy implications of trends in agribusiness value chains. The European Journal of Development Research, 18(4), 572592. Jackson, J.E. (1991). A User's Guide to Principle Components. New York: John Wiley and Co. Ltd. 24 Jaffee, S. (2005). Transaction costs, risk and the organization of private sector food commodity systems. In: Jaffee, S. and Morton, J. (eds). Marketing Africa's HighValue Foods: Comparative Experiences of an Emergent Private Sector. Dubuque, Iowa: Kendall/Hunt Publishing Company. Jaffee, S. and M. Sewadeh (2006) The Changing Composition of Developing Country AgroFood Trade and the Changing Standards Landscape. Paper prepared for World Bank/USAID Elearning Course on Trade and Standards. Washington DC: World Bank. Kennedy, E. T. and Cogill, B. (1987). Income and Nutritional Effects of the Commercialization of Agriculture in Southwestern Kenya (Research Report No. 63). Washington, DC: International Food Policy Research Institute Key, N. and Runsten, D. (1999). Contract farming, smallholders, and rural development in Latin America: The organization of agroprocessing firms and the scale of outgrower production. World Development, 27(2), 381401. Key, N., Sadoulet, E. and De Janvry, A. (2000). Transaction costs agricultural household supply response. American Journal of Agricultural Economics, 82(2), 245259. KIT (2006). Chain Empowerment: Supporting African Farmers to Develop Markets. KIT, Amsterdam. Kolk, A. and van Tulder, R. (2006). Poverty alleviation as business strategy? Evaluating commitments of frontrunner multinational corporations. World Development, 34(5), 789801. Little, P. D. and Watts, M. J. (Eds.). (1994). Living under contract: Contract farming and agrarian transformations. Madison, WI: The University of Wisconsin Press. Maertens, M. and Swinnen, J.F.M. (2007). Trade, Standards and Poverty: Evidence from Senegal. LICOS Discussion Paper 177/2007. Katholieke Universiteit Leuven, Leuven. 25 Masakure, O. and Henson, S. (2005). Why do smallscale producers choose to produce under contract? Lessons from nontraditional vegetable exports from Zimbabwe. World Development, 33(10), 17211733. McCulloch, N. and Ota, M. (2002). Export Horticulture and Poverty in Kenya (Working Paper 174). Sussex, UK: Institute of Development Studies. McDonald, R. (1985). Factor Analysis and Related Methods. Hillsdale: Erbaum. Mebrahtu, E., Pratt, B. and Lönnqvist, L. (2007). Rethinking monitoring and evaluation: Challenges and prospects in the changing global aid environment. Oxford, UK: INTRAC. Messner, D. (2004). Region in the `world economic triangle.' In: H. Schmitz (Ed.), Local Enterprises in the Global Economy: Issues of Governance and Upgrading (2052). Cheltenham, UK: Edward Elgar Publishing, Inc. Minten, B., Randrianarison, L. and Swinnen, J.F.M. (2006). Global Retail Chains and Poor Farmers: Evidence from Madagascar. LICOS Discussion Paper 164/2006. Katholieke Universiteit Leuven, Leuven. Morse, S. (2004). Indices and Indicators in Development: An Unhealthy Obsession with Numbers. London: Earthscan Publications Ltd. Mosse, D. (2004a). Is Good Policy Unimplementable? Reflections on the Ethnography of Aid Policy and Practice. Development and Change, 35 (4), 639671. Mosse, D. (2004b). Cultivating Development: An Ethnography of Aid Policy and Practice. London: Pluto Press. Page, S. and Slater, R. (2003). Small Producer Participation in Global Food Systems: Policy Opportunities and Constraints. Development Policy Review, 21 (56), 641654. Pelupessy, W. and van Kempen, L. (2005). The impact of increased consumer orientation in global agrifood chains in smallholders in developing countries. Competition and Change, 9(4), 357381. 26 Pingali, P., Khwaja, Y. and Meijer, M. (2005). Commercializing small Farms: Reducing transaction costs. In The Future of Small Farms: Proceedings of a Research Workshop. International Food Policy Research Institute, Washington DC. Poulton, C., Dorward, A. and Kydd, J. (2005). The future of small farms: New directions for services, institutions and intermediation. In The Future of Small Farms: Proceedings of a Research Workshop. International Food Policy Research Institute, Washington DC. Poulton, C., Kydd, J. and Dorward, A. (2006). Overcoming market constraints on pro poor agricultural growth in SubSaharan Africa. Development Policy Review, 24(3), 243 277. Reynolds, L. T., Murray, D. and Taylor, P. L. (2004). Fair trade coffee: Building producer capacity via global networks. Journal of International Development, 16, 11091121. Reardon, T. and Barrett, C. B. (2000). Agroindustrialization, globalization, and international development: An overview of issues, patters, and determinants. Agricultural Economics, 23, 195205. Reardon, T. and Timmer, P. (2005). Transformation of markets for agricultural output in developing countries since 1950: How has thinking changed? In: Evenson, R., Pingali, P. and Schultz, T.P. (Eds.). In Handbook of agricultural economics: Agricultural development, farmers, farm production and farm markets. Amsterdam: Elsevier Press. Reardon, T., Timmer, P. and Berdegué, J. (2004). The rapid rise of supermarkets in developing countries: Induced organizational, institutional, and technological change in agrifood systems. Electronic Journal of Agricultural and Development Economics, 1(2), 168183. Ruben, R., Slingerland, M. and Nijhoff, H. (2006). Agrofood chains and networks for development: Issues, approaches and strategies. In R. Ruben, M. Slingerland, and H. Nijhoff (Eds.). In Agrofood chains and networks for development (pp. 125). Wageningen, The Netherlands: Springer. 27 Salganik, M.J. and Heckathorn, D.D. (2004). Sampling and Estimation in Hidden Populations Using RespondentDriven Sampling. Sociological Methodology, 34, 193239. Shepherd, A. (2007). Approaches to Linking Producers to Markets. Agricultural Management, Marketing and Finance Occasional Paper 13, Food and Agriculture Organisation, Rome. Simmons, P., Winters, P. and Patrick, I. (2005). An analysis of contract farming in East Java, Bali, and Lombok, Indonesia. Agricultural Economics, 33, 513525. Singh, S. (2002). Contracting out solutions: Political economy of contract farming in the Indian Punjab. World Development, 30(9), 16211638. Spector, P.E. (1992). Summated Rating Scale Construction. Newbury Park: Sage Publications. Staal, S., Delgado, C. and Nicholson, C. (1997). Smallholder dairying under transaction costs in east Africa. World Development, 25(5), 779794. Temu, A.E. and Temu, A.A. (2005). HighValue Agricultural Products for Smallholder Markets in SubSaharan Africa: Trends, Opportunities and Research Priorities. Paper prepared for International Workshop on `How Can the Poor Benefit from the Growing Market for HighValue Agricultural Products?' CIAT, October 35, 2005. Van der Meer, C. (2006). Exclusion of smallscale farmers from coordinated supply chains: Market failure, policy failure or just economies of scale? In R. Ruben, M. Slingerland, and H. Nijhoff (Eds.). In Agrofood chains and networks for development. Wageningen, The Netherlands: Springer. Von Braun, J., Hotchkiss, D. and Immink, M. (1989). Nontraditional Export Crops in Guatemala: Effects on Production, Income, and Nutrition. (Research Report No. 73). Washington, DC: International Food Policy Research Institute Weatherspoon, D., Cacho, J. and Christy, R. (2001). Linking globalization, economic growth and poverty: Impacts of agribusiness strategies on SubSaharan Africa. American Journal of Agricultural Economics, 83(3), 722729. 28 Wenar, L. (2006) Accountability in international development aid. Ethics and International Affairs, 20 (1). 123. Wilk, E. and Fensterseifer, J. (2003). Towards a national agribusiness system: A conceptual framework. International Food and Agribusiness Management Review, 6(2), 99110. Williams, S. and Karen, R. (Eds.). (1985). Agribusiness and SmallScale Farmers: A Dynamic Partnership for Development. Boulder, CO: Westview Press, Inc. WinterNelson, A. and Temu, A. (2005). Liquidity constraints, access to credit and pro poor growth in rural Tanzania. Journal of International Development, 17, 867882. Winters, P., Simmons, P. and Patrick, I. (2005). Evaluation of a hybrid seed contract between smallholders and a multinational company in East Java, Indonesia. Journal of Development Studies, 41(1), 6289. World Bank (2006). Selecting Monitorable Indicators for Agricultural and Rural Development Programs: Measuring Results in Lessthan Ideal Conditions. World Bank (2007). World Development Report 2008: Agriculture for Development. World Bank, Washington DC. 29 Annex Table 1. Results of principle components analysis of scaling data on potential impacts of smallholder participation in highervalue supply chains: Potential Impact PC1: PC2: PC3: Access to Economic & Environmental Resources & Social Gains & Health Knowledge Impacts Understanding of market requirements 0.8240 0.0240 0.1522 Level of technical/agronomic learning 0.7131 0.1021 0.0308 Access to improved technologies 0.7001 0.2655 0.1373 Business management skills 0.6907 0.0009 0.3532 Access to technical assistance 0.6781 0.2931 0.0383 Ability to adapt to changes in market conditions 0.6388 0.1835 0.2064 Access to finance 0.6489 0.1828 0.2585 Access to inputs 0.5897 0.3529 0.0814 Shortterm agricultural income 0.2001 0.7435 0.1246 Longterm agricultural income 0.1797 0.7423 0.1305 Ability to be involved in viable farmerbased organizations 0.2107 0.6949 0.0508 Economic impacts on wider community 0.1493 0.6718 0.3983 Food security 0.0021 0.6329 0.2303 Stability/predictability of income 0.3095 0.6079 0.3572 Environmental sustainability of the wider community 0.2531 0.2540 0.7902 Environmental sustainability of farming enterprise 0.3928 0.2653 0.7686 Health risks to farm workers 0.2881 0.1815 0.7677 Participation of women 0.0948 0.3058 0.7336 30 Annex Table 2. Results of principle components analysis of scaling data on importance of constraints limiting African smallholder participation in highervalue supply chains: Constraint PC1: PC2: PC3: PC4: Weakness Prevailing Access to Lack of of Farming Finance & Producer Support Structure Risk Knowledge Services Management Weak/absent farmer organizations 0.8181 0.0275 0.2473 0.0196 High costs of standards certification 0.7522 0.2007 0.0763 0.0051 Weak/limited availability of technical advisory 0.6964 0.0663 0.0448 0.4766 services Weak bargaining power vs. commercial 0.6632 0.2907 0.3676 0.1168 agribusiness entities Weak/inadequate rural transport infrastructure 0.6123 0.2529 0.4243 0.0880 Lack of pertinent market information 0.5902 0.0351 0.4798 0.3658 Fragmentation of household land holdings 0.2054 0.7784 0.0446 0.0187 Lack of scale 0.1512 0.6681 0.1603 0.1658 Insecure land tenure 0.0799 0.6597 0.1978 0.2288 Lack of access to competitively priced finance 0.1474 0.0053 0.7631 0.2495 Risk management concerns 0.1699 0.1022 0.6705 0.0653 Limited technical knowledge of farmers 0.0649 0.0092 0.1084 0.8552 Lack of farmer `quality' consciousness 0.0007 0.2426 0.1819 0.5997 31 Annex Table 3. Results of principle component analysis of scaling data on importance of constraints inhibiting African processing, export trading and modern retail distribution companies from sourcing from smallholders: Issues PC1: PC2: PC3: Access & Cost Supply Chain Food Safety & of Management Quality Infrastructure Controls Weak/inadequate rural transport infrastructure 0.8533 0.0750 0.1728 High transport costs for raw materials 0.7475 0.0842 0.1964 Lack of/limited farmer irrigation capacity 0.6346 0.0624 0.2619 Weak communications infrastructure/facilities 0.5910 0.4311 0.0324 Inability to enforce commercial contracts 0.0046 0.8441 0.0326 Inability to recover loans from farmers 0.0477 0.7697 0.0826 Lack of trust in market intermediaries 0.2624 0.5273 0.3248 Inability/high costs of product traceability 0.0186 0.0976 0.8080 Lack of confidence in enforcement of regulations 0.2252 0.0381 0.7252 related to pesticides and veterinary drugs High cost of grading and testing delivered products 0.4078 0.1590 0.6388 32 Annex Table 4. Results of principle components analysis of scaling data for importance of indicators used to measure success of project or policy interventions linking smallholders to highervalue supply chains: Issues PC1: PC2: PC3: PC4: Managerial & Productivity Technical Number of Marketing Gains Assistance Recipients Capacity and Support Measures of reduced transaction costs faced by agribusiness firms 0.7711 0.0330 0.2124 0.0792 Indicators of reduced transaction costs faced by farmers 0.7667 0.0047 0.2005 0.1106 Indicators of ability of farmers to continue supplying target markets in 0.7564 0.2431 0.0104 0.0610 the medium and long term Evidence of increased viability and capacity of farmer organizations 0.7015 0.1201 0.2245 0.4135 Measures of improved natural resource management by farmers 0.6823 0.2100 0.1367 0.0815 Returns to land and labour invested in focal commodities 0.1793 0.7628 0.0984 0.0904 Measures of technical farm productivity (eg. crop yields) 0.0307 0.7211 0.1764 0.2393 Gross value of sales of smallholders from focal commodities 0.2239 0.5886 0.2104 0.3431 Evidence of improved access to technical or business advisory services 0.2106 0.1510 0.7631 0.0420 Quantities of technical assistance provided (for example of 0.1281 0.0977 0.7497 0.1299 people/groups trained, etc.) Measure of public sector administrative/technical capacities 0.1732 0.0145 0.7910 0.0228 Evidence of increased capacity of local providers of technical/business 0.2982 0.0302 0.7132 0.0112 services Number of farmer groups formed 0.0228 0.1334 0.2110 0.8351 Number of targeted farmers supplying the highervalue crop/livestock 0.1877 0.3655 0.0104 0.6866 product Number of farmers/groups certified for a particular standard 0.0512 0.0518 0.1009 0.7009 Distribution of participation based on gender 0.3629 0.2929 0.1376 0.6126 33 Annex Table 5. Results of principle components analysis of scaling scores for importance of factors in success (or otherwise) of interventions linking smallholders to highervalue supply chains: Factor PC1: PC2: PC3: PC4: PC5: PC6: Prevailing Project Effective Role of Farmer Ability to External Management Institutions Private Experience Adapt to Conditions Sector Local Conditions Prevailing local interest rates 0.8522 0.0248 0.1384 0.1336 0.1423 0.0187 Positive trends in pertinent commodity prices 0.6935 0.1464 0.1360 0.0083 0.3143 0.1722 National sociopolitical stability 0.6137 0.0572 0.1342 0.0345 0.1275 0.3597 Commercial expertise of project advisory service provider 0.0144 0.8368 0.0467 0.0470 0.0280 0.0128 Appropriate management of financial support to farmers', ` 0.1092 0.6766 0.0739 0.1971 0.0253 0.2459 Technical expertise of project advisory service provider 0.2770 0.6673 0.0059 0.0707 0.0399 0.2210 Appropriate intervention timing in the `lifecycle' of an industry 0.2446 0.4678 0.2220 0.3729 0.1201 0.2790 Welldefined property rights 0.2213 0.1220 0.8323 0.0812 0.0297 0.0504 Coherent ownership structure for farmer organizations 0.0526 0.3121 0.7513 0.1936 0.1988 0.0158 Strong enforcement institutions 0.1902 0.2301 0.5421 0.0711 0.0384 0.2142 Project interventions initiated by the private sector 0.1490 0.0559 0.0104 0.8610 0.0576 0.1662 Project interventions in collaboration with the private sector 0.0162 0.0624 0.2443 0.8080 0.0283 0.0487 Sharing of project intervention costs with benefiting firms 0.0056 0.1242 0.2167 0.7431 0.0895 0.1342 Targeting farmers with prior positive experience of producer organizations 0.0737 0.0345 0.1064 0.2082 0.8351 0.0067 Targeting farmers with prior experience producing any commodity for a 0.0815 0.0913 0.0280 0.2577 0.8305 0.0515 discerning, qualityfocused market Targeting farmers with prior experience in producing focal or similar 0.1342 0.0893 0.1452 0.1174 0.8134 0.1003 commodity Agroecological conditions in targeted producer areas 0.3056 0.0099 0.0534 0.1170 0.1240 0.7334 Security situation in targeted production areas 0.0337 0.0669 0.1751 0.3092 0.2898 0.6597 Flexible project design enabling modifications and redirections according to 0.0521 0.1667 0.1854 0.2762 0.2154 0.5303 local conditions 34