Peasant agency in global value chains : an examination of how small-scale oil palm farmers in Ghana exercise agency through peasant agriculture

(1)

Peasant Agency in Global Value

Chains

An examination of how small-scale oil palm

farmers in Ghana exercise agency through peasant

agriculture

Ruth Manley (Student ID: 11126558)

ruthemanley@gmail.com International Development Studies

University of Amsterdam 24th June 2016

Supervisor Local supervisor Second reader

Dr. Y.B.P. van Leynseele Dr. M.A.A Derkyi Dr. M.A.F. Ros-Tonen

University of Amsterdam, University of Energy and University of Amsterdam,

(2)

1

Abstract

Food sovereignty is an emerging concept in critical agrarian studies that is premised on ideas and practices related to small-scale family farming as a means of exercising agency and resisting the strong forces of globalisation. The food sovereignty movement rejects the pursuit of development through increased market integration and value chain collaborations. This thesis takes an actor-oriented approach, combining critical agrarian and value chain approaches, to analyse smallholders’ potential for agency in the face of increasing market integration in global value chains, concentrating on the oil palm sector in Ghana. The evidence is drawn from a mixed method research approach including surveys, observations, focus group discussions, and individual interviews with smallholders and oil palm buyers.

The analysis shows that oil palm market relations are not marked by a single set of conditions produced by global value chains. A number of value chains exist which present a plurality of terms of engagement to which smallholders can respond. Within this context, two main modes of farming are brought to the fore – entrepreneurial and peasant – whereby most smallholders are between the two with varying ‘degrees of peasantness.’ The findings show different trajectories among oil palm smallholders whereby contract farmers resemble the entrepreneurial mode of farming – characterised by relations of dependency and constraint – and independent farmers (those who are not part of a contracting scheme) mirror the peasant mode of farming – characterised by autonomy, resilience, independent production and selectively engage with the markets in a process wherein agency is manifested. Nuances were also found within the dichotomous groups as some contract farmers willingly give up their autonomy, and some independent farmers do not have the autonomy to choose their buyers due to their lack of access to assets and resources and subsequent exclusion from global value chains.

The thesis provides a strong claim for including market integration into the food sovereignty discourse as smallholders display agency through selective engagement in global and local value chains. This underlines the importance of global markets to many smallholders and highlights weaknesses in the critical agrarian hypothesis that smallholders exercise agency through distancing from global markets. Secondly, the thesis advocates that an actor-oriented ‘beyond the chain’ approach is necessary for better informed policies regarding truly inclusive value chain collaborations that incorporate principles of food sovereignty.

(3)

2

Acknowledgements

This thesis would not have materialised without the help of a coutless number of people I can only hope are aware of my infinite gratitude. I would like to thank some in particular. First, I wish to express my sincere appreciation to my supervisor, Dr. Yves van Leynseele, whose continuous advice and enthusiasm has guided me through the thesis writing process in the smoothest way possible. His timely, meticulous feedback and genuine interest is much appreciated. I would also like to extend my gratitude to Dr. Mirjam Ros-Tonen, not only for taking the time to be my second reader, but for making us feel like part of the team from the beginning. Secondly, I wish to thank all those involved in my field research in Ghana, your kindness and generousity will never be forgotten. My particular thanks goes to Dr. Mercy Derkyi, whose care for us during our time in the field was second only to a mother, and to Marta Ataa-Asantewaa, who helped ease many of my fieldwork concerns. I would also like to thank the staff of the MOFA who provided me with much needed practical support during my stay in Kade, and Mary, for being my Ghanaian sister. To my dear research assistants, Emmanuel Jesse Sarbeng and Sam Asiedu, thank you for your help, your hard work, your patience during my firey outbursts in the heat, and, most of all, your friendship. Thirdly, I wish to extend my deepest appreciation to all my friends. Thanks especially to Lars and Marie for an unforgettable adventure in the field, and to the rest of the IDS crew for keeping me sane during the intense months that preceeded it. Finally, thanks to my beloved family for their love and support, this year and always.

(4)

3

List of Acronyms and Abbreviations

BOPP – Benso Oil Palm Plantations Limited

COVE – Corporate Village Enterprises

CPO – Crude Palm Oil

FFA – Free Fatty Acid

FFB – Fresh Fruit Bunches

FSM – Food Sovereignty Movement

GOPDC – Ghana Oil Palm Development Company Limited

GVC – Global Value Chain

MASDAR – MASDAR International Consultants

MNC – Multinational Corporation

MOFA – Ministry of Food and Agriculture, Ghana

NPP – New Patriotic Party

NWO – Netherlands Organisation for Scientific Research (based in the Netherlands)

Obooma – Obooma Farms Limited

OPRI – Oil Palm Research Institute (based in Ghana)

PPP – Public-Private Partnership

PSI-Oil Palm – President’s Special Initiative for Oil Palm

Serindipalm – Serindipalm Company Limited

TOPP – Twifo Oil Palm Plantations Limited

UENR – University of Energy and Natural Resources (based in Ghana)

UvA – Universiteit van Amsterdam (based in the Netherlands)

VC – Value Chain

VCC – Value Chain Collaboration

(7)

6

List of Figures, Tables and Pictures

Figures

2.1 Interfaces and degrees of peasantness

2.2 Conceptual scheme

3.1 Research location

3.2 Visual model for mixed methods sequential explanatory design procedures

4.1 Restriction from resources in the community

5.1 Land tenure

5.2: Crop combinations

Tables

3.1 Overview of methods and respondents

Pictures

3.1 Focus group discussion with women farmers in Damang

4.1 Small scale oil palm processing machine and medium-scale oil palm processing plant

(8)

7

Chapter 1: Introduction

Agriculture remains predominantly a small-scale affair. A significant proportion of the world’s food is produced by farms of under two hectares managed by some of the world’s most vulnerable families (Vorley et al., 2012; IFAD 2013a). Efforts to reduce poverty have therefore been closely linked to small-scale farming. A growing tendency sees ‘market-based’ development whereby smallholders are linked to global markets through ‘inclusive business’ and value chain collaboration (VCC) models (Vorley et al., 2012). Connecting smallholders to global markets can present opportunities for increased productivity, income and innovation capacity (Swinnen et al., 2013; Burnett and Murphy, 2014). However, it also raises the question of how the right to self-determination and ability of local people to be autonomous in the food chain is impacted (Arthur, 2012). The food sovereignty movement (FSM) has emerged in response to such concerns, emphasising the need for resilience and autonomy among smallholders and rejecting the pursuit of development through increased market integration (Altieri, 2009).

Against the backdrop of food sovereignty, this thesis explores the potential for peasant agency in the face of increasing market integration in global value chains (GVCs) controlled by large multinational corporations (MNCs). Taking an actor-oriented approach, it looks ‘beyond the chain’ to include non-commodity production, social relations in the local setting, and alternative marketing of oil palm (Ros-Tonen et al., 2015: 542). Agency is used to understand farmers’ strategies and constraints vis-à-vis GVCs and the context within which they operate (Vorley et al., 2012). The oil palm sector in Ghana is the chosen case study as oil palm is grown both as a cash crop for sale in global markets and as a food crop for local consumption. Furthermore, the sector provides a number of upgrading opportunities to farmers in the form of contract farming schemes, certification programmes, and local palm oil processing. Contract farming schemes and certification programmes are common models of VCC, while local palm oil processing portrays an alternative market that can be linked to the FSM. Ghana’s oil palm sector therefore presents an interesting case study of how agency can be exercised in different markets. The research question addressed is how do smallholders

display agency through patterning their production activities and market relations with different value chains in the oil palm sector in Ghana?

It follows that the thesis adopts two theoretical strands: a value chain perspective and a critical agrarian approach (from which the FSM emerged). Inclusive business and value chain models often treat smallholders as beneficiaries of external initiatives (Vorley et al., 2012). The FSM, on the other hand, advocates for the view of smallholders as active agents. The thesis thus focuses on agency as a possible means of reconciling inclusive value chain approaches with a food sovereignty perspective. This is relevant to the debate as it explores how smallholders can have an active, as opposed to passive, role in VCCs. It therefore helps overcome the ‘top down’ focus of GVC analysis by focusing on producers and bridge the weakness of the latter’s sole focus on vertical chain relations by introducing a horizontal ‘beyond the chain’ contextual understanding into the analysis. Simultaneously, it overcomes the widely held critique of peasant agency theories regarding their lack of consideration for the importance of

(9)

8

markets. Moreover, the thesis contains societal relevance as an actor-oriented perspective can inform more appropriate and inclusive policies regarding smallholders’ market integration through VCCs.

The research is based on 60 surveys, 26 interviews, 2 focus groups and observations conducted in the Eastern region of Ghana over a period of 4 weeks. The field research was carried out within a broader framework of a four-year action research project titled Inclusive

business – smallholder partnerships for greater food sovereignty and sustainable landscapes in Ghana and South Africa funded by the Science for Global Development (WOTRO)

division of the Netherlands Organisation for Scientific Research (NWO).

This chapter has aimed to provide the rationale behind the research. Chapter two outlines the two main theoretical strands of the research: value chain analysis and peasant agency as defined by critical agrarian studies. After an introduction of the theories of peasant agency, focus is specifically placed on the ways in which smallholders pattern their production activities and market relations with different value chains, and how this enhances their agency. Chapter three predominantly discusses the methodology and research design of the research, including ethical considerations and limitations of the research. The development of the oil palm sector is outlined in chapter four, using a value chain analysis to map out current market relations. An examination of horizontal relations is also provided. Chapter five then considers peasant agency by examining the capabilities, constraints and degrees of peasantness of respondents. This enables the categorisation of farmers according to how they farm and the identification of farmers who display agency as according to peasant agency theories. This group of farmers is discussed in chapter six, which analyses how they organise their production and marketing so as to enhance agency. Chapter seven concludes the thesis, providing a summary of the findings, a theoretical reflection and suggestions for policy and further research.

(10)

9

Chapter 2: Theoretical Framework

This thesis adopts a critical agrarian approach combined with a value chain analysis in examining the agency of smallholder farmers. It follows that this section contains two sub-sections; agency and value chain analysis. A third sub-section discussed strategies for patterning production and marketing.

2.1 Value Chain Analysis

Value chain analysis is increasingly used in the literature as an analytical tool exploring the international expansion and globalisation of production and exchange (Gereffi, 2014; Kaplinsky, 2000). It provides a framework for mapping the activities along a chain, the distribution of power and rewards, and the location of entry barriers. Ghana has experienced the expansion of scale export-oriented production of oil palm with the result that large-scale private companies currently play a large role in the sector. This is typical of developing countries as they often follow strategies of exporting products to developed markets in the belief that their industries will upgrade to global standards (Navas-Alemán, 2011). Such practise has led to the emergence of a literature that studies global value chains (GVC), within which suppliers in developing countries are tightly governed by global buyers (ibid). The issues of upgrading and governance are central to GVC analysis, where governance is the top down view focusing predominantly on lead firms and the organisation of global industries, while upgrading constitutes the bottom up perspective focusing on the strategies used by economic stakeholders to maintain or improve their position in the value chain (Gereffi, 2014). This underlines the importance of a value chain (VC) analysis to this study as issues of governance and upgrading provide insight into the structured context within which smallholder agency is either enhanced or constrained. As the thesis takes an actor-oriented approach focusing on farmer perspectives, the research focuses predominantly on forms of upgrading in the research area and the effects of governance on local actors as opposed to power distribution within the whole VC.

The interconnected aspects of governance and upgrading in VCs can affect smallholders’ agency as they determine inclusion, exclusion or adverse inclusion in value chains (Hickey and du Toit, 2007). Governance structures are often created by lead firms, which are defined as the firms that control chain activities. They determine the terms of chain membership, incorporating or excluding certain actors and allocating the value-added activities. For smallholders, this can lead to either marginalisation and exclusion or upgrading and participation (Ponte and Gibbon, 2005). While upgrading and participation are often seen as positive outcomes, highly insecure and exploitative employment can mean smallholders have been adversely included in a chain and are not necessarily better off as a result (Barrientos et al., 2011). As explained below, agency is defined in the thesis (partially) as autonomy over marketing of crops, making this an important factor for analysis as inclusion may not be advantageous and exclusion may therefore be by choice. For example, smallholders may decide to produce for less lucrative but less risky local markets if they feel they lack leverage in larger VCs (Bolwig et al., 2010). Thus, households can be involved in hybrid arrangements

(11)

10

of inclusion in certain chains and exclusion from others (du Toit, 2009). They may participate in multiple VCs and may be only partially integrated in the chains (Ploeg, 2014) – an important consideration for this thesis which analyses the different trajectories of farmers regarding their participation in different markets.

The latter point highlights the need to move beyond a GVC analysis to include an analysis of domestic and regional VCs. Due to the emphasis on GVCs, the type of chain governance that exists in smaller VCs and the effect it has on upgrading has been largely neglected in the literature (Navas-Alemán, 2011). However, following the food sovereignty discourse, this thesis proposes that these chains can offer different possibilities for exercising agency as they are more loosely governed and offer separate opportunities for upgrading. Furthermore, the thesis emphasises that many farmers are simultaneously involved in several VCs and so are exposed to different governance structures and upgrading patterns (ibid). As will be described further on, the options created by following a multi-chain strategy are important considerations for agency as they can reduce dependency on markets and increase the resilience of the farm. A VC analysis thus contributes to this research as it allows for the analysis of how and why households are integrated in markets in different ways. Analytically, this requires determining the effect of upgrading and governance structures on smallholders and their ability and/or desire to participate in a certain VC through an examination of the terms of engagement within that chain.

Accordingly, it is useful to distinguish between ‘tight’ and ‘loose’ VCs regarding governance structures and upgrading opportunities. Tight VCs operate through clearly established relationships and a single channel, usually involving contract agreements (Cuevas and Pagura, 2016). Such agreements between farmers and processing companies involve the production and supply of agricultural products by farmers and the supply of inputs and technical assistance by companies (Eaton and Shepard, 2001). They result in the commoditisation of farmers and their subsequent dependency on the buying company. Such agreements can increase farmers’ exposure to risks as outputs must meet specific standards as specified in the agreement. However, they also present opportunities for farmers in the form of access to improved seeds, inputs, finance, and secured markets (Christen and Anderson, 2013). The export-oriented VCs governed by multinational corporations (MNCs) operating on contract schemes are a prime example of tight VCs in Ghana. Contract farming is therefore an important background to this research. By contrast, loose VCs are usually locally embedded, self-governed chains. They offer farmers a variety of marketing options and, therefore, opportunities for competition. However, farmers often have less opportunity to access credit, inputs, extension, and sector knowledge flow between participants in the chain (Cuevas and Pagura, 2016). Limited access to such resource can expose farmers to risks. The alternative market for oil palm generated by the presence of small-scale processing plants is an example of a loose VC. It is important to note that these typologies are flexible and smallholders can be engaged with both to varying extents which change over time (Cousins, 2014). However, making the distinction can be helpful to analysing how and why smallholders interact with different markets and if and how they can pattern market relations so as to find an

(12)

11

arrangement best suited to them. The distinction is therefore used in this research based on the terms of participation and terms of engagement of the different oil palm markets.

Analysing different VCs that operate in the same locality calls for a ‘bottom-up’ producer perspective. Conventional VC analysis, however, focuses predominantly on the lead firm (Laven, 2010). Furthermore, it is only concerned with governance arrangements and power relations within the chain. Critical observers call for the consideration of the context in which VCs are embedded (Bolwig et al., 2010; Ros-Tonen et al., 2015). The vertical linkages in VC analysis must therefore be incorporated with ‘horizontal’ interactions and effects ‘beyond the chain’ (Ros-Tonen at al., 2015: 527). Such horizontal embeddedness is relevant to understanding households’ engagement with VCs as it broadens the range of issues considered to include power relations and inequality in the local setting as well as power relations within the chain (Bolwig et al., 2010). While an in-depth study of such is largely outside the scope of this research, the research identifies gender relations and marginalisation as two particular points of interest. Several studies have found that labour in oil palm harvesting is clearly divided by gender (Osei-Amponsah, 2013; Adjei-Nsiah et al., 2012; Koczberski, 2007). Women are predominantly engaged with the processing and sale of produce (Osei-Amponsah, 2013; Adjei-Nsiah et al., 2012), while men carry out the bulk of harvesting activities (Koczberski, 2007; Koledoye and Deji, 2014; White and White, 2012). The division of labour in turn effects gender relations and is important to consider in this study as men and women will therefore differ in their abilities to exercise agency. Marginalisation is also relevant to VC analysis as people may be deeply incorporated in a particular VC, but excluded in local society. It is considered in this research following studies that have shown contract farming can lead to socioeconomic marginalisation as it is generally the better-off smallholders who are recruited to them, which increases rural equalities (Key and Runsten, 1999; De Schutter, 2011).

Focusing on these horizontal aspects is important for this study as it takes an actor-oriented approach and looks at the effects of different VCs operating in the same location. Indeed, looking ‘beyond the chain’ is central to understanding peasant patterning processes, as described below. It allows for the analysis of the role of external networks and resources in determining participation in different VCs. Agency can thus be understood in terms of why farmers differ in their engagement (or their capacity or desire to engage) in VCs and how this affects processes of vertical VC inclusion (Ros-Tonen et al., 2015).

2.2 Peasant Agency

Before going into the debate on peasant agency, it is important to determine what is meant by peasants and how, in this study, they are defined vis-à-vis small-scale farmers (or smallholders). The concept of food sovereignty is emerging in critical agrarian studies and it is behind this backdrop that the notion of agency is developed. Food sovereignty can be defined as the “right of people to produce, distribute and consume healthy food in and near

their territory in an ecologically sustainable manner” (Altieri and Toledo, 2011: 588). Its

roots lie in the peasant response to the globalisation of agrarian processes (McMichael, 2014). Indeed, the concept is premised on ideas and practices related to small-scale family farming as

(13)

12

a means of exercising agency and resisting the strong forces of globalisation (Edelman, 2014). Such ideas and practices are often referred to as peasant agriculture in the food sovereignty discourse. Accordingly, this study defines peasants in relation to smallholders where smallholders are “farmers who produce goods and services for both markets and subsistence,

based mainly on family labour and limited access to land” (Ros-Tonen et al., 2015: 523).

It follows that this thesis adopts an actor-oriented approach which conceptualises smallholder agency in terms of strategies used by farmers to strengthen their resilience and food sovereignty. Subsequently, agency is defined as the “knowledgeability, capability and social

embeddedness associated with acts of doing (and reflecting) that impact upon or shape other persons’ actions and interpretations, and recursively, of course, one’s own” (Long, 2001: 240

- 241). Smallholders are therefore viewed as potential reflective agents whose actions are carried out in relation to other people or things. This will be shown below in relation to how market relations and production activities are patterned through the aforementioned strategies. Importantly, agency is a temporally embedded process of action based on the past but also oriented towards the future and the present (Emirbayer and Mische, 1998; Long, 2008). Considering the assets and resources of people is essential to understanding the ability of different actors to exercise agency. The research distinguishes between assets and resources by recognising that assets can be considered resources, though not all resources are assets. They can also include other things of use or help to actors that they have knowledge of and access to but do not own (Dijk, 2011). Assets and resources can be created through interaction with other individual and collective actors, including social networks and family relations, and the mobilisation of resources to build strategies. They describe the capability to take action and form the basis of power of an agent (Long, 2001). Accordingly, it is important to also consider ‘non-agency’ whereby there is a lack of access to assets and resources and therefore an incapability to take action (Ploeg cited in Long, 2008: 80).

A growing body of literature outlines the re-emergence of peasant agriculture as a way through which smallholders exercise agency. The broader debate focuses on how globally-managed circuits of food are displacing production factors of smallholders through ‘accumulation by dispossession’ (McMichael, 2006: 407). This is caused by commoditisation relations that incorporate farmers into global industrial-retailing circuits through, for example, contract farming schemes. In these circuits, resources are combined into socio-technical networks which externalise factors of production and knowledge (Ploeg, 2008). They are negotiated through commodity relations with the result that smallholders, while still able to generate an income by producing for the market, are stuck in relations of dependency and constraint. The reconsolidation of the peasantry is arguably a direct response to the crisis engendered by dispossession and commodification brought about by the corporate food regime and represents the possibility of a peasant modernism based on a politics of ecology and food sovereignty (McMichael, 2006). Central, then, to the peasantry is what Van der Ploeg describes as “the struggle for autonomy that takes place in a context characterized by

dependency relations, marginalization and deprivation” (Ploeg, 2008: 23). Autonomy is a

central aspect to peasant agency as it defines the room for manoeuvre within which agency is manifested (Long, 1985).

(14)

13

The re-emergence and strengthening of peasant agriculture, or what is also termed ‘repeasantisation,’ is one of three development trajectories that categorise global agriculture (Ploeg, 2008: 1). The respective processes of industrialisation and deactivation constitute the remaining trajectories, where the former signifies large-scale export-oriented expansion of agricultural production while the latter involves the reduction or containment of levels of agricultural production (ibid). Such trajectories pertain to certain segmentations of agriculture, which can be conceptualised as three modes of farming; peasant farming, entrepreneurial farming, and capitalist (or corporate) farming. Each mode can be defined as a “distinctive and

valid way of farming that is shared by a large group of farmers” (Ploeg, 2010b: 3). In

capitalist agriculture, all resources, including the labour force, are commodified and the production process is based on capital-labour relations and profit maximisation, even if this leads to a reduction of the total value-added (Ploeg, 2014). Entrepreneurial farming is mainly grounded on far-reaching commoditisation of the resources. It is built on financial and industrial capital, and ongoing expansion is a necessity. Production is highly specialised and totally oriented towards markets, and autonomy, self-sufficiency and the demographic cycle contained within the farming family is not considered relevant. The takeover of the resources of others is as important as the production of value-added with the available resources. Finally, peasant farming is grounded on low levels of commoditisation of the main resources, multi-functional farms, and distancing from the main markets (Ploeg, 2008; Ploeg, 2014).

It must be recognised that the aforementioned modes of farming are ideal-typical and in reality there are no definitive lines of demarcation between modes of farming. As visualised in the figure 2.1, the ‘interfaces’ between the different modes represent overlap and ambiguities where boundaries are crossed both backwards and forwards (Ploeg, 2008: 3). Within such interfaces, degrees of peasantness emerge as analytical distinctions which signify the different development trajectories outlined above and recognise the different ways of being integrated into markets and trade networks (Ros-Tonen et al., 2015). On the one hand, farmers may follow a repeasantisation pathway whereby they pattern their production in ways that give them the agency to re-work their market relations on terms best suited to their own food sovereignty (ibid). It is this cyclical process of patterning in peasant agriculture that the research explores, and will be described in detail in the proceeding section. On the other hand, farmers may follow a de-peasantisation pathway whereby peasant practises are eroded, leading to increased dependency on and adverse incorporation into external markets (Ploeg, 2008; Ros-Tonen et al, 2015).

(15)

14

Figure 2.1: Interfaces and degrees of peasantness

Source: Ploeg, 2008: 37

The interconnections between the three modes of farming and greater society are patterned in different ways. One dominant pattern is the highly centralised, often global operations of large food processing and trading companies. Another is the construction and re-production of short and decentralised circuits that link farming and regional society (Ploeg, 2008). While Van der Ploeg speaks of ‘Empire’ and ‘short food circuits’, the notions of tight and loose value chains can also be introduced here. The peasant mode of farming is often, though not exclusively, associated with loose value chains. Capitalist and entrepreneurial farming are, on the other hand, associated with tight value chains. This is an important consideration when analysing opportunities for agency within different value chains.

2.3 Strategies for Patterning Production and Market Relations

The aforementioned peasant mode of farming derives from the ‘peasant condition,’ which grounds the peasantry in its contemporary context while at the same time acknowledging the agency inherent to it (Ploeg, 2008: 23). Such agency emerges in the form of multiple and internally coherent strategies used by peasants to pattern their production activities and market relations in order to become resilient against a hostile structuring environment. More specifically, agency is manifested through the creation and strengthening of a self-managed resource base which allows for co-production between man and nature. This enables farmers

(16)

15

to re-work market relations and feed back into the resource base enhancing autonomy and, thus, reducing dependency (ibid). These strategies may be strengthened through engaging in other off-farm activities, through patterns of co-operation among peasants, and/or through the creation of nested markets. Due to time and space constraints, each element in the process is not discussed in detail. Several strategies are elaborated on and reference is made to how they interact.

Co-production and strengthening of resource base

A defining feature of the peasantry is the process of co-production between human beings and living nature whereby social and material resources are constantly transformed, and are continuously changing agriculture (Ploeg, 2010). The natural resource base is both a condition and an outcome of co-production. Indeed, the construction and maintenance of a self-controlled resource base is crucial to co-production. Peasants strive to create and strengthen an endogenous, resilient and multiple-use resource base through the internalisation of assets in the production unit (Schneider and Niederle, 2010). The process can be further expressed through the conversion of produce into more refined products. In this view, land is considered as ecological capital. A well-cared for ecological capital enables patterns of growth that are independent of the main markets for factors of production and non-factor inputs (Ploeg, 2010). Co-production thus constitutes the intertwining processes of production and reproduction (Ploeg, 2008). It includes the reproduction of the household, a point which is specifically relevant to this thesis given the gendered nature of the division of labour in the oil palm sector. Patterns of production and (social/material) reproduction feed into one other allowing for a more resilient farm and the subsequent emancipation of the peasantry. This generates progress and stimulates endogenous forms of development. It also shapes the social sphere as it integrates the centrality of family farming, craftsmanship, and the artisan nature of production processes (ibid). It is for these reasons that co-production is increasingly a choice of farmers, reflecting agency and preference, which both uses and enlarges their autonomy.

Diversification of crops

One specific strategy used by peasants in developing and maintaining a self-controlled resource base is the diversification of crops for use in commodity and non-commodity circuits. Diversification is therefore central to patterning of both production and market relations. Commodity relations constitute engagement with VCs, and the level of dependency is determined on whether the latter are loosely or tightly governed. Non-commodity relations, on the other hand, operate at the horizontal level and involve production that is not sold. Self-provisioning thus allows for the operation of the farm unit as a whole as opposed to merely the family consumption of food (Ploeg, 2008). It involves reducing dependency on external resources while simultaneously enlarging and improving the stock of internal resources, including ecological capital. This can result in the reduction of monetary costs while overall levels of production are maintained or even slightly improved. Unlike specialisation, which increases dependency relations, diversification increases farmers’ autonomy (Ploeg, 2010). Temporal and variable combinations of production in commodity and non-commodity circuits

(17)

16

are crucial in understanding smallholders’ inclusion in and (often) deliberate exclusion from value-chain relations (Ros-Tonen at al., 2015).

Pluriacivity

Another strategy employed by peasant households to strengthen their resource base is the engagement in pluriactivity (Ploeg, 2014). Pluriactivity occurs when agricultural activities are combined within and outside the production unit, or when agricultural activities are carried out within the production unit and non-agricultural activities are carried out in other sectors (Schneider and Niederle, 2010). For example, households may employ partial engagement in the labour force in order to diversify their incomes. Unlike other forms of market integration, pluriactivity does not lead to a reduction in autonomy. Indeed it is part of the circuit of reproduction as it can allow households to invest in their farms and thus avoid reliance on banking circuits and moneylenders (Ploeg, 2008). For some it is a means of getting the better of two worlds, for others it is a way to get started, while for others again it enables them to continue farming in a relatively autonomous way (Ploeg, 2010).

Patterns of co-operation

Forms of co-operation are often needed to face a hostile environment. Harsh ecological environments are better dealt with through, for instance, peasant-managed irrigation systems or socially regulated exchange patterns (Ploeg, 2008). Similarly, adverse institutional environments require co-operation between peasants in order to defend both individual and collective interests. Co-operation among peasants is a strategy for constructing autonomy at a higher level of aggregation and room for manoeuvre within a hostile environment.

Creation of nested markets

Peasants may also resort to the creation of nested markets as a response to their subordination to the world market and the agribusiness MNCs working within it (Hebinck et al., 2015). Nested markets consist of market segments that are embedded in wider commodity markets but have different dynamics, prices, redistribution of value added, and relations. They are constructed through cooperative relationships among local actors and act as an alternative to the global markets controlled by MNCs who aim to ‘squeeze out’ as much value as possible (Ploeg, 2010: 18). Some defining characteristics which distinguish them from conventional markets include the following: they are seen as specific places where transactions take place according to specific rules, rather than as systems; their construction is induced by changes on the sphere of production; and they are based on short food circuits. Furthermore, they go beyond the conventionalisation thesis which argues that changes in production and marketing will be appropriated by larger industries (Hebinck et al., 2015). Indeed, nested markets provide an alternative to these industries and potential for achieving food sovereignty (Schneider and Niederle, 2010). An important factor is that the construction of these new markets helps create relative autonomy for smallholders (Ploeg, 2010). This is due in part to the re-integration of skills and competences, such as the arts of processing and marketing, which have been externalised to other value chain actors. This thesis posits that the existence

(18)

17

of nested markets is important for selectively engaging with markets as they provide farmers options of following a multi-chain strategy, build resilience through less dependency on GVCs, and help balance commodity and non-commodity relations

Synthesis

In sum, the above strategies represent a cyclical process whereby peasants’ autonomy is enhanced and dependency on external markets reduced. Cycles of production use resources that were produced and reproduced during previous cycles or through socially regulated exchange. This does not only refer to the physical production of resources on the farm, as in the case of self-provisioning, but also the conversion of one’s own resources (such as savings) into the required ones, as in the case of pluriactivity. A sharp contrast to market-dependent reproduction is presented in this pattern as resources do not enter the production process as commodities. They are created by farmers who therefore have the autonomy to do with them as they please. Such patterning of production activities strengthens the resilience and reduces the dependency of the farm on global markets. This autonomy allows peasants to retire or expand their relations with, and thus avoid entrapment in that market, a particular market at any given time. It follows that peasants can work their market relations according to their own individual preferences vis-à-vis the terms of engagement in the different VCs. As outlined in section 2.1, this often involves partial inclusion in both tight (export-oriented) and loose (nested) VCs. It is through this process of coherent patterning on the production and marketing level that peasant agency is displayed.

2.4 Conceptual Scheme

The conceptual scheme visualises how the main concepts of the theoretical framework are connected. It shows how the degrees of peasantness of farmers – as based on their assets and resources, the struggle for autonomy, and objective of production (which can also be affected by horizontal relations) – determines whether they follow a process of de-peasantisation leading to an entrepreneurial mode of farming or a process of repeasantisation wherein the peasant condition is strengthened. The latter, it is hypothesised, is achieved through strategies of co-production, diversification, and patterns of co-operation, which enable farmers to selectively engage with oil palm VCs based on their opportunities and risks. The thesis conceptualises this process as peasant agency. Theories of peasant agency hypothesise that the entrepreneurial mode of farming is associated with tight VCs, which are considered to constrain agency. The thesis tests both hypotheses.

(19)

18

Figure 2.2: Conceptual scheme

Source: Original material for this thesis 2.5 Conclusion

The theoretical framework has adopted theories of peasant agency as derived from the food sovereignty discourse along with a value chain perspective. Both theoretical strands are identified as providing a complementary insight into how farmers organise their production and marketing activities vis-à-vis the structuring environment present in the research area. Combining the theories aims to identify which farmers exercise agency and how, and which farmers’ agency is constrained and how.

(20)

19

Chapter 3: Research Methodology and Methods

The purpose of this chapter is to describe the steps taken in addressing the main research question specified in chapter one. It begins by introducing the sub-questions that guide the research. An explanation of how the major concepts presented in the theory were operationalised follows. A brief justification of the research location is also provided. The chapter then turns to a detailed description of the research methodology and methods. Considerations regarding ethics and limitations bring the chapter to a close.

3.1 Research Questions

The purpose of this research is to answer the following question:

How do smallholders display agency through patterning their production activities and market relations with different value chains in the oil palm sector in Ghana?

In order to do so, four sub-questions are asked. Sub-questions one and two are based on the VC literature while sub-questions three and four are based on theories of peasant agency. Accordingly, the sub-questions are formulated as follows:

1. What are the characteristics of the different value chains in the oil palm sector regarding the terms of inclusion, risks and opportunities?

2. What are the horizontal relations in the local community regarding gender relations and marginalisation?

3. How do smallholders differ in their peasant agency concerning capabilities, constraints and degrees of peasantness?

4. How does peasant agency determine household strategies at both the production and marketing level?

3.2 Operationalisation of Major Concepts

The central concepts of the research questions are operationalised in order to transform the theory outlined in the previous chapter into a measurable form. As derived from the theoretical framework, the theories operationalised are structuring environment, peasant agency, and patterning strategies. They are broken down into concepts with several dimensions, each of which is in turn divided into a number of variables. This section briefly outlines why each was chosen. A full operationalisation table is presented in the appendix which includes the indicators of all variables.

Structuring environment

Structuring environment is split into two sub-concepts; vertical value chain (VC) inclusion and horizontal relations. Following global value chain (GVC) analysis, the former is broken

(21)

20

down into the two central dimensions of VC analysis: governance and upgrading. Chosen variables for governance are entry barriers and contract arrangements. They aim to indicate whether a given VC is tight or loose (Cousins 2014). The chosen variables for upgrading are

risks of inclusion and opportunities. These variables can give an indication of why farmers are

engaging in certain chains (or why they are not), and whether adverse incorporation is occurring. Horizontal relations analysis is broken down into the dimensions marginalisation (with variables contract farming and other groups) and gender relations (with variables

gender roles and decision-making). Peasant agency

Given that this research focuses on peasant agriculture as a specific analytical category, and patterning strategies derive from theories of peasant agency, peasant agency is also operationalised. Following the framework provided by Cousins (2014), it is useful to distinguish between smallholders degrees of peasantness in order to establish an analytical category of the mode of farming they mirror (peasant, entrepreneurial, or capitalist). The factors outlined by Cousins are hence adopted as variables: objective of production;

proportion of output marketed; contribution to household income; organisation of labour; mechanisation; capital intensity; and access to finance. The theory also outlines access to

assets and resources as important factors in peasant agency. Resources include the variables

material resources, use and ownership of resources, local infrastructure, social resources and knowledge. Variables for assets are land tenure, physical assets and financial assets.

Patterning strategies

The theoretical framework describes five patterning strategies commonly used by peasants to develop agency in the face of structural constraints; co-production, diversification, pluriactivity, patterns of co-operation, and the creation of nested markets. It follows that they are operationalised as the dimensions of this concept. Co-production is achieved through strengthening the resource base. This is represented by the variables access to resources and

strategies for using resources. Diversification includes the variables crop choice, produce marketed and produce used for household consumption in order to determine commodity and

non-commodity circuits. Pluriactivity includes the variable multiple jobs to indicate off-farm diversification of livelihood. Patterns of co-operation are operationalised into the variables

collective action and room for manoeuvre. Finally, the variables location, produce, rules and values and social contact are the variables for the creation of nested markets dimension as

they are seen as the principle characteristics of these markets that distinguish them from others.

3.3 Research Location

The research was carried out in the Kwaebibirem district of the Eastern Region of Ghana. This district was chosen as it hosts a number of different production systems (conventional and organic) and scales (large, medium and small) of processing (Osei-Amponsah et al., 2012). Farmers involved in contract schemes with different companies and those who operate

(22)

21

independently could therefore be included in the research. With this in mind, three specific villages within the district were selected. Damang hosts Obooma Farms Ltd., a medium-scale conventional oil palm processing company. Asoum is where the medium-scale organic and Fairtrade certified company Serindipalm Ltd. operates. Finally, within the environs of Kwae, the Ghanaian Oil Palm Development Project (GOPDC) Ltd. can be found – one of Ghana’s largest plantations and oil palm processing plants. All three villages host a number of small-scale and artisanal processing mills, making them perfect locations for this research.

Figure 3.1: Research location

Source: Domenico-de-ga 2016/Google 2016 3.4 Research Methodology

The epistemology of this thesis is pragmatism. Research was collected based on practicality in answering the research question and methodological choices were guided by a pragmatic research philosophy (Creswell and Plano-Clark, 2011). Building on this positioning, the thesis adopted a mixed methods approach whereby both qualitative and quantitative data collection methods were used. Using both methods together helped compensate for the limitations of the respective methods and increases the internal validity of results by triangulating the data (Bryman, 2006). It was especially useful to this research to use both quantitative and qualitative methods as the former allows for an examination of the relationship between agency and the structural context while the latter allows for the explanation of why certain farmers display more agency than others (by examining the strategies they use to do so) (Ivankova et al. 2006).

The research thus adopted an explanatory sequential research design whereby findings from the initial stages of research were used to inform later stages (Creswell and Plano-Clark, 2011). Accordingly, quantitative data was first collected and analysed. The results were then used to inform the proceeding qualitative data collection stage. Figure 3.2 visualises each step involved in the research design. The research sought a causal explanation of the relationship

(23)

22

between various degrees of peasantness and both VC and horizontal relations. The qualitative phase of the research then aimed to explain the survey results by determining what patterning strategies farmers use to enhance their agency (through repeasantisation) in relation to different VCs (Ivankova et al., 2006). Case studies were used in this phase, wherein certain households were chosen as examples of farmers who are following a pathway of either repeasantisation or de-peasantisation. Case studies are useful to this research as they highlight aspects of the theory without a representative population sample (Yin, 2009). While both quantitative and qualitative methods are important to the research, priority was given to the qualitative methods as the focus of this research is on uncovering patterns and processes which can lead to farmer agency. Moreover, qualitative methods are crucial for emphasising context (Bryman, 2006).

Figure 3.2: Visual model for mixed methods sequential explanatory design procedures of the research

Source: adapted from Ivankova et al., 2006 3.5 Methods

3.5.1 Data Collection

Data collection was conducted in the aforementioned villages over the course of four weeks, beginning on the 11th of February 2016. Table 1 provides an overview of the methods used to gather the data. The research began with a survey of 60 smallholders which was conducted with the aid of three survey enumerators. The data gathered from the survey was

(24)

23

predominantly quantitative and included ordinal, nominal and scale variables. Questions focused on respondents’ assets and resources, their degrees of peasantness, and their perspectives on VC and horizontal relations in the local setting. The survey was therefore aimed at answering the research sub-questions 1, 2 and 3. It provided a broader context of the research subject and allowed tests for relationships between variables. Furthermore, it explored differences between contract farmers and independent farmers. While an even spread between the two categories was originally planned, difficulties arose for local gatekeepers in determining which farmers were contract farmers and which were not. Ultimately, 23 contract farmers and 37 independent farmers were surveyed.

The survey results were inputted and analysed during the fieldwork (see below for detail on data analysis). This enabled the identification of farmers who displayed features of the peasant condition. 10 of such farmers were selected as case studies in the qualitative data collection phase. One farmer was unavailable on the day of interviews, however, and her son was interviewed in her place. As significant and interesting contrasts were found between independent farmers and contract farmers, 10 of the latter were also selected as case studies. This enabled an exploration of why such differences exist between the two groups. Indeed, the qualitative data aimed at taking an exploratory approach to discovering the patterning of production and marketing by smallholders. It was also used to explain results from the survey in terms of horizontal relations and vertical VC relations and how they relate to degrees of peasantness and access to assets and resources (Ivankova et al. 2006). Semi-structured interviews were conducted with the selected farmers, focusing on the reasons behind different market engagement and the strategies carried out at the production level which helps the farmer pattern their market relations in this way. They were thus aimed at answering sub-question four. The interviews followed an interview guide, which was developed based on the results of the survey (see appendix 3). The guide helped ensure the comparability between independent and contract farmers and retain focus on the central concepts (Bryman, 2008). One interview was conducted by the researcher in English, while the remainder were carried out in the local language, Twi, by two research assistants. Interviews were recorded and later translated and transcribed in English by the same assistants.

Interviews were also held with representatives of the main buyers of oil palm in the research area. Included were the three companies located in the research villages, namely GOPDC, Obooma Farms Ltd. and Serindipalm Company Ltd. These interviews were less structured than farmer interviews, allowing an in-depth exploration of specific issues of each company. For example, the interview with Serindipalm Company Ltd. focused on certification programmes, while the interview with GOPDC Ltd. focused on contract farming schemes. The interviews were conducted, and later transcribed, in English by the researcher. Additionally, a full-time small-scale processor and an independent trader (middleman) were respectively interviewed. While they examined separate issues to the company interviews, some similar themes were explored, most notably their relationship with the farmers. The research assistants carried out these interviews in Twi and later translated and transcribed them in English. Finally, an interview was held with a representative from the Oil Palm Research Institute (OPRI) in order to gain insight into the historical development of the oil

(25)

24

palm sector in the district. It was conducted in English by the researcher and recorded through note-taking.

In order to gain a deeper insight into horizontal relations (sub-question 2), and for triangulation purposes, focus group discussions (FGDs) were held. Such triangulation measures improve internal validity by checking the accuracy of the findings (Creswell, 2009). Participatory rural appraisal methods were used to determine the constraints of farmers in the research area (sub-question 3). Two FGDs were held, one with 10 respondents from Darmang and Kwae and the other with 6 respondents from Asoum. Most participants had been previously interviewed. However, in order to have a gender balance, several farmers who hadn’t been interviewed, but had been involved in the survey, were invited to join. The first half of each FGD included all participants together. They were then divided into two groups according to gender for the second half. Issues regarding gender relations and marginalisation could be more accurately explored in this way. The research assistants conducted the focus groups in Twi (see picture 3.1). They were recorded and later transcribed in English.

Table 3.1: Overview of methods and respondents

Field notes were taken throughout the entirety of the research based on observations by both the researcher and the research assistants. The aim of such observations was to obtain an in-depth insight into the inner workings of the oil palm sector, and to discover issues and strategies that farmers may not have considered relevant in interviews (Mack et al., 2005). Furthermore, they allowed the researcher to become intimately familiar with how households are vertically integrated into VCs and the horizontal relations within which this takes place.

(26)

25

Picture 3.1: Focus group discussion with women farmers in Damang

3.5.2 Unit of Analysis

The unit of analysis in this research is the ability of farmers to exercise agency in relation to different oil palm value chains.

3.5.3 Units of Observation

The primary units of observation in this research are peasant households as the research aims to analyse how the farm unit operates as a whole. This is crucial to the patterning of production and marketing strategies as it recognises that households often share resources and include a range of different activities carried out by different members (through pluriactivity). Accordingly, this analysis adopts a gender-sensitive perspective. Difference in the ability to exercise agency between men and women is acknowledged. Secondary units of observation in the research are the buyers of oil palm. This enables more accurate conclusions to be drawn regarding the relationship between farmers and the buyers.

3.5.4 Sampling Method

Purposive sampling was employed when conducting the survey as respondents were selected based on certain different criteria (Bryman, 2012). First, and most importantly, respondents had to produce oil palm. Thus, when the thesis speaks of ‘oil palm farmers’ it refers simply to farmers who grow the crop. Criteria regarding quantity or proportion of oil palm grown were

(27)

26

not requested. Second, farmers were selected based on their status regarding contracting. As previously mentioned, both contract farmers and independent farmers were selected. Respondents were found with the help of local gatekeepers. One assembly man1 in each village, each of whom had a list of farmers working in the area, recruited farmers according to the requested criteria. Thus, convenience sampling was used (Bryman, 2012). Purposive sampling was also used in the qualitative part of the research as respondents were selected from the survey participants based on specific requirements. Here, farmers that display signs of the peasant condition are the chosen sample as it is central to the theories of peasant agency. Moreover, farmers showing signs of compulsion due to their contract were selected for comparison. Within each group, diverse case selection was used (in terms of age, gender, etc.).

3.5.5 Data Analysis

The survey provided quantitative information that presents statistically supported conclusions within the research context. It was analysed using IBM SPSS Statistics 23 for descriptive statistics, which gave specific demographic information and an indication of the different relations farmers have to the markets and differences within the degrees of peasantness. In addition, SPSS was used to run correlation tests in order to establish whether there is a relationship between some of the ratio variables measured. Examples of such are the relationship between contract farming and dependency, the relationship between contract farming and autonomy, the relationship between land and autonomy, etc. In order to conduct such tests, perceptions on autonomy and dependency were measured on a 10-point scale and contract farming was converted into a dummy variable. This analysis was carried out in the field in order to inform and provide the sampling for the qualitative research.

Qualitative data was analysed with the aid of Atlas-ti software, which evaluates common trends in data. Interview transcripts were predominantly analysed using axial and selective coding as codes were pre-determined in the operationalisation table (Creswell and Plano-Clark, 2011). This was useful to the research given the short time spent in the field. However, open coding was also used in identifying more specific codes and unanticipated themes that emerged from the data. Such themes were incorporated into later interviews and FGDs.

3.6 Ethical Considerations and Limitations of Research

In line with the ethical considerations proposed by Bryman (2012), all respondents’ identity and records are kept confidential. Respondents were made aware that the research was taking place and given the option to partake in it. Consent for participation was obtained verbally and respondents were made aware that consent could be withdrawn at any point throughout the data collection process. It was also ensured that the research assistants were proficient in both English and Twi and could be trusted both by respondents and the researcher in obtaining consent (Alaei et al., 2013). Respondents were fully briefed on the subject and purpose of the research before formal data collection began. Such measures are especially

(28)

27

important to this research as it concerns the agricultural sector, which gets much attention from the government and other organisations. It was therefore important that it was made clear to respondents that the research is for a master’s degree at the University of Amsterdam, and not for governmental or any type of aid purposes. As a small token of gratitude, respondents were offered something to eat and drink before or after interviews and FGDs. Their travel costs were also reimbursed. Finally, the research takes a gender sensitive approach, which is especially important considering the highly gender segregated nature of labour in the oil palm sector. Accordingly, gender disaggregated data was produced in order to allow for a gender analysis (Leduc, 2009).

The research was designed in such a way as to maximise the quality of the data collected. Nonetheless, certain limitations are inevitable and there were a number of methodological issues that ensued. Due to the fact that three enumerators were involved in the survey, reliability was an issue. This was further complicated by the fact that the enumerators were translating the questions in the survey from English into Twi. As a result, issues of internal validity and consistency emerged (Bryman, 2008). In order to overcome this issue enumerators were trained together and the meaning of each question was explained in depth. They were also encouraged to work together on the translation of the questions in order to ensure they were translating it the same way. The same strategy was employed with the interviews as internal reliability was also an issue for similar reasons (ibid). Again, the two research assistants and the researcher worked closely together, discussing at the end of each day to make sure everyone had the same understanding of the research. Reliability was further strengthened through the use of an interview guide, which allowed cases to be compared equally (ibid).

Another limitation was presented by the short time frame spent in the field, which left little opportunity to gain an in-depth understanding of the local context (Creswell, 2009). This can lead to a bias in information through misconceptions or misunderstandings as the researcher may have an impact on the data presented by participants and how this data is interpreted. Furthermore, using translators inevitably leads to certain nuances becoming lost in translation. These limitations lead to issues of credibility and reliability of findings. However, the research assistants involved in the research were highly educated students from the University of Energy and Natural Resources (UENR) who were proficient in both English and Twi. Furthermore, they perfectly understood the purpose of the research at hand and had previous experience collecting data in the area. Finally, the subjectivity of the research assistants is noted, as is the subjectivity of the researcher, allowing for confirmability of the research (Bryman, 2008).

Another way credibility and reliability were strengthened in the research design was by means of triangulation. Firstly both quantitative and qualitative data collection methods were used. Using both methods together compensates for the limitations of the respective methods. It also increases the internal validity of results by triangulating the data (Bryman, 2006). Secondly, within the qualitative data collection a number of different methods are used namely interviews, FGDs and observations. The FGDs were held after the interviews had

(29)

28

been completed and they allowed the research to be triangulated with communal discussion whereby respondents challenge each other’s opinions. This helps improve internal validity by checking the accuracy of the findings (Creswell, 2009). Credibility was also enhanced through the creation of trust between researcher and respondents. As many of the respondents participated in the research on three different occasions, a relationship was established between themselves and the researchers, leading to the development of trust and understanding.

The research design also raises the question of transferability. As purposive and convenience sampling was used, sample bias is inherent to the research. The restricted research area and the limited number of respondents add to the inability of the research to be transferred to other contexts or times. However, it has been made clear from the beginning that the objective of this research is not to be representative, but rather illustrative. Indeed, transparency has been employed on all levels of the research with all choices, methods and problems indicated and explained. As a result, dependability of the research is enhanced (Bryman, 2008).

3.7 Conclusion

This chapter has provided a detailed description of the methodology and methods used in the collection and analysis of data in the research. Justification for each decision is presented with a view to ensure transparency and transferability of the research. In can be concluded that, while the research has faced challenges in generating high quality results - especially due to langue barriers and use of more than one assistant, - a number of measures have been employed to ensure that data is nonetheless of high quality and is both valid and reliable.

Peasant agency in global value chains : an examination of how small-scale oil palm farmers in Ghana exercise agency through peasant agriculture