Measuring individual social capital and estimating its returns in small-scale communities of farmers and retailers in the tropics

-- WORKING PAPER --

Authors:

Alain Desdoigts, IEDES Université Paris 1 Panthéon-Sorbonne

Francesco Cordaro, Mars-Catalyst Sr. Scientist

Bruno Roche, Mars-Catalyst Managing Director

 

January 2016

Abstract:

Is there a link between an actor's social capital and that individual's economic performance? This study seeks to create a three-dimensional capital space and calculate a social capital coordinate vector for each actor within that space. Questionnaires have been administered among households living in small communities in Ivory Coast, Papua New Guinea, Tanzania and Vietnam. First of all, the main structure of this space is found to be much the same across countries. The three dimensions or attributes of the social capital structuring the space are, in descending order of importance: i) Social exclusion, ii) Trust and behaviours and iii) Collective action. Secondly, we show how actors and subgroups of actors are located in this space in the light of their socio- demographic (religion, ethnicity, indigenous status) and economic (size of farmed land) characteristics in line with a series of solar systems that may nonetheless be overlapping within the same galaxy. Thirdly, an examination is made of two kinds of occupations accounting for a large majority of the workforce in these countries: family farming and micro and small enterprises. We will show that, all other things being equal, an actor's level of productivity rises according to the individual's trust in the members of the community and the individual and community involvement in collective actions. Being a member of an organisation (cooperative, diaspora) is also conducive to an actor's productivity and access to formal credit.

Key words: Social capital measurement, family farming and small business, economic performance, formal and informal credit access, meta-analysis.

Preface
1. Introduction

CONTENTS

2. Preliminary remarks about the actors, their economic activities and the institutions that govern them

2.1 The actors and their main economic activities

2.1.1 Small farmers and family farming

2.1.2 Informal sector and micro and small enterprises

2.2 Formal and informal institutions

2.2.1 Institutions and macroeconomic development

2.2.2 Institutions: a microeconomic perspective

3. Social capital
3.1 Different approaches to the concept

3.1.1 Definitions: French and American schools

3.1.2 Club good and externalities

3.2 Measuring social capital

3.2.1 Measuring social capital: a brief review of the literature

3.2.2 World Bank Social Capital Assessment Tool (SoCAT)

4. Surveys and methodology
4.1 Surveys: geographies and communities
4.2 Geometric data analysis (multiple correspondence analysis)

5. Constructing a space for social capital and structural factors 5.1 Social exclusion, trust and behaviours, and collective action

5.1.1 Social exclusion
5.1.2 Trust and behaviours
5.1.3 Collective action
5.1.4 Inequalities and distribution of social capital 5.1.5 Conflicts and social capital

2

5.2 Structural factors: village, ethnicity, religion, farm size, productivity

5.2.1 Intercommunity differences in terms of social capital
5.2.2 Religion and social capital in IC
5.2.3 Plantation size (ha), productivity (kg/ha) and social capital in IC
5.2.4 Group membership, profit change and social capital of Vietnamese retailers

6. Economic performance and social capital of actors
6.1. Model and methodology
6.2 Agricultural practices, agricultural productivity and social capital in IC

6.2.1 Horizontal networks, weak versus strong ties and individualism

6.2.2 Individual social capital and agricultural productivity

6.3 Retailers’ profits, credit access and social capital in VN

6.3.1 Divide between the provinces in the north and those in the south 6.3.2 Levels of and trend in profits and social capital
6.3.3 Access to credit both formal and informal and social capital 6.3.4 General and job satisfaction and social capital

7. Conclusions, limitations and perspectives 8. Bibliography

3

PREFACE

“Social capital generally refers to trust, concern for one’s associates, a willingness to live by the norms of one’s community and to punish those who do not. These behaviours were recognised as essential ingredients of good governance among classical thinkers from Aristotle to Thomas Aquinas and Edmund Burke. However, political theorists and constitutional thinkers since the late eighteen century have taken Homo œconomicus as a starting point and partly for this reason have stressed other desiderata, notably competitive markets, well-defined property rights, and efficient, well- intentioned states. Good rules of the game thus came to displace good citizens as the sine qua non of good government. "Bowles and Gintis, “Social capital and community governance”, Economic Journal, 2002.

"Alongside the things that are sold and those that are given there are some that are not to be either sold or given, but which must be kept and passed on, and these things are the supports of identities that survive better than others over time.” Maurice Godelier, Au fondement des sociétés humaines, 2007.

Thanks to:

  • -  Papua New Guinea (2011): Paige West (Columbia University) + PNG IBR

  • -  Tanzania (2011): Amy Stambach (University of Oxford / University of Wisconsin-

    Madison + BUMACO)

  • -  Côte d’Ivoire (2012) : Socio anthropologists from the Department of Sociology,

    Université FHB + ENSEA

  • -  Vietnam (2013): DEPOCEN

4

Introduction

As an actor's social capital is a highly complicated issue, resulting from a combination of several items (networks, leadership, trust, behaviour, beliefs, compliance with norms and rules,...) none of which are tangible, it is understandable that an assessment of social capital along the lines of human capital creates a genuine challenge.

However, this is the initial aim we are seeking in the context of this research project, as a resulting of adopting an approach akin to Bourdieu (1984), which, ultimately, is found to be particularly useful and illuminating. Our main concern is small communities of actors in the tropics (Ivory Coast, Papua New Guinea, Tanzania and Vietnam). We construct a three- dimensional space for social capital providing an opportunity to locate each actor, assigned a three-coordinate vector as a reflection of the individual's social capital. These coordinates obviously do not have any absolute sense. They are the Euclidean distances between the individuals in one community that matter here: the distribution of the social capital of actors in a community or a religious, ethnic subgroup or membership of an organisation (e.g. cooperative, diaspora,...)

We next estimate the returns to an actor's social capital in terms of the individual's productive or commercial activities or use of credit for the two types of activities, which account for a large majority of the workforce in the countries where the surveys were conducted: family farming (cash crop) and retailers and street traders (retailers & street traders).

Consequently, this study focuses on measuring the social capital of an actor and the actor's private return, rather than the aggregated social capital of all the actors in the same community. Hence, we are more concerned with the impact of social capital as the attribute of an actor on the individual's economic performance than the process involved in its formation and accumulation or than the impact of a community's aggregated social capital on the outcomes of its members. For example, an actor joining a club is one of the most common modalities for investing in social capital. What we are mainly concerned with in this study is the fact that this actor is a member of a network or organisation rather than the actor's reasons and motives for becoming a member (see, under this heading, the thoughts of Glaeser et al. 2002 and Sobel 2002). Focusing on the outcome instead of the process and

5

preferring parsimony to complexity, we favour the economic approach to the anthropological one. This is not because the former is automatically preferable to the latter. It all depends upon the aims in view. Combining the two approaches would no doubt be illuminating in several respects but this multidisciplinary aspect is still subject to methodological obstacles and resistances to be found in the early stages of this study in the field of social science research (Bardhan and Ray 2006).

2. Preliminary remarks about the actors, their economic activities and the institutions that govern them

This section is intended to describe the key components of this study: first of all the actors surveyed and their main activities, and secondly, the colonial history of the communities within which these actors interact and institutions they have inherited, on the one hand, and those they wish to retain and existed prior to colonisation, on the other.

6

2.1 The actors and their main economic activities

Conducted between 2010 and 2014, our survey applies to two main types of actors: i) small farmers whose main occupations are related to cash crops, such as coffee and/or cocoa and ii) micro-entrepreneurs selling end-use products, such as chewing gum or cereals or chocolate bars

Family-oriented economic activities involving the production of goods and services not only for consumption purposes but also for selling purposes are what the majority of households in the poorest countries are engaged in, thereby representing the backbone of their economic development. These activities have to contend with labour markets, credit opportunities or incomplete land and, consequently, inefficiencies that have serious implications in various ways, such as risk management, access to sources of information, and failing to keep up-to-date with the technological practices adopted.

2.1.1 Small farmers and family farming

At the instigation of the United Nations, the year 2014 was proclaimed the International Year of Family Farming (FAO 2014). This is because the family farm differs in many ways from the agricultural production processes to be found in developed and industrialised economies. A large number of development actors (states, institutions, multinationals,...) would like to be able to rely on more productive systems of farming, generally seen from the perspective of intensive farming and based on market reforms and, in particular, land tenure arrangements (World Bank 2007).

At field level, Western visitors continue to be struck by the lack of equipment and materials, the so-called physical capital available to small farmers in the tropics. An experienced eye is also surprised by the technological backwardness of the agricultural practises used. Small farmers in this region do to tend to be poorly educated and are very often unable to read and write. Some do complete primary school but only a few attend secondary school. Almost no-one goes on to higher education. In short, human capital is also lacking in this case, including from a health perspective, although huge improvements have been made in recent years thanks to the momentum provided by various development actors.

7

To sum up, there is barely any accumulation of physical, human and technological capital to underpin agricultural activities in these regions, whereas economic development is dependent upon building upon these production factors, which continue to be key sources of growth (Durlauf, Johnson and Temple 2005).

If you stay here for a few days, it does not take you long to become aware of the lack (or dire shortage) of paid workers from outside. The workforce is derived from within the farm household: the family in the more or less broad sense of the term. Sometimes western observers who have forgotten that their grandparents, too, helped out on the farm when they were children, will feel uncomfortable seeing school-going children (or otherwise) taking part in farm activities.

As for what remains as a production factor, that is the land, or, more precisely, access to land. In the tropics this means mother earth because mother earth is sacred and hence inalienable. Economists refer to collective goods. Underpinning the very principle of the free market economy, and forming the foundation of the individual-based society(Godelier 2007), the principle governing the right of private property does not therefore apply in these areas of the world to the primary production factor: the land. In this respect a distinction has to be made between a right of use or usufruct and transfer and administration rights, i.e. the right to define the rights of others (Schlager & Ostrom 1992). This means someone may possess land but not be the owner ... What is more, land tenure in these regions, and specifically in Sub-Saharan Africa, is based on the relationships individuals enjoy within the community and kinship. Consequently, the land itself gives rise to many interrelated claims and constant discussions about these claims (Berry 2001). It should be stressed, however, that this form of rural land management system has for a long time prevented the emergence of landless populations, inequalities (Clastres 1976).

Land management is therefore inextricably linked to human governance. It is not a question of private property in the de facto sense of the term, while economists claim land security is supposed to: i) increase small farmers' ability to reap the benefit of their investments without the threat of expropriation; ii) improve access to credit with land being available as collateral iii) improve trading gains or, similarly, reduce the opportunity costs when weighing one activity against another, e.g. paid employee and/or small farmer; iv) be more conducive to innovation when the collective management of a resource may hinder

8

innovation as a result of free rider behaviour, for example (Besley 1995). Land insecurity could therefore be the cause of the low level of investment and, by the same token, of agricultural productivity in these regions (see recent economic literature reviews by Place 2009, Fenske 2011 & Udry 2012).

The findings of the studies have not yet managed to produce a consensus. For example, Brasselle et al. (2002) referred to in Fenske (2011, p. 139) asserts: “No clear-cut conclusion emerges... whether their authors deal with areas dominated by systematic titling procedures, optional titling, or informal enforcement of land rights. "More generally speaking, we see a dependency on the context and the way the studies were conducted. For example, some customary rights systems may provide the land security necessary for investment. In the case of Ivory Coast, the lack of property rights has not prevented the further development of perennial crops along the lines of coffee and cocoa and, nowadays, rubber tree production. The series of laws published since colonial times (the latest law was issued in 1998) seeking to switch from custom-based land tenure to a statutory title deeds system failed to change anything. However, even though customary rights are very often confined to a right of use, small farmers rarely have the plots assigned to them expropriated if they are being ostensibly farmed.

What has to be borne in mind at this stage is that rural land management in the tropics is based first and foremost on a social relationship. This refers to ties between people in the case of access to land and its resources available in ever-decreasing quantities. In common with Pierre Gourou (1991), many anthropologists and geographers have enabled us to gain a better understanding of the overlapping of economic, human and social issues in the context of land management in these communities. As for the possibility of an African defining feature, Gourou writes (p.8): "African farming is intelligible only when regarded in the context of African civilisation... (Its) reform is dependent not only on the provision of top- quality seeds, fertilisers and power-driven machinery. It is dependent first and foremost on the development of supervisory techniques."

Family farming practised in the small rural communities surveyed in the tropics therefore produces a profile that is a far cry from the ideal most Western economic actors have about the so-called higher efficiency of farms, which tend to be bigger, preferably mechanised, relying on modern intensive farming practices (equipment, chemical inputs,

9

seeds obtained through transgenesis,...) and primarily based on paid labour.

2.1.2 Informal sector and micro and small enterprises

The informal economy's relevance in developing countries is particularly high, indeed even higher than that of the formal economy. This is particularly true in the small business sector. Street traders have been acknowledged as tourist-pullers for several decades now, while it has to be said that the fast-paced development in these countries has played a key role in the ever-increasing numbers of these micro entrepreneurs in the urban environment. Consequently, they now represent a major challenge for the economic development of their countries, hence the importance of gaining a better understanding of their organisational systems and their socio-economic operations.

We agree with Banerjee & Duflo (2011) and Cling et al. (2012) as regards the specific character of this field of activity, where small traders and street traders opt for this line of business more by default than by design. This does not necessarily mean they do not enjoy their work but the origins and reasons for this extensive informal economic activity, now universally recognised as a fact of life, lie outside the scope of this study.

We should nonetheless stress that it is questionable at this stage whether the informal sector should be equated with an illegal sector that must be eliminated (Hodgson 2006). The term "unorganised sector" is, according to Guha-Khasnobis et al. (2006), preferable, involving informal systems of governance and organisations that are crucial for the local economies, thereby complementing and even replacing the formal regulatory and organisational institutions as a result of being more credible. As Meagher notes (2005), a sentiment recently echoed by Totolo (2013) in his examination of the social networks available within small companies in Nairobi: “Many leading commentators on the informal economy have abandoned the informality paradigm in favour of a focus on the organizational role of social networks... Rather than representing economic informality in terms of an absence of regulation, social networks portray the informal economy as alternative forms of regulation operating outside the framework of the state.”

Accordingly, informality may be the rule rather than the exception, while the economic and social ties the actors have willingly or unwillingly with their families and the communities where they live and/or pursue their economic activities, should be determining

10

factors for their daily professional activities. Not only in terms of access to credit, sharing risks and information about any opportunities there may be or their trust in their peers but also in terms of learning how to keep their businesses afloat or even how to expand them. More generally, this raises the question of what strategies the actors are able to develop together to cushions the shocks for which they have no formal protection. These strategies could be based, for example, on their networking or collaborative activities, or via the institutional arrangements developed in the communities to which they belong.

As in the case of farming households, it is a bit of an understatement to say their accessible resources are limited. The physical capital available to them and skill levels are often low, or even non-existent. First of all, the sustainable livelihoods approach regards social capital as one of the most important assets for these actors, who are amongst the most vulnerable and least well-protected operators in the developing countries (Lyons & Snoxell 2005). Secondly, the new information economy approach based on social networks also highlights the relevance of ties and social networks operating as a bridging arrangement in the absence of formal institutional and regulatory structures (see, for example, Dixit 2004 and Meagher 2005, 2006).

11

2.2 Formal and informal institutions

2.2.1 Institutions and macroeconomic development

All the communities we surveyed (Ivory Coast, Papua New Guinea, Tanzania and Vietnam) used to be colonies of European powers (France and Great Britain) over a period of decades or even several centuries but have now all gained independence. They all form part of a group of over 200 nation states belonging to a globalised world from which colonial empires and the Soviet Union have disappeared within the space of less than 50 years. The people living in these former colonies are not necessarily residing within the borders and structures of the states inherited from the colonising countries. Several of these "artificial" states are therefore made up of a large number of local societies that are not always willing to give up their identities, thus building up tensions with each other. They also have to come to terms with a form of state based on the separation of politics and religion (Godelier 2007).

It is worthwhile noting that the distribution of linguistic diversity is effectively represented by a Gaussian curve centred on the equator. The distance to the equator is also a good indicator of economic development, thus suggesting to certain influential UN economists, such as Jeffrey Sachs and his team (1999, 2006), that there used to be a tropical determinism and solely massive amounts of aid would enable these former colonised countries to escape the poverty trap. Which means in this context achieving their green revolutions and attaining industrial development in the light of an intelligent, targeted coordination of investments. This approach is based on the so-called Big Push theory originally proposed by Rosenstein-Rodan (1948) and supported in the mid-2000s by key figures such as Tony Blair or Bono the singer with the rock group U2. However, this geographical theory was unable to withstand the introduction of variables describing the institutions1 that are now the key variable conducive to the accumulation upstream of production factors and innovation.

1 The concept of institutions or "good governance" is one that is complicated and fairly abstract in 12

In the field of economic science, Acemoglu, Johnson & Robinson (2001) have made a key contribution to the debate about the economic development roles played by institutions. These authors first of all highlight the close link between the high level of settler mortality and the contemporary level of economic development, which they attribute to a third variable: the risk of being expropriated. They then make what is potentially a very instructive link. Regions that experienced high settler mortality rates saw the emergence of states with short-term visions and a policy based on extracting local wealth, whereas in places where settlers survived, institutions more conducive to economic development were apparently created.

The huge amount of work accomplished by Acemoglu and his team since the early 2000s (e.g. 2005, 2012) resulted in a huge upsurge in the amount of work focused on the role economic and political institutions play in the case of development. With the upshot that the relevance of this research work is now universally accepted. Consequently, the history of European colonialism is now exploited as a natural experience to be used to investigate how institutions may affect a country's economic development owing to factors such as their tendency to grant priority, or otherwise, to eradicating inequalities (Engerman & Sokoloff 2006). However, the "black box" still has to be broken open, i.e. there has to be a better understanding of what kinds of institutions are the most conducive to the economic development of a group (village community or state), and at what stage of its cultural, political social and economic development?

2.2.2 Institutions: a microeconomic perspective

This concern for the institutions as a cause of contemporary inequalities between countries has helped to rekindle microeconomic research interest (e.g. game theory and contracts theory), where the focus has long been on trying to understand how the institutions may have both a positive and negative impact on the incentives for the various actors. It is no coincidence that so much of the work on economic development applies to agricultural institutions and the importance of networks where our actors engage in

practice. Accordingly, special attention has to be paid to this concept to avoid it becoming an all- embracing concept. Conversely, there is a clear recognition of what the institutions do not cover: climate, landforms, natural resources (i.e. the natural environment). Nor are these objects produced by human beings using natural resources. It will subsequently become evident that a similar problem has to be overcome in the case of social capital.

13

business and carry out their economic activities (see above and the thoughts of Bardhan 2010, Besley & Jayaraman 2010).

However, let us return for a while to the definition of institutions given by the Nobel Prize Winner Douglas North (1994): “... the humanly devised constraints that structure human interaction. They are made up of formal constraints (rules, laws, constitutions), informal constraints (norms of behaviour, conventions, and self-imposed codes of conduct), and their enforcement characteristics." He also refers more briefly to "ground rules". Formal institutions are those codified and created by political institutions, while, in line with social norms, informal institutions are under the supervision of peers and the actors are subject to moral obligations2. Moreover, the constraints to which North draws our attention help to cut down transaction costs, i.e. those associated with buying and selling goods, investment, credit,... The more complex the products and markets where they are traded, the more important the institutions. So it is easier to understand why the lack of property rights for land (a property that according to custom, as mentioned above, is inalienable because it is sacred) is interpreted by economists as an obstacle to economic development.

To sum up (see Roland 2014), the institutions may be said to help to solve typical problems: i) information asymmetry; ii) hold-up; iii) cooperation; iv) coordination; v) enforcement. Each problem is matched with a formal or informal solution, or indeed a mix of formal and informal solutions. For example, information asymmetry is the source of two problems: i) adverse selection where one of the agents, e.g. the buyer or the "client ", does not have full information about the good the party would like to acquire, e.g. a second-hand car (lemon); ii) moral hazard where the action of one of the agents, e.g. the insured person or wage earner, cannot be verified by the one who is called the principal, the insurer or employer respectively, or the "patron "in the renowned "patron-client "relationship. This is a well-known type of problem on the labour and insurance markets, markets which are very incomplete when they exist or are accessible to actors in the communities we surveyed.

And yet, our small farmers in Ivory Coast and Papua New Guinea, in common with our

2 Four years previously, North (1990) wrote: “I wish to assert a fundamental role for institutions in societies: they are the underlying determinants of the long-run performance of economies – Third World countries are poor because the institutional constraints define a set of pay-offs to political/economic activity that do not encourage productive activity... The inability of societies to develop effective low-cost enforcement of contracts is the most important source of both historical stagnation and contemporary underdevelopment in the Third World.”

14

micro-entrepreneurs in Vietnam, have to contend with risks they no doubt wish to insure themselves against. Nor are they always able to verify the efforts made by workers they may have hired to work on their farms. Or they have to be able to trust their partner in order for the transaction to take place. These types of problems could be solved as a result of: i) laying down rules on the disclosure of or even the obligation to make any relevant information about the transactiontransparent; regulating admittance to the profession; providing appropriate product quality guarantees. These rules are fairly reliable when an efficient legal contract enforcement system is available. In the case of informal institutions, these tend to be recurrent interactions, with the reputation or the contractual ties helping to remedy the information asymmetry problem. The hold-up problem may be overcome on the basis of an enforceable contract binding the parties. The formal contract continues to be the best solution. As in the case of land transactions, the informal approach generally involves the presence of a witness to ensure the parties honour their commitments

We are not claiming here to have covered the entire issue but these examples do help to show just how resourceful small rural communities or traders in tropical areas can be in coming up with completely rational answers to economic problems, such as information asymmetry or the struggle trying to get all sides to meet their commitments. Once again, what are missing are not the institutions, which are more informal than formal in this case, but the completeness of the markets (when they exist).

Development economists are therefore compelled to behave like anthropologists to some extent by stepping away from the actors being studied and their environment. They have to draw upon economic theories that are different or at least more suited to the context and the development issue being investigated. This will be discussed later on. This has been undertaken by Fafchamps (1992), for example, who tells us about importance of these mutual assistance networks in pre-industrial rural societies and their role in insurance matters, or Greif (1994), who considers the part played by cultural influences and the social networks of Mahgribi traders in terms of providing incentives to meet their commitments (contract enforcement). More recently, Munshi (2014), too, pointed to the networks or ties between actors as an answer to the flawed markets.

15

3. Social capital

When markets are incomplete or even non-existent and formal institutions are weak, and the contemporary independent state (weak) has not completely succeeded in becoming a component of traditional civil society, informal institutions are quite likely to take precedence or carry more weight than formal institutions in cultural, political, social and economic interactions, on the one hand (Bowles and Gintis 2002), consistent with the development of social norms within civil society, on the other (Ostrom 2000).

At the very least, what we are seeing is stacking if not overlapping in the case of two types of institutions. What matters then are the more or less lasting ties the actors in a community establish between themselves, their rights and their duties, not only as members of an actual group (e.g. nuclear family, clan, ethnic group) or out of choice (e.g. association, cooperative) but also sometimes as spokespersons for the group to which the actors belong or represent. We will be returning to this subject but a reference may be made here to the concepts of weak and strong ties introduced by Granovetter (1973, 1985) and echoed by Putnam et al. (1993) in their discussions of horizontal and vertical networks. What has to be remembered here is that the latter agree with the former that the density of the weak ties established by the members of the community and included within its midst (e.g. being a members at the same time of several groups, associations or cooperatives and playing an active and legitimate role within this context) is a) more important than establishing intense personal ties (e.g. extended family, clan) for the purpose of maintaining social cohesion and b) more conducive to productive activities and collective action. (see, for example, Ostrom & Ahn 2003, a work containing some of the most important texts published during the 20th century and, more specifically, since the mid-1980s.)

3.1 Different approaches to the concept

The concept of social capital is a highly complicated one, as underscored by the many definitions provided by researchers from the different humanities and social science disciplines. This section will first of all offer a brief survey of the huge variety available, then specify the concept of social capital as envisaged by economists and in its most relevant version in the light of our approach.

16

3.1.1 Definitions: French and American schools

- French school:

As mentioned very early on by the sociologist and anthropologist Marcel Mauss in his research programme introducing his article Essai sur le don. Forme et raison de l’échange dans les sociétés archaïques (The gift. Form and reason for exchange in archaic society) (1923-1924), the concept of social capital appears to be highly complicated. He writes: "Everything intermingles in them, everything constituting the strictly social life of societies that have preceded our own, even those going back to protohistory. In these ‘total’ social phenomena, as we propose calling them, all kinds of institutions are given expression at one and the same time—religious, judicial, and moral, which relate to both politics and the family; likewise economic ones, which suppose special forms of production and consumption, or rather, of performing total services and of distribution..."

This "everything intermingles" is an underpinning for the concept of embeddedness developed by Karl Polanyi in his work The Great Transformation (1944) then by Granovetter (1985). This concept refers to the economic sphere being integrated into the social sphere or, conversely, as was the case during the 19th century, its autonomy when the term ‘disembedded’ is used and a reference is made to decentralised and fragmented markets comprising actors operating independently of each other while pursuing their own interests.

No development economists dispute this state of affairs nowadays or are reluctant for their analysis to include (when relevant) actors seeking an admittedly individual but socialised objective, where the reason and cognitive skills are limited. Looking back, this is probably one of the most significant implications of the concern shown for the concept of social capital in recent decades. Uniting different disciplines of the humanities and social sciences within the same research topic.

At this stage, a useful parallel may be drawn between the idea of embeddedness and custom-based management of rural land in the tropics, which, as we have seen, cannot be undertaken independently of the formal and informal ties established between the members of a community (e.g. nuclear and extended family, indigenous people versus migrants,...), whose rights and duties in relationship to certain pieces of land are embedded in the very structure of the social and political relationships they enjoy within the

17

community (Chauveau & Colin 2010). Land administration is therefore consistent with a first tier of embeddedness via: i) the inclusion of economic activities within networks (e.g. a network's role in seeking land to be farmed); ii) the cognitive and cultural constructs regarding the objects being traded (e.g. what rights are legitimately transferred?) and the parties involved in the transaction (e.g. who may legitimately trade with whom?).

In Les formes de capital (1986), the sociologist Pierre Bourdieu makes a distinction between three types of capital: economic, cultural and social. He defines social capital as "the sum of the resources, actual or virtual, that accrue to an individual or a group by virtue of possessing a durable network of more or less institutionalized relationships of mutual acquaintance and recognition [...] The volume of social capital possessed by an agent depends in particular on the size of the network of connections the agent can actually mobilise and the volume of capital (economic, cultural or symbolic) personally owned by those with whom the agent is connected."

- The American school:

The definition put forward by the American philosopher and economist Francis Fukuyama (2000) and cited in Durlauf (2002) makes it possible to identify many of the insights underlying social capital that we will be calling upon later on this study: “Social capital can be defined simply as an instantiated set of informal values or norms shared among members of a group that permits them to cooperate with one another. If members of the group come to expect that others will behave reliably and honestly, then they will come to trust one another. Trust acts like a lubricant that makes any group or organization run more efficiently.”

It is also recognised that social capital makes agents more trustworthy because an individual is less inclined to cheat another person if the latter is part of the former's network. The wider and more extensive the network, the more people you can trust, which in all likelihood should promote productive activities (see Putnam 2000 and the concepts of bonding versus bridging social capital). This is illustrated by the example Coleman (1988) gives of the New York diamond market and its Hassidic Jewish actors. He concludes (p. 119): “In explicating the concept of social capital, three forms were identified: obligations and expectations, which depend on trustworthiness of the social environment, information-flow capability of the social structure and norms accompanied by sanctions.”

18

All of these definitions lead to the conclusion that social capital refers to: i) the relevance of the social structure available for each and every member of the community; ii) the social networks of actors that are both informal (family, contacts, diaspora) and formal (membership of organisations and associations); iii) the quality of the relationships individuals have with each other and society. The relationship between individuals, social norms (solidarity, reciprocity), and culture (values and beliefs) are therefore regarded as factors with the potential to contribute to the coordination of individual actions and cooperation enabling individuals and institutions to achieve shared goals (Ostrom & Ahn 2003).

3.1.2 Social capital: club good and externalities

A point of reference for understanding the concept of social capital as studied by economists is the chapter written by Durlauf & Fafchamps for the Handbook of Economic Growth (2005, chapter 26).

For example, how can social capital be equated with capital? In common with physical capital, social capital may be accumulated by individuals subject to a certain cost. It has a private return and needs to be maintained otherwise it could depreciate. In common with human capital, which depreciates when the individual leaves an enterprise, an actor's social capital may depreciate upon leaving a community to join another one. It is often claimed that this is not an individual's property in the same way as a tool is for a farmer or a stock of goods is for a trader, and it cannot be sold. This is true only to a certain extent (e.g. an address book has an exchange value in addition to its user value) and can be transferred from one individual to another.

Physical capital and human capital are in fact private goods, whereas social capital may be regarded as a club good. This is a non-rivalrous good such that one individual using it does not prevent another from doing do, while social capital is partially excludable, so an individual may be deprived (excluded from) access to this capital. Information is an example of a type of good that is both non-rivalrous and partially excludable.

Accordingly, social capital or, more specifically, the intensity and nature of interactions between actors, may generate positive externalities for club members and produce

19

spillovers through such things as sharing technologies and information among members. An unfortunate consequence of the externalities produced is that actors generating social capital do not reap all the benefits of their investment, resulting in a suboptimally balanced social capital level/stock being available to the community. This therefore raises a question for development actors: is it possible to work towards the development of this specific form of capital?3

Coleman (1988, pp. 100-101) draws a parallel between these three forms of capital: physical, human and social. We will be invoking this parallel in the subsequent part of our analysis, in the light of our objectives and this is going to be particularly relevant: “If physical capital is wholly tangible, being embodied in observable material form, and human capital is less tangible, being embodied in the skills and knowledge acquired by an individual, social capital is less tangible yet, for it exists in the relations among persons. Just as physical capital and human capital facilitates productive activity, social capital does as well. For example, a group within which there is extensive trustworthiness and extensive trust is able to accomplish much more than a comparable group without that trustworthiness and trust.”

3.2 Measuring social capital

This section offers a short review of the literature and a brief presentation of the survey derived from a project the World Bank initiated in the late 1990s.

3.2.1 Measuring social capital: a brief review of the literature

It should be stressed that as no single definition of social capital is available and it is unrealistic to be able to sum up all its dimensions in just one indicator, the measurements being proposed will always involve grey areas with all the obvious implications for statistical inference. However, as the two Noble Prize winners Robert Solow and Elinor Ostrom respectively stress: i) “If social capital is to be more than a buzzword [...] social capital should somehow be measurable, even inexactly”; ii) “Social capital, with only a decade of history of

3 What also has to be borne in mind is the negative factors that may sometimes be associated with social capital (corruption, mafia, democratic elections whose outcomes are mainly based on the ethnic criterion,...). In the wake of Portes (1998), many observers have highlighted these negative externalities resulting from some group activities to the detriment of non-members

(i.e. excluded).

20

empirical applications and attempts at measurement, does exhibit serious problems of measurement. [...] It would not be wise at all to dismiss the concept on the grounds that it is difficult to measure”.

A problem has been encountered and has still not resolved for the many variables economists use (favouring the quantitative approach) as with human capital, for example. Looking back and with the perspective of roughly 50 years, it is debatable whether human capital research has been pointless in terms of the benefits for scientific knowledge and the development actors and populations covered. The social sciences and development actors should obviously show as much concern for social capital as other forms of capital, and particularly human capital, say, but we do realise that an actor's social capital conflates many more factors than the actor's human capital, for example. One of the key challenges of this study was to construct a parsimonious measurement approach that is also as exhaustive as possible.

The standard practise for estimating the benefits of social capital with a dependent variable of interest involves running regressions of the latter on a set of control variables and a social capital measurement. There is now a significant body of empirical literature on social capital. One of the most influential studies at macroeconomic level is the one completed by Knack & Keefer (1997), where the aim is to explain growth within a wide cross- section of nations using a measurement of the degree of trust obtained on the basis of national survey data; data derived from the World Value Survey, where each observation is the outcome of the responses given to a questionnaire by roughly 1000 individuals representative of the overall population in each country in the sample, and where trust is measured according to the percentage of responses for which, ‘most people can be trusted’ and the argument ‘you can’t be too careful in dealing with people’ (from roughly 8% in Brazil to 60% in the Scandinavian countries). Their findings show that trust is higher in countries with a higher level of development, fewer inequalities and more ethnically homogenous populations. Are these findings applicable to the tiniest communities of small farmers engaging in family farming and the family micro-enterprises so prevalent in the tropics?

Social capital may be regarded as the glue binding a society's members together, emerging as soon as individuals interact within an organisation or a group (association, cooperative, family, friends, religion,...). This is explains why most social capital indicators

21

are of the microeconomic variety in common with the membership rate and level of participation in organisations or institutionalised or non-institutionalised networks, degree of social trust, ability to coordinate around collective actions, level of solidarity,... Readers interested in the relatively recent state of the art in terms of empirical literature should refer to Durlauf & Fafchamps (2005), who make a distinction in empirical studies on social capital in developing countries between studies about its measurement and impact on a variable of interest, such as household expenditure or a micro entrepreneur's profit, and those concerned more particularly with the formation of social capital, i.e. its determining factors. Our objective fits into the context of the first study group. Presented below is key information Durlauf & Fafchamps (Table 1, pp. 1674-75) have gathered together for a selection of studies already published and which our study resembles: Narayan & Pritchett 1999, Maluccio et al. 2001, and Fafchamps & Minten 2002.

{TABLE 1}

3.2.2 World Bank Social Capital Assessment Tool (SoCAT)

To these studies should be added those undertaken within the framework of the World Bank. Academic interest in social capital was soon extended to development actors. This is in line with the World Bank, which proceeded in 1996 to roll out a large-scale qualitative and quantitative data-gathering programme in small rural and urban communities in tropical areas, seeking to develop appropriate and relevant social capital measurements4 (Dasgupta & Serageldin 2000, Grootaert & Van Bastelaer 2002, Grootaert et al. 2004, www.worldbank.org/socialdevelopment). The integrated questionnaire of Grootaert et al. (2004) has a qualitative component, where socio-anthropologists hold talks with particularly important or influential members (even supposedly) of the community: village leader, land chief, leaders or managers of an /association, subgroups of members of the same organisation/association, subgroups of actors who are not members of this organisation/association,... There is also a quantitative dimension with a questionnaire targeted on households in the surveyed community making a distinction between two forms

4 The complicated nature of the social capital concept and the difficulty of measuring the outcome has already been pointed out by Grootaert &Van Bastelaer (2002): “a concept that encompasses too much is at risk of explaining nothing.” Mindful of this risk we worked on preparing measurements we developed and then used to estimate the returns to capital social (see below).

22

of social capital: cognitive social capital (inherently subjective) and structural social capital. A short description of the various dimensions of capital covered in this questionnaire is provided in table 2 below.

{TABLE 2}

4. Surveys and methodology

4.1 Surveys: geographies and communities

Four surveys were conducted on three continents (Africa, Asia and Pacific) and in four countries between 2010 and 2013: Papua New Guinea 2001, Tanzania 2011, Ivory Coast 2012 and Vietnam 2013. A more or less large number of communities (villages) and actors (households, organisation representatives, chiefdoms,...) were surveyed in each country. The target of the surveys in the first three were rural village communities whose main activity was family farming involving perennial crops (cocoa and coffee) while in Vietnam the actors surveyed were primarily small traders or micro-entrepreneurs. Consequently, it is apparently reasonable to assume that communities for which we have information available in this study have identity and cultural characteristics, are consistent with the norms and are engaged in very different agricultural and commercial practices. The formal (consequences of colonial arrangements, then the period of independence) and informal institutional arrangements governing these populations are also obviously very different. However, is it also possible to discover similarities so that an individual social capital organised in an area shared by all these communities can be observed?

At this stage it is important to stress that the quantitative data in the questionnaires for the actors were supplemented by detailed observations and accounts of conversations between socio-anthropologists in the field and key actors in the community, such as the village leader, heads of organisations and associations, the land chief, women,...

In the case of Papua New Guinea (PNG), the survey was conducted in three village communities in the Eastern Highlands. The community members' main occupation is coffee growing, in common with all the small village communities surveyed in Tanzania (TZA) in the

23

regions of Kagera and Mount Kilimanjaro. A total of 447 and 561 questionnaires were collected in PNG and TZA respectively. The choice in Ivory Coast (IC) fell on the former forest area now used for growing the perennial crops cocoa, coffee and, more recently, rubber trees. Seven communities covering the cocoa area from east to west were surveyed, while 2020 questionnaires were collected. In the case of Vietnam (VN), the survey covers nine provinces from north to south and 2465 questionnaires were collected.

This study is focused primarily on two geographies (Ivory Coast and Vietnam) for which large and economically representative samples are available, because the data gathered covers all of IC's cocoa production areas and VN's national territory. In contrast to PNG and TZA, we also have information in IC and VN about economic activities, farm practices and returns for small farmers in IC, on the one hand, and profits and turnovers for micro- entrepreneurs in VN, on the other. Lastly, the experience built up in PNG and TZA in 2011 was used to adjust the quantitative questionnaire for households and the method for gathering both qualitative and quantitative data in the field.

4.2 Geometric data analysis (multiple correspondence analysis - MCA)

How can a quantitative understanding been gained of the complicated nature of the social capital concept and the many dimensions of the questionnaire (see table 1) for actors in the communities surveyed? How can a space for social capital be created where the actors operate and be compared cross-sectionally, i.e. international, intercommunity and over time?

It should be stressed that no society, be it traditional or modern, is homogenous, while the idea of communities in which the individual is dissolved is nothing but a myth of "traditional collectivism" (Olivier de Sardan 1995). We apply the approach adopted by Bourdieu and apply a multiple correspondence analysis (MCA) to the questionnaire summed up in table 1. This analysis is designed to create a map or space for social capital for locating not only the actors in the communities surveyed but also the different villages and subgroups of actors, for example, according to their religion, ethnicity, size of farmed land, productivity,... (see also Le Roux et al. 2008). Consequently, our study is mainly intended to identify a middle way between the single variable generally chosen on an agreed hoc basis

24

to measure an actor's social capital and the theoretically inherent and even almost holistic complexity of the concept.

Within this study MCA should be understood as a geometric representation of data, converting Euclidian distances of statistical proximities between elements, as for example in this case actors or specific groups of actors (religious identities, ethnic identities,...). It allows the emergence of groupings, oppositions, trends, which cannot be identified directly, all the more so in this mix of objects constituting an individual's social capital. This value of this method is the ease of interpreting its findings, where the spatial proximity between two actors reflects the resemblance between the pair that can be identified to define the profile relative to the mean profile within the population, represented by the origin (Escofier & Pagès 2008). This will enable us to construct for a given community not a social capital indicator but a distribution of the social capital on the basis of synthetic variables revealed by an investigation of the collected data. It is therefore possible to consider comparing the distributions of social capital with each other. The analysis will be focused on the dispersion, deviations from the mean, for example, of the social capital of members of a community or indicators of inequality, such as the Gini inequality index.

Let us take the example of a question from the questionnaire in section ‘4. Exclusion’ (see table 1): “To what extent do differences in wealth and material possession tend to divide people in your (business) community, i.e. living neighbourhood)?”. The question here is a variable of the space of origin whose interpretation is being sought, and there are as many variables as questions. In other words, several dozens to be multiplied by as many individuals who replied to the questionnaire.

Another example, section ‘1. Trust’, features the variable: “Do you think that in this neighbourhood/business community, people generally trust one another in matters of lending and borrowing?”; or “In this neighbourhood (business community), one has to be alert or someone is likely to take advantage of you?”. Questions for which the survey has to choose from a number of set (coded) and mutually exclusive responses: ‘Do trust’ and ‘Do not trust’, and ‘Strongly agree’, ‘Agree’, ‘Disagree’ and ‘Strongly Disagree’, which represent what are called modalities. In the case of several dozen variables this soon adds up to 100 or so modalities ... Hence the need to reduce this space, while losing a minimum amount of information. This is what a MCA does when it processes all the responses to a survey to

25

which several thousand individuals have replied.

These questions for themes ‘1. Trust’ to ‘5. Collective actions’ are used as active variables, i.e. as variables for recording the perceptions, feelings and behaviours actors expect on the part of their peers. These questions / variables are associated with the quality of life in terms of solidarity, the responsibilities of actors, sources of exclusion, trust, solidarity,... They apply to actors' behaviours with and towards each other. They are inherently subjective. The questions for themes ‘6. Organizational density’ to ‘8. Leadership’ and the variables characterising the socio-demographic identities of actors are treated as passive or neutral variables. These are not impressions or opinions. Mention is also made of structural factors projected into the space for social capital, as what are involved are realities, such as geographic location, membership of one or more organisations or the decision-making method within these organisations. Particularly relevant variables within the context of our study are economic variables, as with, for example, the economic performance of actors or their access to credit.

5. Constructing a space for social capital and structural factors

Within this space towards which the actors gravitate, we now project structural factors as with socio-demographic and economic variables. Is social capital systematically organised between the various geographies and independently of the fields of activities? What do they have in common and what significant differences can be seen, for example, between the communities in Ivory Coast and Papua New Guinea? What is the nature of the relationships between social capital structural factors? For example, do men and women systematically reply differently? Or do people living in the same village have the same perceptions of trust and solidarity? How do small farmers regard the social cohesion within their communities compared with their neighbours farming larger amounts of land? The multiple correspondence leads to data-mining that makes it possible to answer these kinds of questions as a result of calculating the coordinates of the (average) typical members of each subgroup in the space constructed on the basis of the information each party provides in terms of themes ‘1. Trust’ to ‘5. Collective action’.

26

{TABLE 3}

5.1. Social exclusion, trust and behaviours and collective action

5.1.1 Social exclusion

An analysis of the data points to a common structure between the communities of actors while allowing the data to speak for themselves, i.e. without making any a priori assumptions about their distribution nor about a specific form of dependency. Table 3 sums up our findings for the five surveys conducted between 2011 and 2014. The chronology of the survey is seen by reading from left to right.

It should be noted first of all that the total inertia of the first three axes is roughly 80% for the five geographies. They are central to the discussion below. The first axis primarily reflects theme ‘4. Exclusion’, which contributes by 80% to the construction of this axis in TZA and VN, by 96% in IC and by only 42% in PNG. Secondly, what are important here are the questions about the sources of exclusion: “To what extent do differences in ... tend to divide people in your community?" Various modalities for sources of exclusion are proposed: landholdings, wealth and material possession, age, gender, political, ethnicity, old and new generations, education,... The comparative significance of these sources of exclusion nonetheless differs from one country to another. The wealth and material possession- related differences are a social of social exclusion that comes high up on the list (top 3) in all the countries but it is the landholdings, political and ethnic-related differences that take precedence in IC and PNG, while differences in terms of gender, generations and education are apparently relatively more significant in KEN. The social status differences dominate in VN. Thirdly, the modalities contributing to the construction of this synthetic variable we call ‘Social exclusion’ are the ‘Not at all’ and ‘Somewhat’ modalities located on the left or right,

27

as the case may be, on axis 15.

It is worthwhile noting one or two specific geographical characteristics. In IC for example, the question “Do these differences cause problems?" plays a key role in the construction of the synthetic variable corresponding with axis 1, where the reply ‘Yes’ or ‘No’ is located on the right or left, as the case may be, of this axis; the modality ‘Yes’ having obtained 34% of the answers. Lastly, over two-thirds of the latter answered ‘Yes’ to the question: “Do such problems ever lead to violence?" Also noteworthy in PNG is the amount of ‘Yes’ versus ‘No’ answers to the question: “If a community project does not directly benefit your neighbourhood, but has benefits for others in the community, do you think your neighbour would contribute time/money for this project?" Roughly 40% of the actors in PNG replied ‘No’ versus only 20%, for example, in TZA.

5.1.2 Trust and behaviours

The following axis is one whose synthetic variable primarily reflects the cognitive social capital and particularly the questions for themes ‘1. Trust’ and ‘3. Behaviour and attitudes’. First of all, the sum of the contributions these two themes make to the construction of this synthetic variable we call ‘Trust and behaviours' is equal to: PNG 21+43=64%, TZA 46+13=59%, IC 66+13=79% and VN 28+36=64%. The variables of theme ‘1. Trust’ that matter in the construction of this synthetic variable are: “If I have a problem, there is always someone to help me”, “Most people in this community are willing to help if you need it”, “Most people in this community are basically honest and can be trusted”, and “If you lose an animal, someone in the community would help look for it or return it to you”. Secondly, the modality ‘Strongly agree’ is the one located on the right-hand side of the axis in TZA and IC 6, whereas the modalities ‘Strongly disagree’ and ‘Disagree’ produce this axis in PNG. There may an axis reflecting trust in TZA and IC but it would be more accurate to refer to an axis of mistrust in PNG. Thirdly, in the case of behaviours, the question “To what extent do most

5 The modality ‘Very much’ when accounting for less than 5% of the answers has been merged with the modality ‘Somewhat’. This applies in a large majority of cases but two exceptions should be pointed out in view of the political context and the civil war that took place in the country during the decade preceding the survey. In CIV, the modality ‘Very much’ was chosen by 10% and 15% of actors in the case of differences concerning landholdings and politics respectively. In PNG, between one- quarter and one-third of the actors surveyed chose the modality ‘Very much’ in the case of differences in terms of landholding, education and wealth and material possession.

6 In CIV, no actor replied ‘Disagree’ or ‘Strongly disagree’.

28

people of this community believe they are responsible for their own words and actions?" appears as discriminatory in PNG, TZA and IC. In IC, the axis is therefore produced by the two modalities ‘Less than half the people’ located in its negative part and ‘Most of people’ located on the right-hand side of the axis. Equally important in the production of this axis in IC are the variables “I do not pay attention to the opinions of others in the village" and “people are always only interested in their own welfare”, whereas in PNG and TZA, the modality ‘No’ to the question “If people violate social norms, do they try to remedy their moral transgressions?" plays a key role in the construction of the synthetic variable.

We must once again stress that this descriptive analysis is not intended to be a replacement for a comparative cross-country analysis of the different institutional and cultural environments where trust takes shape and develops within a community.

5.1.3 Collective action

This final axis above all reflects theme ‘5. Collective action’ particularly in IC (45%) and VN (83%), where it is even more important than the axis ‘Trust and behaviours'. The modality ‘Yes’ is the one contributing to the construction of the axis for the question: “How often in the past year, have you personally actively participated in an information campaign?" This is followed by the same response for the options: ‘actively participated in an election campaign’, ‘actively participated in an association’, ‘contacted your community organization elected representative’, and ‘made a personal contact with an influential person’. Equally important, the response ‘Never’ to the question “How often have you joined together with others in your neighbourhood to address a common issue?”, a modality located on the left-hand side of the axis. A similar message, albeit in drastically different contexts, is delivered by IC and PNG in the case of the variable: “In the past two years, how often have members of this community got together and jointly petitioned government officials or political leaders with the village development as a goal?" Next comes “whether any of these actions was/were successful?" Actors in PNG and TZA provided extremely different responses to this question. Half of the actors surveyed in PNG replied ‘Never’ versus less than 5% in TZA, while over 40% of actors in TZA replied ‘Frequently’ compared with less than 3% in PNG.

5.1.4 Inequalities and distribution of social capital

29

The MCA allows a reasonably sized space to be constructed for us to measure and explain: the distribution of social capital within a population of actors. This distribution of social capital in IC and VN according to its three most important dimensions is shown in graph 1.

{GRAPH 1}

5.1.5 Conflicts and social capital

It is worthwhile emphasising at this stage the rifts between actors in the same community or violence and conflicts, which are closely correlated with the more or less favourable and optimistic perceptions actors have of the social cohesion and climate of trust within the community, or the ability of their peers to mobilise their efforts to pursue common goals for promoting the community's development. All societies have to contend with conflicts that are part and parcel of any social life, on the one hand, and are divided and split, on the other. Lastly, conflicts do not necessarily mean societies are doomed to become fragmented or beset by anarchy but, on the contrary, they may help to strengthen social cohesion (Olivier de Sardan 1995 referring to the Manchester school and the functionalist paradigm).

Our intention here is not to subject the findings of our surveys to an anthropological development analysis, entering here via the channel of conflicts. This would also mean going back to the various fields of endeavour to confirm any hypotheses. Our less ambitious goal is to confirm the relevance of the social capital measurement methodology adopted by this research programme. Consequently, it may not be just a coincidence if actors in IC who think the differences in terms of landholdings and ethnicity are a source of exclusion are the same ones replying ‘Yes’ to the question: “Do these differences cause problems?" and two-thirds of the latter also answer ‘Yes’ to the question: “Do such problems ever lead to violence?”. Combined with this is the prevalence of theme ‘3. Behaviour & Attitudes’ in the construction of the axis ‘Collective action’ for IC. In this case the modalities ‘Rarely’ and ‘Yes, are systematically’ found to be left and right respectively of the axis and replying to the question: “If people in this community violate social norms (e.g., insults, rubbish someone, stealing), do they try to remedy their moral transgressions?”. This same question also contributes to the construction of the synthetic variable ‘Trust and behaviours' in PNG and

30

TZA; There are 34% or twice as many to have answered ‘No’ in PNG compared with the percentage in TZA. Equally divisive in PNG and TZA, the two questions: “In your opinion, is this community generally i) peaceful or conflictive, ii) harmonious or disagreeable?" While less than 6% of actors apparently regard their communities as conflictive or suffering from a somewhat disagreeable climate, they account for 20% in PNG. None of these questions appear to be divisive in VN.

This analysis offers a possibly confusing view of social capital, wavering between the holistic "everything mixed together" of Mauss (i.e. social globality) and a more local approach, where the analysis is focused on its measurement, albeit partial, on the one hand, and the emphasis on actors and groups of actors, on the other. However, we will see that this space for social capital is associated with socio-demographic characteristics, such as not only religion and the ethnic group but also economic variables, as with the size of farms, agricultural practices, access to credit or the economic performance of actors.

5.2 Structural factors: village, ethnic group, religion, farm size, productivity

The analysis in the previous section reflected an investigation of the ties between social capital modalities as defined in the quantitative survey for actors without any a priori hypothesis about their socio-demographic or economic characteristics. However, is there a relationship between our space for capital in which the actors operate and the communities (villages) surveyed? Or, alternatively are the perception actors have of social cohesion and their peers' observance of social norms systematically linked to their ethnic origins or religion? A more specific question in the light of our study concerns whether an actor's economic performance is linked to an actor's social capital. Information about the agricultural practices of actors in IC and the economic performance of micro-entrepreneurs in VN has been collected specifically for the purpose of answering these types of questions. Let us now consider the situation.

5.2.1 Intercommunity differences in terms of social capital

The mean coordinates in the space for social capital of actors belonging to a community, i.e. one of the seven villages surveyed in Ivory Coast, are calculated and this

31

mean individual or individual representative of the actor's village is now located in the space, as illustrated in graph 2 for Ivory Coast. The seven villages surveyed from east to west, i.e. travelling along the so-called pioneer fronts or the various cocoa belts are: Ebilassoukro close to the Ghana border in the east and in the first cocoa belt located in the first cocoa-growing centres (1890-1910), Ande and Bodo still in the east, where cocoa- growing became established starting from 1920, Belleville in the central-western part, when cocoa-growing started to become established in the 1950s, then Soubré and Dahiri in the south-west, where cocoa production became established from 1970 onwards, and, lastly, Seïleu in the west where cocoa farming became established in 1985. The analysis of this structural factor shows that the villagers in Ebilassoukro, Bodo or Belleville tend, on average, to be located on the right of the Social Exclusion axis. The differences in terms of landholdings and wealth and material possession are, on average, more sources of exclusion for actors in these villages than for their neighbours in the villages of Ande and Soubré. Similarly, it may be said that these differences are relatively more sources of problems and violence in Ebilassoukro, Bodo or Belleville, located to the right of the origin in the left-hand graph related to Ande and Soubré, whose representative individuals are located to the left of the origin. In the second factorial design Trust and behaviours - Collective action, the individuals representative of the villages are seen to be located along the first bisector. Accordingly, the villagers in Seïleu (north-east on the design) have more trust in their peers and are more inclined to get together to undertake collective actions than people living in Dahiri or Belleville (south-west on the design)...

{GRAPH 2}

At this stage we will refrain from drawing any hasty conclusions about the quality of life in these communities. These significant differences apparent between villages are quite simply rather reassuring and should incite us to learn more about the location where clear differences are expressed, as a reflection of the interplay among actors and collective forms of certain organisations and the inclusive and community functions should not be underestimated nor should the leaders/members or "patron/client" divide be underestimated (Olivier de Sardan 1995). In any event, the relevance of a multidisciplinary approach is obvious here and the outcomes should encourage us to ensure more interaction

32

between economists and socio-anthropologists.

These differences in IC, as shown in graph 2, are also reported between villages in PNG, regions in TZA and provinces in VN. Fixed-effects (villages, regions, provinces) are almost systematically included in microeconometric studies of development, which are then only a measurement of what we do not know. It will be interesting in the following section to investigate whether the coordinates for the social capital of actors may replace these silent variables ...

5.2.2 Religion and social capital in Ivory Coast

This is another case where we do not wish to supplant a meticulous, in-depth investigation of the relevance of religion in a community's collective organisation. All we want to do is show the importance of all other things being equal reasoning. Four religions / beliefs are present in Lower Ivory Coast, where the survey was conducted: Animism, Catholicism, Protestantism and Islam. Graph 3 shows in the second factorial design, Trust and behaviours – Collective action, the location of the individual representative of each belief. The four main religions can be broken down as follows: Islam (34%, primarily ethnic groups in the northern part of IC and Burkina Faso), Catholicism (26%, primarily the Akan group, in the east close to Ghana's border, to which the Baoulés from the centre of the country belong), Protestantism (17%) and Animism (15%, primarily in the west, close to the border with Liberia).

{GRAPH 3}

It is worthwhile noting the distribution of the individuals representing each religion (shown in blue) along the first bisector of the design Trust and behaviours – Collective action, where the Animists are located in the north-east quadrant and Muslims in the south- east one. The others, Christians and those without any religion, are very close to the origin. Supposing that all the Animist live in Seïleu and the Muslims in Belleville and Dahiri, a socio- anthropologist could conceivably be sent to Seïleu then to Belleville and Dahiri and be instructed to conduct a survey for each community, the one Animist and the other Islam. The distributions are not so clear, unfortunately. The various pioneer fronts have been the

33

focus of migratory flow into villages originally inhabited by indigenous people with different customs and traditions, as clearly illustrated by the graph on the right! Consequently, the coordinates of individuals representing Seïleu, on the one hand, and the Animist group, on the other, are not combined in the space for social capital. The ratio of Animists is higher in Seïleu than in Belleville and vice versa in the case of Muslims but that is not all! Hence the importance of all other things being equal reasoning ...

5.2.3 Size of the plantation (ha), productivity (kg/ha) and social capital in IC

A significant stylized fact in family farming in the tropics is the inverse relationship between the size of the land farmed and the productivity of farm household. In average terms, the smaller the size of the land worked by a farm household the higher its output. The debate about this relationship and the reasons goes back several decades (Bardhan 1973, Eastwood et al. 2010). The two explanations offered up to now are: i) a more efficient use of inputs by smaller farms and ii) errors occurring when measuring the production factors and, in particular, the size and quality of the land being farmed. What does our study have to say on this subject? Is there a relationship between an actor's social capital and the size of the actor's largest cocoa plantation?

Towards this end, we divide the actors into four identically sized groups (quartiles) according to the size of their largest sized (hectarage) plantation (each actor operates fewer than two plantations, on average) and productive measured as kg/ha. The four groups are as follows: size of the plantation, less than 2 ha, between 2 and 3 ha, between 3 and 5 ha, 5 ha and more, and the plantation's productivity, less than 125 kg/ha, from 125 to 220 kg/ha, between 220 and 400 kg/ha and over 400 kg/ha. Graph 4 shows the perceptions of the actor representative of each subgroup projected onto the second factorial design.

{GRAPH 4}

The most optimistic trust-related perceptions are obviously those expressed by the smallest growers (shown in red in the graph on the right). Conversely, those inclined to express feelings of mistrust are growers in the bottom quartile (over 5 ha, in yellow). It should also be stressed that these quartiles are distributed (as with the religious groups)

34

along the first bisector of the second factorial design, Trust and behaviours – Collective action, where individuals representative of the group of large (small) growers are located in the south-west (north-east) quadrant. In the case of productivity, the least productive individuals apparently express some degree of mistrust towards members of their communities, whereas the most productive individuals apparently express having a feeling of trust.

Accordingly, a positive correlation emerges between our synthetic variable, reflecting the actors' having a feeling of trust in the other members of their communities and their agricultural productivity, and a negative correlation between the former and the size of the cocoa plantation being operated. A third variable therefore has to be introduced to take account of the negative farm size/productivity relationship.

5.2.4 Group membership, profit change and the social capital of Vietnamese retailers

The survey we conducted in 2013 focused on nine provinces in North and South VN featuring roughly 70% of women and 30% of men with nine out 10 being married. 1214 actors reported being members of an organisation compared with 1222 who do not belong to any organisation, or a 50-50 split. An actor will, on average, be a member of less than two organisations. The two most important organisations are the Women's Union and the Farmers' Association. It comes as no surprise to observe, for example, that 70% of a group's members are in Women's Union groups, which are established in almost all areas of Vietnam, be they rural or urban. This kind of trade union may have originally operated on a quasi-governmental basis for the purpose of devolving resources and power to the regions but their present-day activities are much more akin to those of an NGO with a microcredit function, for example. These trade unions are undoubtedly of key importance owing to their willingness to create what Granovetter (1973) calls weak ties (Poon et al. 2012). Their membership is made up of both women and men. Next are the members of a farmers association (roughly 25%). Communist party members account for less than 5% of the total.

It is interesting here to project onto the first factorial design, Social exclusions – Trust and behaviours, the coordinates of the average individual member of an organisation compared with those of an individual who does not belong to any organisation (see graph 5,

35

on the left). The coordinates of these two (hypothetical) actors are aligned along the second bisector of the factorial design. In other words, in terms of averages, actors who are members of an organisation express having a higher feeling of trust and see less social exclusion (north-west quadrant on the graph), compared with their peers who do not belong to any organisation (south-east quadrant).

{GRAPH 5}

We are now going to consider the trend in the profits of actors during the years prior to the survey. The level of profits may have increased, decreased or remained unchanged. Profits rose for 17% of actors, fell for 48% and therefore remained unchanged for 35%. These retailers had been pursuing their activities for 10 years on average (standard deviation 9 years). Projecting the modalities ‘increased’, ‘decreased’ and ‘unchanged’ onto the second factorial design shows they are aligned from right to left on the axis Collective action. Those claiming their profits increased are, on average, those who also reported having being involved over the past year in an information campaign, election campaign or in associations. The modality ‘increased’ is the one located the furthest to the right along the axis Collective action when the modality ‘unchanged’ is just on its left, still in the positive part of the axis. The modality ‘decreased’ is even further left but now in the negative part of the axis.

6. Economic performance and social capital of actors

Now that we have constructed a space for social capital in two activities representative of developing countries (family farming and small business) and geographies located on three continents (Sub-Saharan Africa, Asia and Pacific), it is time to consider the hypothesis that an actor possessing a large stock of social capital would be more or less economically and professionally successful compared with the average. This question raises a number of methodological problems, particularly in terms of identification (Durlauf 2002, Durlauf & Fafchamps 2005, Fafchamps 2006).

6.1. Models and methodology

36

What impact does an individual's social capital have on the outcome of that person's economic activity? Or Yi the economic outcome of actor i. In IC this refers to productivity

measured as kg/ha for the operator's largest cocoa plantation. This compares with VN, where we use a discrete variable showing whether the retailer's profit has increased, decreased or remained unchanged compared with the profit in the past year, and whether the retailer had access to formal and/or informal credit during the past three years. Our model is as follows:

Yi =α+Social_Exclusionβ1 +Trust&Behaviorsβ2 +Collective_Actionβ3 +WΟ(i)γ+Xiφ+εi

where WΟ(i) shows whether actor i is a member of one or more organisations and, if

the answer is yes, which one (s), Xi is a vector of socio-demographic characteristics

applicable to the actor such as the latter's age, gender, educational level, origin,..7. So we have three groups of variables: i) the actor's social capital measured by the individual's coordinates in the space derived from the MCA, ii) membership of an organisation or even more specific information about the type of organisation, iii) a set of information about the socio-demographic characteristics of the actor.

According to this specification, the social capital variables are regarded as exogenous. The model thus estimated by the ordinary least squares does not take account of any interactions there may be between actors within a community, between members of an organisation and between members of two organisations. Nor does our model consider the behavioural spillovers of peers an actor may interact with or otherwise. However, Manski (1993, 2000) and Durlauf (2002) have alerted us to these two potential biases that may result from omitting these effects on the endogenous variable of interest in our model. First of all, our study does not at this stage take account of the characteristics specific to groups of which the actor is a de facto member (e.g. family) or related to the actor's action (e.g. cooperative). For example, the educational level, the main occupation of the members of the actor's nuclear and/or extended family are likely to have an impact on the individual's economic outcome, directly via support (e.g. information about sound agricultural practices) or indirectly via the actors social capital coordinates, i.e. where the actor is located in the

7 It should be stressed that an actor's social capital changes during the individual's life cycle and the probability of migrating from one community to another. Accordingly, the actor's age and origin may be correlated with our social capital variables.

37

space for social capital. Secondly, in addition to the characteristics of a group's members, their behaviours (e.g. agricultural practices, compliance with norms,...) may affect the actions and efforts made and, consequently, an actor's economic outcomes. Bearing in mind, the limitations of our model and these potential endogeneity biases, we would now like to consider some of the preliminary findings for IC and VN.

6.2 Agricultural practices, agricultural productivity and social capital in IC

It has to be remembered that the data gathered in IC covers roughly 2020 actors established in seven villages located from east to west in the various cocoa belts that have seen the arrival of significant waves of internal (e.g. Baoulés from the centre) and external (mainly Burkina Faso and Mali) migration in the light of the labour required for cocoa farming, leading to the exhaustion of the "forest rent"(forest resources) immediately afterwards (see graph 6). Consequently, one-third of the actors are not Ivorian nationals (important in view of the 1998 Law on rural land tenure) and only half were born in the village where they pursue their activities. More specifically: 38% are indigenous, 27% non-indigenous and 35% are aliens. Amongst these, 76% (1527) have at least one cocoa plantation. Roughly one-quarter of them harvested crops in more than one plantation during the previous season and one-third have one or more plantations that have not yet begun to produce. It should also be noted that 47% are illiterate and less than half of the others are educated beyond primary school level.

38

6.2.1 Horizontal networks, weak versus strong ties and individualism

In addition to the three synthetic variables for social capital (Social exclusion, Trust & Behaviour and Collective action), we add a variable to record an actor's membership of an organisation. 80% answered ‘Yes’ to the question “Are you a member of any group or association?" and, on average, one actor is a member of less than two organisations at the same time (see table 4). By way of comparison, 90% and 83% reported being a member of one organisation in TZA and PNG respectively. This relatively high figure does not clearly reflect the concerns Putnam (1995) expresses about the disengagement of Americans from involvement in groups or organisations, particularly in the religious sphere. However, this should not foster the myth of an "Africa of villages and a continent of the collective "(Oliver de Sardan 1995, who adds, on page 61: "The systematic exploitation of the "village collective "level is highly ideological to some extent (derived from) two Western traditions, the Christian movement and Socialist movement."). It is nonetheless a reflection of a social life where social networks and the ties between actors and their interactive areas have to produce institutional arrangements for replacing and/or complementing or otherwise competing with the formal regulations created by the modern state.

{TABLE 4}

In view of the status of religions in the space for social capital in IC (see above), it is worthwhile noting the importance of religious organisations in the social life of households in Lower Ivory Coast. Family associations or diasporas apparently play a key role, in terms of solidarity and reciprocity. Particularly in the light of the relative position of organisations with a primarily economic objective, such as cooperatives or economic interest groupings (EIG) representing small farmers. Consequently, the "strong ties "forged within a nuclear and extended family or an ethnic or cultural group are apparently relatively more significant than the "weak ties" that may be established between actors. Granovetter (1973, 1895) and Putnam et al. (1993) agree that overlapping weak ties in several organisations are more significant than intense

39

personal ties for maintaining the stability of social capital and collective action (see table 5).

The level of interactions between actors is therefore relatively high and in the same way as the temptation of the consensual village "community" myth should be resisted, small farmers should not be regarded as individual, independent and rational "small business persons" maximising their profits and seeking efficiency based solely on purely functional and abstract criteria (Olivier de Sardan 1995). In the light of our quantitative approach to social capital and its connection with an actor's economic performance, it is worthwhile weighing these comments borrowed from an anthropologist from Africa against the methodological question discussed in the previous section and investigated by Manski (1993, 2000 – where he decided not to mention social capital), on the one hand, and applied to social capital by Durlauf (2002), on the other.

{TABLE 5}

This distinction between weak ties and strong ties and their individual and collective relevance is also referred to by Putnam (2000), when he points to the difference between "bonding and bridging social capital ", with the first type granting priority to network density and akin to strong ties, while the second one favours a network's scope. Putnam claims "bridging social capital "creates social cohesion and is the most conducive to collective action. At this stage it is understandable that the statistics featured in table 4 are much too aggregated to teach us very much about this subject. That is why it is important to combine quantitative and qualitative approaches and apply an even more micro-level analysis. An analysis of organisations within each village, their social and economic activities and their cultural and political roles, of their members, too, and their socio-demographic characteristics, and of their method of decision-making8.

8 See also Burt 1992 (discussed in Sobel 2002) who focuses on individuals making the connection between two networks, which, without this, would be completely disconnected.

40

6.2.2 Individual social capital and productivity

The model we believe is very similar to the one developed by Haddad & Maluccio (2003) as it simultaneously adds to the group of explanatory variables information about membership of a group or groups, on the one hand, and our synthetic variables reflecting the social capital of actors (Social Exclusion, Trust and Behaviours and Collective action), on the other. The estimation of the model, whose outcomes are featured in table 6, mainly shows two things. First of all, solely the coordinates of the axis Trust & Behaviours in the space for social capital correlate positively with productivity. The coordinates of actors in the factorial design Social Exclusion – Collective action do not apparently affect their productivity. Secondly, being a member of an organisation is also associated positively with productivity (see models (2), (3) and (4) in table 5). A consideration of a household's two most important groups (see model (5)) shows that being a member of a farm cooperative (cooperative variables 1 and 2) clearly has a positive and statistically solid impact on an actors' productivity. As a function of whether the cooperative also represents the most important organisation for the household, note should be taken of the relative significance of the family or diaspora association as the second most important organisation for the household. This does not work the other way. The productivity of an actor claiming the most important organisation for the individual's household is the family or diaspora association is not any higher than that of an actor who does not belong to any association, unless the second most important organisation for the actor's household is a farm cooperative. Accordingly, groups that may be described as market-oriented, on the one hand, and community-oriented, on the other, (see Bernard et al. 2008) are apparently complements more than substitutes. Lastly, being a member of a religious group has a negative or zero impact on productivity depending on whether this organisation is the first (negative) or second (zero) most important organisation for the household.

{TABLE 6}

These correlations are robust against differences between actors in terms of rights of administration on the farmed plantation, expropriation risks and the fact of being born in the village. Reflecting an actor's land tenure security, these three variables show three different things. First of all, an actor claiming to have rights of administration (or

41

transfer) rather than just the right to operate a plantation (a right the actor may not necessarily possess but it is important for the individual to believe this!) is more productive than the actor claiming the opposite. Secondly, those saying they have no rights of administration and, moreover, replied they ran a high risk of being expropriated or evicted from their plantations are more productive than the others, in terms of averages. In order to understand this finding, which at first sight seems to be contradictory, we have to bear in mind that the more land is farmed ostensibly, the lower the risk of being evicted (see, for example, Udry 2012)9. Thirdly, being born in the village also has a positive impact on productivity irrespective of the actor's social capital, even if being born in the village represents social capital in itself.

Two aspects concerning the plantation more than the actor personally should also be noted. First of all, during its life cycle the cocoa tree's level of productivity rises until reaching maturity (an optimum), after which it tends to become less productive. In other words, the tree's productivity has a bell-shaped curve. The previous outcomes are valid whatever the age of the plantation and the life cycle of its trees. Lastly, it is interesting here to see a confirmation of the aforementioned stylized fact as applicable to the farm size-productivity relationship. Small plantations have the best outcomes.

We believes that these findings are quite exciting, irrespective of the huge variety of social ties, which a lack of relevant information prevents our study from specifying. First of all, they demonstrate something happening in cooperatives tending to favour the economic performance of an actor within their ranks (e.g. exchanges of information, access to inputs). The performance is even higher when an actor is also a member of a family association or diaspora (e.g. access to credit). Secondly, the trust actors place in their peers and compliance with values such as individual responsibility, altruism and the solidarity actors have with their peers tend to make the actors more productive, irrespective of the membership of one or more organisations. The outcome is interesting in the sense that the body of empirical literature on social capital often uses the participation of a community's actors in group activities to measure the climate of trust within that community. At individual level, it is also proposed that the intensity of an

9 Félix Houphouët-Boigny, the President of Ivory Coast from its independence in 1960 until 1993, states that the "land belonged to the one farming it."' He encouraged migratory flows, seeking to promote the perennial crops cocoa and coffee in the 1970s and 1980s. Rubber tree production is now taking over. Almost half of the cocoa growers surveyed (n=1527) are cultivating another perennial crop. Rubber tree production (40% of growers) follows just after coffee (60%).

42

actor's interactions with other members of the community, e.g., within organisations, affects the development of the feeling of trust in the actor's peers and the actor's beliefs about the behaviour of these peers being more or less compliant with the social norms.

The direction of causality, if it exists, has, admittedly, not been established in statistical terms but actors are likely to become members of a cooperative to improve their performance rather than the reverse10. What would be the point of actors who are amongst the most productive members of their community joining a cooperative, whose members are less productive and might not welcome these better performers? Individual success (within the village itself, i.e. outside "frameworks" in IC) is not prized nor much appreciated in traditional African societies, as this is regarded as a threat to social cohesion and the traditional system of solidarity. That is why young people in villages drift away from farming as a main occupation and have for a long time tended to move to an urban environment if they aspire to a certain degree of social mobility (Durlauf 1992). This kind of conservatism is even hindering the creation of organisations whose aim is the economic development of its members (Bernard et al. 2010 in Senegal and Burkina Faso).

6.3 Retailers’ profits, credit access and social capital in VN

A total of 2465 questionnaires were collected from micro entrepreneurs in nine Vietnamese provinces from north to south11. In each province, nine municipalities were selected at random: six in a rural area and three in an urban one. Just over 250 micro and small entrepreneurs were surveyed in each province (graph 7).

The socio-demographic characteristics of the actors are as follows: 71% are women, 96% belong to the Kinh group, 90% are married and 74% do not have a religion. An actor's social capital changes during the individual's life cycle. Accordingly, the actors are broken down by age group: 48% are aged between 21 and 40, 46% between 41 and 60, while only 26 of them are under 20 and 112 are over 60. Lastly, the educational level tends to be higher because 38% of them completed secondary school and 37% completed high school. In the

10 Here it should be noted that a plantation's output is visible and is easy for the other growers to see but less so in the case of a grower's income or expenditure or that of the grower's household (the endogenous variable most frequently cited in this literature)!
11 Five provinces were surveyed in the north: Vinh Phuc and Bac Ninh nieghbouring Hanoi then Thai Binh and NGhe An towards the south. The four provinces chosen in the south are, moving southwards: Lam Dong, Ba Ria-Vung Tau, then Tra Vinh and Soc Trang.

43

case of IC it has already been mentioned that the mobility of actors is a key parameter for an actor's social capital. In the case of this sample, roughly 50% of them have parents from outside the province where they are pursuing their entrepreneurial activities. Lastly, as a guide, 35% (11%) replied ‘Totally satisfied’ (‘totally-unsatisfied’) to the question: “How satisfied are you today in terms of overall job satisfaction?" The others replied ‘neutral’.

{GRAPH 7}

Below we will be investigating whether there is any relationship between our social capital variables and the subjective satisfaction of actors, on the one hand, and their economic performance and access to credit, on the other. As with Ivory Coast, we single our three groups of explanatory variables. First of all, the socio-demographic variables, including age, occupational experience (number of years in the community of residence) and if an actor is an insider or more of an outsider (the actor's father is from the community where the individual lives). We also control for region (north or south) and community of residence and professional occupation according to the rural or urban aspect. Secondly, in addition to three social capital variables, we include variables characterising the size of the business (annual income), membership or otherwise of a group, and the experience the actor has acquired over the years through the individual's business activities.

Making a distinction on the basis of whether the actor lives in the north or south of the country is justified by a clear divide between the five provinces in the north and the four in the south.

6.3.1 Divide between the provinces in the north and those in the south

Bear in mind that the most important axis is the Social Exclusion axis (44%), where, according to what the actors say, the most significant divisive issues are the differences in ‘social status’ and ‘income and material wealth’12. This is followed by the Collective Action axis (24%), where questions about actors' involvement in collective actions contribute by

12 (The modality ‘Not at all’ was selected by roughly 80% of actors, while the modality ‘very much’ was mentioned by less than 5% of them. Consequently, the modalities ‘not at all’ in the negative part of the axis, on the one hand, and the modality ‘somewhat’ located to the right of the axis, on the other, are the ones that contribute to the construction of this synthetic variable we have called Social Exclusion.

44

83% to the construction of this axis. The Trust & Behaviours axis contributes by only 11% to the construction of the space for social capital. In common with other countries, there is a specific geographical characteristic, such that the provinces and their representative actors are completely distinct within this space (graph 8).

First of all, the graph on the left shows the differences according to social status and income as divisive issues are apparently more telling, on average, in the southern three provinces (Ba Ria-Vung Tau, Soc Trang and Tra Vinh located in the positive part of the x-axis) than in the north (Bac Ninh, Hanoi, NGhe An, Thai Binh and Vin Phuc to the left on the same axis). Secondly, actors in the provinces of NGhe An and Vin Phuc in the north (to the right on the x-axis of the graph on the right) are apparently more ready for collective action than their southern neighbours in the provinces of Ba Ria-Vung Tau and Soc Trang (to the left on the same axis). Lastly, the climate of trust is, on average, better in the north (Bac Ninh, Hanoi, Thai Binh and Vin Phuc located in the positive part of the y-axis of graph on the right) than in the south (Ba Ria-Vung Tau, Soc Trang and Tra Vinh located in the negative part of the same axis).

Is this divide between the provinces in the north of the country and those in the south reflected in their economic performance? In the case of per capita income per day in 2011 and denominated in US Dollars, we can distinguish between three categories including: $1.5- 2 (Nghe An in the north and Tra Vinh and Soc Trang in the south; $2-2.5 (Vin Phuc and Thaï Binh in the north and Lam Dong in the south); > $2.5 (Bac Ninh and Hanoi in the north and Ba Ria in the south). As for the three poorest provinces, Soc Trang and Tra Vinh in the south are the provinces where representative individuals experience the most divisions within their communities, while it is the opposite in the Nghe An province in the north (see axis 1). Moreover, these three provinces, whose per capita income per day is the lowest of the nine provinces, are also characterised by the strong feeling of mistrust their inhabitants have towards the other members of their communities (see Axis 3). This finding is a repeat of the one shown above for farming households in Ivory Coast, where trust is also associated positively with their level of agricultural productivity.

{GRAPH 8}

45

6.3.2 Levels of and trend in profits and social capital

It should be recalled that 17% (48%) of actors saw their year-on-year profits increase (decrease). (so this remained the same for 35% of actors.) We consider if the probability of a profit increasing or, conversely, decreasing is linked to age or to origin, depending on whether an actor is pursuing the activity in the father's community of origin. We are interested first and foremost in the social capital variables as described above. We look to see if there is a relationship between these variables and the level of profits (quartiles). Towards this end, we estimate two Logit multinomial models whose findings are shown in table 7.

The probability of having a high profit level (quartiles 2, 3 and 4 relative to the 1st quartile) is higher in the northern provinces and in urban areas. Retailers seen to enjoy a high level of profits are reported to be better educated (high school and university) and more experienced. Pursuing an activity in the father's community of origin is also preferable. The size of the business measured according to annual income also has a positive impact on the profit level. Consequently, a poorly qualified, inexperienced small retailer living in a southern province and located in a rural area that the father is not a native of will, on average, make only small profits. As for the retailer's stock of social capital, those who feel less excluded are not ranked in the top profit quartile. What is more, the variable synthesising the actor's level of involvement in collective actions has an impact on the change in the actor's profit. The more an actor is involved in collective actions, the more likely the actor's profit is to have increased during the previous year. The same applies for an actor who is a member of a group.

{TABLE 7}

On the basis of graph 9, let us take a look, for example, at the marginal probability of a profit increasing, decreasing or remaining unchanged during the past year according to the level of investment in collective actions (north-west quadrant) or the marginal probability of belonging to one of the profit quartiles according to one's feelings in terms of social cohesion (reversed axis 1 and north-east quadrant), professional experience (south-west quadrant) or age (south-east quadrant).

46

Let us now examine the graph for the north-west quadrant. The closer an actor is to origin, the more the individual is inclined to invest heavily in collective action. Clearly, then, if a small effort is made, the profit downward scenario during the previous year (blue line) is a lot more likely than then profit upward scenario (red line). The probabilities show that the likelihood of the first (second) scenario decreases (increases) in line with the collective actions undertaken by the entrepreneur. Let us now move to the south-west quadrant. If the probabilities of being in the top (red line) or bottom quartile (violet line) are not significantly different for two inexperienced actors, the probability of being in the top (bottom) quartile decreases (increases) in line with the experience the retailer builds up. However, it would be wrong to measure the actor's experience by age (see red and violet lines respectively in the graph for the south-east quadrant).

47

{GRAPH 9}

6.3.3 Credit access both formal and informal and social capital

We will now consider whether an individual's social capital is important for obtaining access to formal credit, on the one hand, and informal credit, on the other. In the case of formal credit, the following question was put to the actors, “Have you applied (if yes, received) for bank loans or other formal credit over the last three years?”, where three modalities could be chosen: not applied, applied and got refused, applied and received. Just over two-thirds of the actors in our sample have not sought any formal credit from a financial institution. Almost all of the remaining one-third have received credit. Is this a self-selection process by the actors? What is the situation with social capital?

Are these differences resilient to the introduction of control variables such as the actor's experience or annual income? Similarly, it is reasonable to assume that access to credit differs between the northern and southern provinces. Table 8 features the findings of a logit multinomial applied to access to formal and informal credit.

{TABLE 8}

Several items may be observed. First of all, a candidate entrepreneur has easier access to formal credit as a member of a group. Secondly, access is all the more likely when the individual is involved in collective actions. Conversely, the actor will express a strong feeling of being excluded by other members of the community if the individual's request for credit is turned down. Thirdly, one-quarters of the entrepreneurs surveyed availed themselves of informal credit during the three years preceding the survey. In average terms, they are apparently more involved in collective actions than those that did not have recourse or access to credit. Lastly, in keeping with the aforementioned North-South divide, the probability of seeking informal credit is relatively higher in the northern provinces. (see also graph 10.)

{GRAPH 10}

48

6.3.4 General and job satisfaction and social capital

The two previous sections took a look at objective output measurements, as with the level of profit and access to credit. This section will consider the situation for subjective satisfaction, i.e. as felt by the actor.

The general level of satisfaction is, on average, higher in the south, while the level of job dissatisfaction is lower. The same conclusion may be drawn for the level of trust in terms of general satisfaction or, conversely, the level of mistrust, in terms of job satisfaction. The probability of being (dis) satisfied even approaches (zero) one in the case of community actors expressing a strong feeling of trust (see graph 11 on the left).

{TABLE 9} {GRAPH 11}

49

7. Conclusions, limitations and perspectives

What is most important about the MCA undertaken for the different countries covered by this meta-analysis is the value of the construction of a three-dimensional space for social capital where actors operate that may be located via a three-coordinate vector in this space. The value is primarily found in the parsimony of the social capital variables selected, on the one hand, and the sum of independent information they have, on the other. Two actors close to each other in this space are two actors who replied to questions in a similar way. Accordingly, we may conclude that these two actors have identical experiences of / have the same feeling about: i) social exclusion within their community, ii) the trust they may have in their peers and their behaviours more or less in keeping with the norms and iii) their ability to take action in favour of a shared goal. Each actor who originally replied to several dozen questions is now identified by a three-coordinate vector.

Finally, the analysis shows that only 20% of the original information was lost during the operation. When the importance of degrees of freedom in regression exercises is realised, the relevance of a MCA in this study is easier to understand. The space also includes the coordinates of the replies to the questions they have been asked. Consequently, the proximity of an actor and a modality in this space is a reflection of the answer the actor gives to the question associated with the modality, so it is possible to make an assessment in terms of the distribution of social capital within a population in this space of limited size.

50

8. Bibliography

Acemoglu, D., S. Johnson and J.A. Robinson (2001) “The colonial origins of comparative development: An empirical investigation." American Economic Review, 100: 115-31.

Acemoglu, D., S. Johnson and J.A. Robinson (2005) “Institutions as a fundamental cause of long-run growth." In Handbook of Economic Growth Vol. 1A, ed. P. Aghion and S.N. Durlauf, Elsevier B.V.

Acemoglu, D. and J.A. Robinson (2012) Why nations fail? The origins of power, prosperity and poverty. Profile Books LTD.

Banerjee, A. and E. Duflo (2011) Poor economics: A radical rethinking of the way to fight global poverty. PublicAffairs.

Bardhan, P. (1973) “Size, productivity and returns to scale: An analysis of farm-level data in Indian agriculture." Journal of Political Economy, 81: 1370-86.

Bardhan, P. (2010) “Institutional economics of development: Some general reflections." In Institutional microeconomics of development, ed. T. Besley and R. Jayaraman, MIT Press.

Bardhan, P. and I. Ray (2006) “Methodological approaches in economics and anthropology." Q-squared Working Paper No. 17.

Bauernschuster, S., O. Falck and S. Heblich (2010) “Social capital access and entrepreneurship." Journal of Economic Behavior and Organization, 76: 821-33.

Bernard, T., M-H. Collion, A. De Janvry, P. Rondot and E. Sadoulet (2008) “Do village organizations make a difference in African rural development? A study for Senegal and Burkina Faso." World Development, 36: 2188-204.

Bernard, T., A. De Janvry and E. Sadoulet (2010) “When does conservatism constrain village organizations?" Economic Development and Cultural Change, : 609-41.

Berry, S., (2001) Chiefs know their boundaries: Essays on property, power and the past in Asante, 1896–1996. London: Heinemann.

Besley, T. (1995) “Property rights and investment incentives: Theory and evidence from Ghana." Journal of Political Economy, 103: 903-37.

Besley, T. and R. Jayaraman (2010) “Introduction." In Institutional Microeconomics of Development, ed. T. Besley and R. Jayaraman, MIT Press.

Bourdieu, P. (1984) The distinction. Boston: Harvard University Press.
Bourdieu, P. (1986) “Forms of capital." In
Handbook of Theory and Research for the Sociology

of Education, ed. J.G. Richardson. NY: Greenwood Press.
Bowles, S. and H. Gintis (2002) “Social capital and community governance."
Economic

Journal, 112: F419-36.
Burt, R.S. (1992)
Structural holes: The social structure of competition. Cambridge: Harvard

University Press.

Chauveau, J-P. and J-P. Colin (2010) “Customary transfers and land sales in Côte d’Ivoire: Revisiting the embeddedness issue." Africa, 80: 81-103.

51

Clastres, P. (1976) La question du pouvoir dans les sociétés primitives. Interrogations, 7: 3-8.

Cling, J-P., S. Lagrée, M. Razafindrakoto and F. Roubaud (2012) « Un enjeu majeur de développement : améliorer la connaissance de l’économie informelle pour mettre en œuvre des politiques adaptées. » In L’économie informelle dans les pays en développement, Cling et al. (dir.). Agence Française du Développement (AFD).

Coleman, J.S. (1988) “Social capital in the creation of social capital." American Journal of Sociology, 94: S95-S120.

Dasgupta, P. and I. Serageldin (2000) Social capital: A multifaceted perspective. Washington, DC: World Bank.

Dixit, A.K. (2004) Lawlessness and Economics: Alternative modes of governance. Princeton, NJ: Princeton University Press.

Durlauf, S.N. (2002) “On the empirics of social capital." Economic Journal, 112: F459-79. Durlauf, S.N. and M. Fafchamps (2005) “Social capital." In Handbook of Economic Growth,

Vol. 1A, ed. P. Aghion and S.N. Durlauf, North Holland: Elsevier.
Durlauf, S.N., P.A. Johnson and J.R.W. Temple (2005) “Growth Econometrics." In
Handbook

of Economic Growth, Vol. 1A, ed. P. Aghion and S.N. Durlauf, North Holland: Elsevier. Eastwood, R., M. Lipton and A. Newell (2010) “Farm size." In Handbook of Agricultural

Economics, ed. P.L. Pingali and R.E. Evenson. North Holland: Elsevier.

Engerman, S.L. and K.L. Sokoloff (2006) “Colonialism, inequality, and long-run paths of development." In Poverty traps, ed. S. Bowles, S.N. Durlauf and K. Hoff. NJ: Princeton University Press.

Escofier, B. and J. Pagès (2008) Analyses factorielles simples et multiples : objectifs, méthodes et interprétation. 4ème édition, Dunod.

Fafchamps, M. (1992) “Solidarity networks in preindustrial societies: Rational peasants with a moral economy." Economic Development and Cultural Change, 41: 147-74.

Fafchamps, M. (2006) “Development and social capital." Journal of Development Studies, 42: 1180-98.

Fafchamps, M. and B. Minten (2002) “Returns to social network capital among traders." Oxford Economic Papers, 54: 173-206.

FA0 (2014) The state of food and agriculture 2014. Innovation in family farming. Food and Agricultural Organization of the United Nations, Rome.

Fenske, J. (2011) “Land tenure and investment incentives: Evidence from West Africa." Journal of Development Economics, 95: 137-56.

Fukuyama, F. (2000) “Social capital and civil society." IMF Working Paper 74.
Gallup, J.L., J.D. Sachs and A.D. Mellinger (1999) “Geography and economic development."

International Regional Science Review, 22: 179-232.
Glaeser, E.L., D. Laibson and B. Sacerdote (2002) “An economic approach to social capital."

Economic Journal, 112: F437-58.
Godelier, M. (2007)
Au fondement des sociétés humaines. Albin Michel.

52

Gourou, P. (1991) L’Afrique tropicale, nain ou géant agricole ? Flammarion, Paris. Granovetter, M.S. (1973) “The strength of weak ties." American Journal of Sociology, 78:

1360-80.
Granovetter, M.S. (1985) “Economic action and social structure: The problem of

embeddedness." American Journal of Sociology, 91: 481-510.

Greif, A. (1994) “Cultural beliefs and the organization of society: A historical and theoretical reflection on collectivist and individualistic societies." Journal of Political Economy, 102: 912- 50.

Grootaert, C. and T. Van Bastelaer (2002) Understanding and measuring social capital: A multidisciplinary tool for practitioners. Washington, DC: World Bank.

Grootaert, C., D. Narayan, V. Nyhan Jones and M. Woolcock (2004) “Measuring social capital: An integrated questionnaire." World Bank Working Paper No. 18. Washington DC: World Bank.

Guha-Khasnobis, B., R. Khambur and E. Ostrom (2006) “Beyond formality and informality." In Linking the formal and informal economy: Concepts and policies, ed. B. Guha-Khasnobis, R. Khambur and E. Ostrom. Oxford: Oxford University Press

Haddad, L. and J.A. Maluccio (2003) “Trust, membership in groups, and household welfare: Evidence from KwaZulu-Natal, South Africa." Economic Development and Cultural Change, 51: 573-601.

Hodgson, G.M. (2006) “What are institutions?" Journal of Economic Issues, 40: 1-25.
Knack, S. and P. Keefer (1997) “Does social capital have an economic payoff? A cross-country

investigation." Quarterly Journal of Economics, 112: 1251-88.
Le Roux, B. and H. Rouanet (2004)
Geometric data analysis: From correspondence analysis to

structured data analysis. Kluwer Academic Publisher.
Le Roux, B., H. Rouanet, M. Savage and A. Warde (2008) “Class and cultural division in the

UK." Sociology, 42: 1049-71.
Lyons, M. and S. Snoxell (2005) “Creating urban social capital: Some evidence from informal

traders in Nairobi." Urban Studies, 42: 1077-97.
Maluccio, J.A., L. Haddad and J. May (2000) “Social capital and household welfare in South

Africa 1993-1998."Journal of Development Studies, 36: 54-81.
Manski, C. (1993) “Identification of endogenous social effects: the reflection problem."

Review of Economic Studies, 60: 531-42.
Manski, C (2000) “Economic analysis of social interactions."
Journal of Economic

Perspectives, 14: 115-36.

Mauss, M. (1923-24) “Essai sur le don. Forme et raison de l’échange dans les sociétés archaïques." L’année sociologique, Tome 1 : 30-186. (The gift: The form and reason for exchange in archaic societies. London: Routledge, 1990.)

Meagher, K. (2005) “Social capital or analytical liability? Social networks and African informal economies." Global Networks, 5: 217-38.

Meagher, K. (2006) “Social capital, social liabilities, and political capital: Social networks and 53

informal manufacturing in Nigeria." African Affairs, 105: 553-82.
Munshi, K. (2014) “Community networks and the process of development."
Journal of

Economic Perspectives, 28: 49-76.
Narayan, D. and L. Pritchett (1999) “Cents and sociability: Household income and social

capital in rural Tanzania." Economic Development and Cultural Change, 47: 871-97.
North, D.C. (1990)
Institutions, institutional change and economic performance. New York:

Cambridge University Press.

North, D.C. (1994) “Economic performance through time." American Economic Review, 84: 359-68.

Olivier de Sardan, J-P. (1995) Anthropologie et développement : Essai en socio-anthropologie du changement social. Paris : Karthala.

Ostrom, E. (2000) “Collective action and the evolution of social norms." Journal of Economic Perspectives, 14: 137-58.

Ostrom, O. and T.K. Ahn (2003) “Introduction." In Foundations of Social Capital, ed. E. Ostrom and T.K. Ahn. Edward Elgar Publishing Limited.

Place, F. (2009) “Land tenure and agricultural productivity in Africa: A comparative analysis of the economics literature and recent policy strategies and reforms." World Development, 37: 1326-36.

Polanyi, K. (1983) La grande transformation. Aux origines politiques et économiques de notre temps. Paris : Gallimard (1ère éd. anglaise, 1944).

Poon, J.P.H., T.T. Thai and D. Naybor (2012) “Social capital and female entrepreneurship in rural regions: Evidence from Vietnam."Applied Geography, 35: 308-15.

Portes, A. (1998) “Social capital: Its origins and applications in Modern Sociology." Annual Review of Sociology, 24: 1-24.

Putnam, R.D. (1995) “Bowling alone: America’s declining social capital." Journal of Democracy, 6: 65-78.

Putnam, R.D. (2000) Bowling alone: The collapse and revival of American community. New York: Simon and Schuster.

Putnam R.D., R. Leonardi and R.Y. Nanetti (1993) “Social capital and institutional success." In Making Democracy Work: Civic Traditions in Modern Italy, Chapter 6. NJ: Princeton University Press.

Roland, G. (2014) Development Economics. Prentice Hall.
Rosenstein-Rodan, P. (1943) “Problems of industrialization of Eastern and Southeastern

Europe." Economic Journal, 53: 202-11.
Sachs, J.D. (2006)
The end of poverty. Penguin Books.

Schlager, E. and E. Ostrom (1992) “Property rights regimes and natural resources: A conceptual analysis." Land Economics, 68: 249-62.

Sobel, J. (2002) “Can we trust social capital?" Journal of Economic literature, XL: 139-54. Totolo, E. (2013) “Social networks and institutional change among small enterprises in

54

Nairobi." In Securing livelihoods: Informal economy practices and institutions, ed. I. Hillenkamp, F. Lapeyre and A. Lemaître. Oxford: Oxford University Press.

Udry, C. (2012) “Land Tenure." In The Oxford Companion to the Economics of Africa, ed. E. Aryeetey, S. Devarajan, R. Kanbur, and L. Kasekende. NY: Oxford University Press.

World Bank (2007) World Bank Development Report 2008. Agriculture for Development, Washington DC: The World Bank.

Wydick, B. (1999) “Can social cohesion be harnessed to repair market failures? Evidence from group lending in Guatemala?" Economic Journal, 109: 151-208.

9. Tables

Table 1: Empirical studies on social capital in the tropics

Study

Actors

Outcomes

SC measures

Findings

Narayan and Pritchett (1999)

Households in rural Tanzania

Per capita household expenditure

SC indices constructed for both households and villages. Indices based on memberships in groups, characteristics of the groups, and household values and attitudes.

Village SC dominates individual SC

Maluccio et al. (2001)

Households in KwaZulu-Natal, South Africa

Per capita total expenditure

Index of individual memberships in groups, reflecting number, gender heterogeneity, and performance, based on survey responses. Community social capital levels computed as aggregates of individual indices.

Individual and community social capital measures statistically significantly associated with expenditure.

Fafchamps and Minten (2002)

Food traders in Madagascar

Value added and total sales

Number of traders known, number of relatives in agricultural trade, number of potential informal traders.

Number of traders known and number of potential informal traders statistically significant.

Source: Durlauf & Fafchamps 1995, pp. 1674-75.

55