|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
LINKING FARMERS TO MARKETS: ASSESSING PLANNED CHANGE INITIATIVES TO
IMPROVE THE MARKETING PERFORMANCE OF SMALLHOLDER FARMER GROUPS
IN NORTHERN TANZANIA
JAMES G. BARHAM
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
James G. Barham
To my jamnii: the Barhams, Kellys, and Byrons
TABLE OF CONTENTS
LIST OF TABLES .........._.... ...............7.._.._ ......
LIST OF FIGURES .............. ...............9.....
AB S TRAC T ............._. .......... ..............._ 10...
1 INTRODUCTION ................. ...............12.......... ......
Research Objectives............... ...............1
Structure of the Dissertation ................ ...............14................
Literature Review ................. .. .. ..... ...... .. .......1
Economic Policy and Perspectives in Sub-Saharan Africa ................ ............. .......15
Theoretical Considerations and Hypotheses .............. ...............28....
New institutional economics .............. ...............29....
Feminist economics............... ...............3
Social capital .............. ...............39....
Collective action............... ...............43.
Cultural materialism ................. ...............46.................
2 PROGRAM AND STUDY AREA BACKGROUND .............. ...............51....
Country Background............... .. ..................... .......5
Agricultural Marketing Systems Development Program (AMSDP) ................ ................. 53
The AMSDP Stakeholders ............ ..... ._ ...............55...
Research Sites and Partner Agencies.................. ... .. .... ..............5
Traditional Irrigation and Environmental Development Organization (TIP) ................6 1
FAIDA MaLi (FAIDA) .............. ...............62....
Development Approaches .............. ...............62....
Group Selection Criteria ................. ...............63........ ....
Selected Farmer Groups ................. .... ....... .. .. ... ... ........6
Demographic Characteristics and Farming Systems of Farmer Group Members .........._......67
The PA Training Activities for Farmer Groups............... ...............72.
Study Entry Point............... ...............73.
3 RESEARCH METHOD S ................. ...............76....... ....
Sustainable Livelihoods Approach .....__................. ........_ ......... 7
Conceptual Model for Study............... ...............81.
Research Design .............. ...............82....
Data Collection Methods .............. ...............83....
Study Time Frame .............. ...............84....
Sample Groups............... ...............85.
Dependent Variable: Marketing Performance Rating .............. ...............86....
Explanatory Variables .............. ...............87....
Group Asset Variables............... ...............8
Group Composition Variables ............._. ...._... ...............93....
Group Characteristics Variables............... ...............9
Group Heterogeneity Variables ............._. ...._... ...............95....
Market Access Variables ........._.._... ......___ ...............97....
Matching Research Obj ectives with Hypotheses ................. ........._.__...... 98__._...
Objective 1: Farmer Group Assets .............. ...............99....
Obj ective 2: Farmer Group Composition ....._.__._ ..... ..._. ......._...........0
Obj ective 3: Farmer Group Characteristics ................. ...............100........... ..
Obj ective 4: Market Access ........._.___..... .___ ...............101..
Objective 5: Partner Agency. .......__..........._......._. ....__.. ...............101
D ata A nalysis............... ....... .... .. ..........0
Reliability and Validity of Measurements and Results ................. ................. ........ 102
4 FINDINGS: PART I ................. ...............107........... ...
M marketing Obstacles ................ ................ ...............107......
Gender Specific Marketing Obstacles .................. .. ... ..... .......... .............1
Matching Marketing Obstacles to Livelihood Asset Constraints ................. ...............112
Group Marketing Performance Results ................. ...............114...............
Underlying Factors to Group Marketing Performance ................. ............................114
Group Assets ................. ...............117......... ......
Group Composition ................. ...............119......... ......
Group Characteristics ................. ...............123......... ......
Group Heterogeneity .............. ...............124....
M market Access .............. ...............125....
W health Revisited .............. ...............128....
Multiple Regressions Analysis .............. ...............132....
Conclusion ................ ...............138................
5 FINDINGS: PART II ................. ...............139........... ...
The Effects of Partner Agencies on Group Marketing Performance ................. ................ 139
M marketing Strategies ................. ....... ...... ...............144......
Risk Assessment of Marketing Strategies ................. ......................... ............150
Partner Agency Linkages and Group Initiatives............... ..............15
Marketing Strategies and Explanatory Variables .............. ...............157....
Conclusion ................ ...............166................
6 SUMMARY OF FINDINGS AND STUDY CONCLUSIONS ................. ............... .....168
Summary of Findings: Hypotheses Revisited............... ...............16
Objective 1: Farmer Group Assets .............. ...............168....
Obj ective 2: Farmer Group Composition ................ ...............171..............
Obj ective 3: Farmer Group Characteristics ................. ...............173........... ..
Obj ective 4: Market Access .........__.. ..... ._ __ ...............174..
Objective 5: Partner Agency ........._._.. ....__. ....._. ....._.. .........._....174
Study Conclusions: Theory Revisited .............. ...............178....
7 RECOMMENDATIONS ........._.___......___ ...............185....
Poor Smallholder Farmers and Market-Oriented Interventions .............. ......................8
Programmatic Recommendations .............. ....._ ...............186...
Stages of Agro-Enterprise Development ................. ...............187...............
Issues of Stakeholder Participation .............. ...............194....
Research Recommendations ................. ...............198................
Final Remarks ................. ...............201................
A SELF-ADMINISTERED QUESTIONNAIRES .............. ...............203....
Questionnaire on Trust .............. .... ........... .. .........0
Questionnaire on Individual and Group Information .............. ...............204....
B WEALTH RANKING CRITERIA FOR GROUP MEMBERS ................. ............... .....207
C INSTITUTIONAL CAPACITY AS SES SMENT TOOL ........................... ...............208
D STAGES OF FARMER GROUP S AND THEIR AGRO-ENTERPRISE CAPACITY......2 12
LIST OF REFERENCES ................. ...............213....._.._ ....
BIOGRAPHICAL SKETCH .............. ...............223....
LIST OF TABLES
1-1 Impacts on food security, poverty alleviation, and sustainable development. .................. .27
1-2 Collective action initiatives and improved performance .............. ....................4
2-1 Size and composition of TIP and FAIDA groups ......____ ............ ..........__.....6
2-2 Group membership by partner agency and sex ................. .....__ ............. ......6
2-3 Years of schooling of producer group members ......____ .... ......... ........._.......6
2-4 Maj or religious affiliations of producer group members ................. ................. ......69
2-5 Age range of producer group members ..........._ ..... ._ ....__ ............. ...69
2-6 Maj or ethnicity of producer group members ..........._ .....___ ...............69
2-7 Commodity types by PA farmer groups .............. ...............71....
3-1 Group study sample .............. ...............85....
3-2 Group asset variables by description and value ........._.._.. ....._.. ........_.._.......8
3-3 Group composition variables by description and value ........._.._.. ....._.._ ........._.....94
3-4 Group characteristics variables by description and value ..........._. ......... ..... ........._.95
3-5 Group heterogeneity variables by description and value ................. ...._ ..............96
3-6 Market access variables by description and value ....._____ .... ... .__ ...........__.....9
4-1 Marketing obstacles by PA groups ..........._ .....___ ...............107
4-2 Livelihood asset constraints by farmer group marketing obstacles .............. ........._....113
4-3 Marketing performance rating by partner agency ...._.._.._ .... ... ..._. ......_.._........114
4-4 Test of significance using ANOVA and Pearson' s R ................. .......... ...............116
4-5 Group asset variables ................ ...............117...............
4-6 Group composition variables ................ ...............120...............
4-7 Gender categories by community leaders and providers/partners ................. ................121
4-8 Group characteristics variables ................. ...............123...............
4-9 Group heterogeneity variables ................ ...............125...............
4-10 Market access variables .............. ...............126....
4-11 Group wealth categories by statistically significant variables ................. ................ ...129
4-12 Regression model 1 for determinants of marketing performance............... ...............133
4-13 Explanatory variables categorized by factor domains .............. ...............134....
4-14 Regression models 1 & 2 for determinants of improved marketing performance ...........135
5-1 Partner agency by marketing performance .............. ...............140....
5-2 Partner agency by group composition and characteristics ................. ............ .........140
5-3 FAIDA and TIP farmer groups by group assets .............. ...............142....
5-4 Partner agency by market access .............. ...............142....
5-5 Marketing strategies and corresponding collective actions .............. ......................4
5-6 FAIDA group marketing strategies and corresponding collective actions ...................149
5-7 TIP group marketing strategies and corresponding collective actions ............................150
5-8 Marketing strategies by own initiative and PA linkage ......___ ........._ ..............154
5-9 Marketing strategies by group composition and characteristics ..........._.. .................158
5-10 Marketing strategies by wealth, water, and land............... ...............162.
5-11 Marketing strategies by market access variables ......._................. ............... ....16
LIST OF FIGURES
1-1 Informal and formal internal institutions .............. ...............30....
2-1 Primary AMSDP stakeholders (adapted from IFAD, 2001, pp. 11-12) .........................56
2-2 Communication and budgetary flows of primary stakeholders............ ... .........___...57
2-3 Market chain actors as other key stakeholders .............. ...............58....
2-4 Map of Tanzania with designated research sites (adapted from IFAD, 2001, p.3)...........60
3-1 Sustainable livelihoods framework ................. ...............77................
3-2 Conceptual framework for research study ................. ...............81...............
4-1 Onions packed in rumbesa............... ...............109
5-1 Ansoff matrix of marketing strategies and level of risk ................. ................ ...._.145
6-1 The driving forces and enabling factors of group marketing performance .............. .....177
8-1 Stages of agro-enterprise development and degree of farmer participation ................... .196
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
LINKING FARMERS TO MARKETS: ASSESSING PLANNED CHANGE INITIATIVES TO
IMPROVE THE MARKETING PERFORMANCE OF SMALLHOLDER FARMER GROUPS
IN NORTHERN TANZANIA
James G. Barham
Chair: Anita Spring
Major: Interdisciplinary Ecology
The primary inquiry of my study was to identify and understand the underlying factors that
enable smallholder farmer groups to improve their market situation. My study evaluated a
government-led program in Tanzania that is attempting to increase smallholder farmers' incomes
and food security through a market-oriented intervention. Over the course of 15 months,
research was conducted in collaboration with two local non-governmental organizations that are
implementing partners in this development program. Two rounds of group interviews (6 to 8
months apart) were conducted with 34 farmer groups. Individual data were collected on
approximately 400 farmer group members. My analysis considered how group assets and
characteristics, as well as the larger infrastructural conditions, affect a farmer group's ability to
improve its market situation. Findings suggest that more mature groups with strong internal
institutions, functioning group activities, and a good asset base of natural capital are more likely
to improve their market situation. Structural social capital in the form of community leaders and
ties to external service providers are enabling factors that can enhance a group's ability to access
new resources. Likewise, cognitive social capital in the form of group altruism is an enabling
factor that can enhance a group's ability to maximize existing resources. Gender composition of
groups also factors in group marketing performance. It acts as an enabling factor for male-
dominated and gender-balanced groups and acts as a disabling factor for female-only groups.
Cognitive social capital in the form of group trust and homogeneity of identities (i.e., ethnicity,
age, education, religion) are not significant factors in a group's ability to improve its market
In recent years, the importance of smallholder agriculture has been greatly recognized,
demonstrated by both the donor community and governments' pledge to engage in the requisite
interventions to generate agricultural and economic growth. In post-structural adjustment Africa,
this growing recognition has led to two maj or crosscurrents of theory and practice that now
define the maj or policy directives concerned with boosting Africa' s faltering agricultural
economies. First, agricultural development will not occur without engaging smallholder farmers.
Accounting for the overwhelming maj ority of actors in this sector, smallholder farmers must be
made central to any strategy to revitalize not only the agricultural sector, but also the economy as
a whole (Diao & Hazell, 2004; Resnick, 2004; IFAD, 2003; Magingxa & Kamara, 2003;
USAID, 2002; IFPRI, 2002; Wheatley, 2001; Wiggins, 2000). The second current, which
intersects with the first, is that the maj or obstacle facing smallholder-led agricultural growth is
lack of market access. Thus, the maj or proponents of market-led growth contend that enhancing
market access for smallholders will lead to increased incomes and food security, more
opportunities for rural employment, and sustained agricultural growth (Dorward et al., 2003;
Stiglitz, 2002; Poulton et al., 1998; Bates, 1997). But, there are a number of researchers who
adamantly oppose this market-led growth model for smallholder farmers. They argue such a
model may lead to the opposite effect, with increased food insecurity due to market dependence
and volatile prices, as well as bring about other negative impacts, such as concentration of
landholdings; increased environmental degradation from intensive farming practices; and
reduced biodiversity (Altieri, 1995; Wheatley, 2001; Omamo & Farrington, 2004).
Nonetheless, market access proponents make a strong and attractive case that for small
farmers to thrive in the global economy, it is necessary to create an entrepreneurial culture in
rural communities where "farmers produce for markets rather than trying to market what they
produce" (Lundy et al., 2002, p. 19). From an implementation perspective, this means shifting
the focus from production-related programs to more market-oriented interventions. This has
placed renewed attention on institutions of collective action most often realized through the
structure of farmer groups as an important and efficient mechanism for enhancing the
marketing performance of smallholder farmers (Kariuki & Place, 2005).
This study takes place within the context of these two crosscurrents of theory and practice.
The primary obj ective is to identify and understand the underlying factors that enable
smallholder farmer groups to improve their market situation. This topic is approached through
an evaluation of a government-led program in Tanzania the Agricultural Marketing Systems
Development Program (AMSDP) that attempts to increase smallholder farmers' incomes and
food security through improvements in market access. Within AMSDP, improving market access
includes the following components: (1) reforming the regulatory and taxation system; (2)
improving market infrastructure (e.g., building more roads, post-harvest facilities, market
centers); (3) establishing agricultural marketing information systems; and (4) strengthening
farmer groups and creating market linkages.
This study focuses on this last component of strengthening farmer groups and creating
market linkages. This particular component was to be accomplished through a joint partnership
between the district governments and local non-governmental organizations (NGOs). The main
tasks of the NGOs here termed partner agencies (PAs) as they are referred to in the program -
are to train existing or newly formed farmer groups in capacity building and marketing skills
measures and where possible, to establish market linkages. Through the lens of a livelihoods
approach that will be defined below, this program component attempts to improve the marketing
performance of smallholder farmers by enhancing their stock of human and social capital. The
aim of providing skills training in marketing and entrepreneurship increases human capital,
whereas linking these groups to others in the market chain forges new business partnerships
thereby enhancing their social capital.
To identify and understand the underlying factors that enable smallholder farmer groups to
improve their market situation, the study has the following fiye research obj ectives:
* Farmer group assets: To assess the extent that certain livelihood asset configurations
(i.e., natural, physical, Einancial, human, and social) will affect the group's ability to
improve their market situation.
* Farmer group composition: To assess the extent that certain group composition
attributes will affect their ability to improve their market situation
* Farmer group characteristics: To assess the extent that certain group characteristics,
which include the group's institutional capacity, will affect their ability to improve their
* Market access: To assess the extent that physical market barriers will affect the group's
ability to improve their market situation.
* Partner agency: To assess the extent that partner agency interventions in the form of
group-capacity training and market linkages will affect the group's ability to improve
their market situation.
Structure of the Dissertation
This dissertation is organized into seven chapters including this introduction. Chapter 1
reviews the literature in relation to the research obj ectives reviews. The research obj ectives were
informed by an array of development theory and practice that ranges from the work of
anthropologists and political scientists to the work of ecologists and economists. Chapter 2
provides background to the Agricultural Marketing Systems Development Program (AMSDP)
and describes the study area. Chapter 3 presents the research methods employed in this study.
While the first three chapters set the foundation for the research design and data collection,
chapters 4 and 5 present the results of the quantitative and qualitative data analysis. The results
of the quantitative data analysis (i.e., bivariate and multiple linear regressions) are provided in
Chapter 4. Chapter 5 considers findings from descriptive statistics and qualitative data analysis
techniques (i.e., cross-tabulations, categorizations, classifications) to explore the relationship
between the explanatory variables and the different types of marketing strategies employed by
farmer groups to improve their market situation. The purpose of Chapter 6 is to summarize the
Endings by linking the results to the study hypotheses and to provide conclusions that relate to
the theories guiding this study. The Einal chapter presents conclusions, as well as research and
programmatic recommendations specifically tailored to non-governmental organizations (NGOs)
engaged in agro-enterprise development initiatives.
There are two principal sections that follow. The first section discusses the economic policies
and perspectives concerning sub-Saharan Africa. Although no hypotheses are derived from this
section, it remains important because it contextualizes this study within the broader debate of the
role of agricultural development in the post-structural adjustment period in Africa. The second
section provides the overarching theories that have guided my research questions, providing a
critical lens within which to analyze a planned change process, and allowed for the formulation
of testable hypotheses. The section includes discussions on the following schools of thought:
New Institutional Economics; Feminist Economics; Social Capital; Collective Action; and
Economic Policy and Perspectives in Sub-Saharan Africa
This section provides an overview of the dominant economic policy and perspectives in the
field of agricultural development in sub-Sahara Africa. It provides the necessary macro-lens
within which to view my study, which though focused on more micro-level analysis, must be
cognizant of the larger policy implications. This section discusses the following: the rationale
for and implementation of the structural adjustment programs, with particular reference to
agricultural reforms and market liberalization enacted under these policies; the impact of these
policies on agriculture; the post-structural adjustment period and emergence of a new policy
agenda; the role of smallholder farmers in this agenda; and the promotion market access and its
possible impacts on the livelihoods of smallholder farmers.
Structural adjustment programs in Africa. The World Bank and Intemnational
Monetary Fund (IMF) implemented structural adjustment programs (SAPs) as an attempt to
stave off the economic crises unfolding in sub-Saharan Africa in the late 1970s. There were both
internal and external reasons for the economic crises. Externally, most of the world, and
particularly oil dependent developing countries, were negatively impacted by the sudden increase
in oil prices in 1973 and 1979. Internally, many of sub-Saharan African' governments,
following their independence from colonial regimes, pursued a dualistic or bimodal strategy that
induced a capital-intensive path of development (Tomich et al., 1995). Many of these state-led
interventions were geared toward an urban bias of industrial expansion and appeasing powerful
constituents. These actions heavily distorted the prices not only of food, but also foreign
exchange, capital, wage earnings of the civil sector, among others none of which in the long
term were economically and politically sustainable (Bates, 1981). These ill-conceived
development strategies led to mismanaged fiscal and monetary policies that effectively distorted
the prices of the relative scarcity and abundance of resources to be found within many of the
African countries (Timmer et al., 1983).
1From this point on, sub-Sahara Africa will be referred to as Africa. Thus, this designation does not refer to the
countries of North Africa, which generally had different trajectory in regard to economic development since the
With their economies stagnating, many African countries were incapable of servicing their
debt to both Western governments and banks. SAPs were the IlVF/World Bank' s corrective
answer to the growing economic crises. The measures enacted under SAPs included:
[The] devaluation of overvalued currencies, increases in artificially low food prices and
interest rates; a closer alignment of domestic prices with world prices, an emphasis on
tradable/exportable and the gradual withdrawal of restrictions on competition from abroad
(trade liberalization), privatization policies (of "parastatals" or large-scale government
monopolies), a decrease in government spending, wage and hiring freezes, reductions in
employment in the public sector or the minimum wage, the removal of food and input
subsidies, and across-the-board reductions in budget deficits as ways to invigorate
stagnating economies (Gladwin, 1991, pp. 3-4).
Taken as a whole, SAPs were implemented to "get the prices right" and thus end what in many
economists' eyes were the apparent "market pathologies" that existed in most developing
countries due to heavy handed state-led interventions.
Agricultural market reforms were at the forefront of structural adjustment programs. The
so-called Berg report (World Bank, 1981) catalyzed support for drastic agricultural reforms in
many of Africa' s worst-off countries. Kherallah et al. (2000, p. 6) point to three maj or types of
agricultural reform measures:
* Liberalizing input and output prices by reducing or eliminating subsidies on agricultural
inputs such as fertilizers and credit, realigning domestic crop prices with world prices,
eliminating pan-seasonal and pan-territorial pricing, and reducing overvalued exchange
* Removing regulatory controls in input and output markets (for example, allowing the
participation of the private sector in agricultural marketing), lifting restrictions on internal
movement of food crops, and relaxing quantitative controls such as delivery quotas and
* Restructuring public enterprises and withdrawing marketing boards from pricing and
marketing activities and narrowing their role to more supportive activities (such as
providing market information services and maintaining security stocks).
With such disincentives removed from the market system, proponents of market liberalization
expected such reforms to stimulate farmers to produce more, leading to higher agricultural
growth, and greater economic development overall.
Impacts of the structural adjustment period in Africa. With the hindsight of over
twenty-five years of SAP implementation, it is generally agreed upon that these policies had
negative, and in some cases, disastrous impacts on a greater number of African countries,
particularly the deleterious effects of these policies on the poor and women (Elson, 1989;
Gladwin, 1991; Sparr, 1994; Spring, 1995). The ineffectiveness of these economic policies
compounded by other known factors, such as low land productivity, inadequate infrastructure,
vulnerability to natural disasters, high levels of political instability, and a high prevalence of HIV
and AIDS, has put much of Africa into a dismal economic and political state (Resnick, 2004, p.
The current economic plight of many African countries is well known. Africa not only has
the highest proportion of its population living under the international poverty line (below $1
dollar a day), but it also has seen this proportion increase from 47.4 percent in 1990 to 49 percent
in 2000. Likewise, with over one third of its population malnourished, Africa has the highest
percentage in the world (Resnick, 2004, p. 7).
African countries are still heavily dependent on traditional export crops (e.g., coffee,
cocoa, cotton, sugar, tea, and tobacco) and a maj or part of SAP agricultural reform was intended
to boost productivity in this area. Unfortunately, the region's share of total world agricultural
exports has fallen from about six percent in the 1970s to three percent by 2003 (Diao & Hazell,
2004, p. 2). This is compounded by the fact that agriculture grew more slowly than overall
population growth from 1965 to 1998, and more slowly than growth in the agricultural labor
force from 1980 to 1998 (Dorward et al., 2003, p. 74). With the dismal market for many of
Africa' s traditional export crops, many smallholder farmers are currently resorting back to food
staples to meet subsistence needs. With further withdraw of the state from remote rural areas,
there is a veritable "scramble" by Africa peasants to find viable livelihood alternatives outside of
agriculture (Bryceson, 2002). Food insecurity also remains a major problem, with shipments of
over 3 million metric tons of cereal food aid required in 2002 alone (Resnick, 2004, p. 8).
The economic situation is further compounded by the high prevalence of HIV and AIDS
that from an economic perspective, "decreases labor productivity, erodes livelihood assets, and
blocks the transfer of knowledge from one generation to the next" (Diao & Hazell, 2004, p. 3).
Based on a report by FAO, the AIDS pandemic reduces national economic growth rates across
Africa by about 2 to 4 percent per year (Diao & Hazell 2004, p. 3).
Post-structural adjustment Africa and the new policy agenda. While certain circles
still debate the degree that SAPs should take the blame for Africa's economic plight, there are
signs of a partial convergence between supporters and critics concerning the post-adjustment
agenda. This convergence comes from a certain resignation by SAP critics that market
liberalization is irreversible if not a necessary prescription to economic development. Thus,
while the post-adjustment agenda takes liberalization as a given, it also accepts that
improvements in rural livelihoods require a wider series of interventions. Divergence still exists,
however, regarding the role the state should play in promoting economic development (Ponte,
2002, p. 3). Poulton et al. (1998, pp. 4-5) suggest that three main positions have emerged:
*The "conventional wisdom" position that argues SAPs were correctly designed but were
poorly implemented by African governments. This position views that further market
liberalization policies are necessary and the state should limit its market function to
enforcing contracts and property rights. The state may also provide some service
provisions, such as health, primary education, and possibly agricultural research.
Supporters of this position include Jayne et al. (2002); Kherallah et al. (2000); and Timmer
et al. (1983).
* The "emerging orthodoxy" position shares much in common with the conventional
wisdom perspective but relaxes some of the more rigid positions of the neoclassical
economic perspective. This position sees the state as playing an "activist" role in
promoting market development. This is done primarily through building the necessary
institutional and governance structures that enhance market efficiency. This position is
shared by a growing number of economists and political scientists who ascribe to the
"institutionalists"' school of development economics. Supporters of this position are many,
including Dorward et al. (2003); Stiglitz (2002); Friis-Hansen (2000); Poulton et al.
(1998); Bates (1997); and North (1990).
* The "critical" position argues that the withdrawal of the state in places where the private
sector involvement is structurally limited has been too quick, and that the options of
designing alternative forms of public interventions and reforming existing ones deserve
more careful attention. Further to the left of this position is the argument that the real
problems do not lie with inappropriate policies but more with exogenous factors (e.g.,
unequal trade relationships) that hinder meaningful agricultural development. Supporters
of this position include Omamo (2003) and Ponte (2002).
Economy of affection. One of the "critical" positions not reflected in the above typology
is the notion of the "economy of affection" as used by political economists and other scholars.
The concept was first developed by Goran Hyden in his seminal work, Beyond Ujamaa in
TaTnzania: thiderdevelopment and an uncaptured peasai~ntry (1980). In this book and subsequent
writings, Hyden argues that economic underdevelopment in Africa is not a matter of too much or
too little state or market intervention. Having witnessed the ultimately disastrous consequences
of a state-led economy in Tanzania in the 1970s, as well as the equally ineffective market
liberalization policies that later followed, Hyden argues that a "third" economy exists in much of
Africa that functions outside the realm of both state and market control. Firmly established in a
peasant mode of production, the economy of affection "denotes a network of support,
communications and interaction among structurally defined groups connected by blood, kin,
community, or other affinities, for example religion. It links together in a systematic fashion a
variety of discrete economic and social units which in other regards may be autonomous"
(Hyden 1983, 8). The economy of affection is based on social and economic exchanges between
affective kin ties that function autonomous of the state and the market, and thus, leave the
African peasantry uncapturedd" by the capitalist system. As Hyden (1980, p. 25) states:
The peasants are the owners of the means of production (or at least they control their use
more effectively than the officials) and thus they can always seek security in withdrawal.
As long as labor rather than land is the real scarce resource, officials will have difficulty in
exercising power over the peasants. The peasants, rather than officials, act from a position
While economy of affection may place peasants in a position of strength and autonomy vis-a-vis
the state, it also fosters a subsistence-mode of production that perpetuates underdevelopment.
The reciprocal exchange relations based on the notions of the economy of affection create
leveling mechanisms that lessen social differentiation and economic stratification by
discouraging capital accumulation and individual advancement. In one of his more recent
writings, Hyden (2003) references a study examining entrepreneurship in Tanzania (Trulsson,
1997) to make this case. As he points out:
[The study] showed that successful businesspersons tended to diversify their activities
without much sense of strategy and certainly without the ability to establish 'dynasties'
that can reproduce themselves by virtue of careful management of assets. One reason for
this preference of diversification was the need to respond to claims for sharing the riches
by members of the extended family or even lineage. Money went for 'leveling' rather than
'growth', an approach that helped keep the economy of affection alive, but did not
contribute to the formalization of capitalist relations (Hyden, 2003, p. 14).
Although Hyden's analysis does not fit neatly within the "critical" position as posited by Poulton
et al. (1998), it does function as a critique of those proponents of the "emerging orthodoxy" by
arguing that efforts to establish institutional and governance structures to enhance market
efficiency will prove largely ineffectual since the economy of affection still dominates much of
the social and economic life of the African peasantry. Much of theory behind the "emerging
orthodoxy" position is embodied in the work of new institutional economists. Given the
influence of this school of economic thought on policy makers promoting economic development
in Africa, the next section deals with some of the underlying principles behind new institutional
economics and the policy implications.
New institutional economics and the emerging orthodoxy. The new institutional
economic (NIE) position on market liberalization and privatization falls somewhere between the
argument for SAPs not being pushed far enough and the position that SAPs were pushed too far,
too fast (Friis-Hansen, 2000).2 Proponents of the NIE agree that liberalization and privatization
of the agricultural sector was, and is, the right path to development, but adamantly disagree with
neo-classical conventionall" that simply removing state-caused market distortions will provide
the private sector with sufficient incentive to take over the functions left behind by the state.
Particularly within the agriculture sector, disappointing production responses to state withdrawal
left many conventional scrambling to explain this condition. The answer has already been
provided: the reforms did go far enough and the state's role in economic development needs to
be even further removed. NIE, however, provides a different viewpoint on this, as detailed by
Poulton et al. (1998, p. 2):
An alternative explanation for the apparently disappointing response to market
liberalization was that enthusiasts such as Berg had held an unduly optimistic view of the
potential of the African private sector to provide the services previously provided by state
organs. The observed vitality of small-scale informal sector activity was insufficient
evidence to prove the existence of private entrepreneurs who could assume roles of
national economic importance, for example the marketing of staple foods to feed rising
urban populations. Private traders, often discriminated against and harassed for decades,
did not possess the managerial skills, business experience or capital to move straight into
Furthermore, in many rural areas, with geographically dispersed farms, poor roads, and weak
communication channels, there is simply insufficient business volume to attract private sector
involvement. Also, a number of services provided by the state have "clear public good
2 A more detailed theoretical discussion on NIE will be presented later in this chapter, with the present discussion
focusing on the NIE policy agenda in the post-structural adjustment period.
properties, which will tend to discourage private sector involvement. There are, in other words,
high probabilities of market failure in key liberalized markets" (Dorward et al., 1998, p. 3).
What becomes clear through the lens of a NIE analysis is that market failure in Africa is more
the norm than the exception. Much of the policy advice for African agricultural economies is
based on unrealistic analysis and assumptions that offer few real solutions to the problems at
hand (Omamo & Farrington, 2004). For NIE advocates, the way through this impasse is to stop
focusing so narrowly on how to "get the prices right" and start considering how to "get the
institutions right." As Dorward et al. (2003, p. 79) explain:
The essence of the "new institutional" argument is the very low level of development in
the institutional environment of poor rural areas, together with a low density of
transactions, leads to a very high transaction risk and costs in financial, input, and output
markets. This is particularly the case with financial markets and to a lesser extent with
input markets. High transaction costs and risks, exacerbated by low population densities
and poor communications, lead to coordination and hence market failures, and as these
market failures depress the level of economic activity, raising per unit transaction costs
(with thin markets) risks of transaction failure, a vicious cycle of underdevelopment
Institutional development, thus, becomes the key ingredient for agricultural development. Many
NIE proponents even argue that direct state interventions at the micro-level may be necessary to
support, and even create, functioning institutions to lower transaction costs and reduce risk and
uncertainty for poor smallholder farmers. The NIE position will be discussed in greater detail in
the theory section. Attention at this point turns to the role of smallholder farmers in agricultural
Importance of smallholder agriculture. Concomitant with the rise of NIE as the
emerging orthodoxy, renewed interest has emerged on the role that smallholder agriculture can
play in revitalizing stagnating agricultural economies. This interest comes on the heels of the
United Nations Millennium Summit and the adoption of the Millennium Development Goals
(MDG). A great deal of attention is being paid toward the MDG goal that hopes to cut hunger in
Africa in half by 2015. Smallholder farmers have become a focal center of how this goal is to be
accomplished. Indeed, if there is one coherent theme that has emerged in the past five years in
agricultural policy, it is this: "Smallholders Can Do It!" The leading development documents
and strategy papers take on such titles extolling the virtues of smallholder farmers, such as
"Ending Hunger in Africa: Only the Small Farmer Can Do It" (IFPRI, 2002); "Small Farmers are
A Big Deal" (Narayan & Gulati, 2002); "Cutting Hunger in Africa Through Smallholder-led
Agricultural Growth" (USAID, 2002).
Beyond the enthusiastic rhetoric, there is a great deal of logic to this present focus on the
role of smallholder agriculture. Agriculture is the primary source of livelihood for
approximately 65 percent of Africans. It represents 30 to 40 percent of African GDP and
accounts for almost 60 percent of Africa' s export income. Small-scale farms account for 90
percent of Africa' s agricultural production and are dominated by the poor. Smallholder farming
households also contain 75 percent of Africa' s underweight children (Resnick, 2004, p. 8). Not
only does it make sense in terms of concentrating finite resources, but smallholder farmers also
offer a number of comparative advantages over large-scale and industrial farming. More often
the case than not, this includes higher productivity per acre of land (Tomich et al., 1995); lower
input costs through family and other labor-sharing arrangements, and utilizing local knowledge
and indigenous farming techniques (Poulton et al., 2005).
Promoting market access. The leading research institutes and development agencies now
advocate that the key to revitalizing the agricultural sector must include promoting market access
for smallholder farmers. In fact, there has been a proliferation of documents since the UN
Millennium Summit on improving market access for smallholder farmers (e.g., Wiggins, 2000;
Wheatley, 2001; USAID, 2002; IFPRI, 2002; Lundy et al., 2002; Magingxa & Kamara, 2003;
IFAD, 2003; Omamo, 2003; Diao & Hazell, 2004; Resnick, 2004). The bottom line is that a lack
of marketing opportunities acts as a powerful disincentive for farmers. In order for smallholder
farmers to thrive in the global economy, it will be necessary to create an entrepreneurial culture
in rural communities, where "farmers produce for markets rather than trying to market what they
produce" (Lundy et al., 2002). This has shifted emphasis away from production-oriented
programs to more marketing-related activities. As two leading policy advisors readily admitted:
Marketing is an aspect of project work that is frequently neglected and that deserves more
attention when designing proj ects ...improving marketing systems can be as important as
introducing farming methods that increase yields; indeed the latter may otherwise be
ineffective (Baum & Tolbert, 1985, p. 94).
With the MDG in mind, market enthusiasts insist that promoting market access for smallholder
farmers holds the potential to increase incomes, reduce food insecurity, create new rural
employment, lower malnutrition, and, most importantly, set the foundation for real positive
growths in agriculture and the overall economy. There is also further emphasis given to ensuring
the poor have a stake in the process. As one IFAD (2003, p. 5) document emphatically states:
Improved market access is not an issue of consequence only to better-off producers, and it
is not relevant only to cash crop, rather than food crop, production. It is of importance to
all rural households, and assisting rural poor people in improving their access to markets
must be critical element of any strategy to enable them to enhance their food security and
increase their income.
Defining markets and market access. The MIT Dictionary of Modern Economics (1999,
p. 266) defines markets as "generally, any context in which the sale and purchase of goods and
services takes place." Although straightforward, the use of "sale" lends a certain monetary slant
to markets that dismisses the inclusion of markets where money is not the medium of exchange.
A more appropriate usage comes from Plattner (1989, p. 171), an economic anthropologist, who
defines markets more inclusively as "the social institution to exchanges where prices or
exchange equivalencies exist." He also defines the "marketplace" as the interactions in a
customary place and time, and "marketing" to denote the buying and selling in a market.
Beyond a definition of markets, there are also different types of markets. Bates (1983, p. 3)
helps to delineate these markets and where farmers find themselves positioned:
Farmers are seen as standing at the intersection of three major markets. Their real incomes
depend upon their performance in these markets. They derive their revenues from the sales
they make in the first of these markets the market for agricultural commodities. Their
profits are a function of these revenues, but also the costs they incur in a second market -
the market for factors of production. And the real value of their profits, and thus the real
value of their incomes, is determined by the prices they must pay in a third maj or market:
the market for consumer goods, particularly commodities manufactured in the city.
Dorward et al. (2003) also delineate between three markets, these being the input market (for
labor, land, agricultural inputs), the financial market (for cash and credit), and the output market
(for selling agricultural products).
As for market access, the concept is largely defined, or at least explained, but what
impedes access to markets. IFAD (2003, pp. 9-11) considers this along three dimensions:
physical barriers; market structures; and information, skills, organization. The definitions
include the following:
* Physical barriers: The distance to markets, including the lack of roads or impassable
roads during certain seasons, is a fundamental constraint for many rural communities. It
undermines farmers' ability to buy their inputs and sell their crops. It results in high
transportation and transaction costs, both to buyers and sellers, and it leads to
uncompetitive, monopsonistic markets.
* Market structures: Along with physical barriers, many rural markets are typified by large
asymmetrical relations where there are a large number of small producers/consumers on
one hand, and a few market intermediaries on the other hand. Such market relations tend
to be uncompetitive, unpredictable and highly inequitable. Such power differentials make
it difficult for poor farmers to get fair prices on what they produce, as well as have
effective demand for inputs and requisite consumption goods.
* Information, skills, organization: As participants in markets, poor farmers are at a
distinct disadvantage. Poor farmers lack access to market and marketing information,
research, education and skill training. Furthermore, many lack the collective organization
that can give them the power they require to interact on equal terms with other, generally
larger and stronger, market intermediaries. As IFAD (2003, p. 11) emphasizes, "with little
experience, no information, and no organization, they have no basis upon which either to
plan a market-oriented production system or to negotiate market prices and conditions.
Ultimately, their lack of knowledge means that they are passive, rather than active, players
in the market; that they can be exploited by those with whom they have market relations;
and that they fail to realize the full value of production."
Given the physical and social barriers to market access faced by smallholder farmers,
development planners have the tendency to be overly optimistic of the potential benefits if access
to markets can be improved. However, unlike many market enthusiasts, Wheatley (2001, p. 4)
provides a more balanced perspective on the positive and the negative aspects of greater market
access and integration.
Table 1-1. Impacts on food security, poverty alleviation, and sustainable development.
Market for surplus production of traditional
Improved storage of fresh produce, reduced
wastes and losses
Increased food entitlements
Reduced cost of imported basic foodstuffs
Improved health and nutrition
Increased income from higher value crops
and added value post-harvest activities
Employment generation in larger rural
Market-based incentives for sustainable use
of natural resources.
Market-based incentives for clean
production technologies, and recycling of
Market based incentives for applying food
quality and safety standards
Switch from food crops
Health problems with
Concentration of land holdings
and productive resources
Reduced crop biodiversity
More intensive production
systems, more use of external
Increased pollution from post
In Table 1-1, he classifies the possible impacts along the lines of food security, poverty
alleviation, and sustainable development. Wheatley's overview of negative and positive impacts
is a reminder that promoting market access for smallholder farmers may bring positive short-
term advantages, but could lead to long-term negative consequences (environmental pollution
and loss of biodiversity being maj or concerns). Thus, the maj or question facing development
planners, in both the short and long terms, remains how to maximize the positive aspects of
market access while minimizing its negative aspects.
Theoretical Considerations and Hypotheses
A fruitful analysis of human action requires us to avoid the atomization implicit in the
theoretical extremes of under- and oversocialized conceptions. Actors do not behave or
decide as atoms outside a social context, nor do they adhere slavishly to a script written for
them by the particular intersection of social categories that they happen to occupy. Their
attempts at purposive action are instead embedded in concrete, ongoing systems of social
relations (Granovetter, 1985, p. 487).
My theoretical and research pursuits share much in common with the ideas expressed by
Granovetter. Beyond the academic polemics of actors having complete agency and self-
maximizing utility to actors being fully guided by societal and ideological structures, there exists
the reality that actors are capable of behaving at both spectrums given the situation and context.
What Granovetter is seeking is a "middle path" where the reality of economic and social action
can truly be studied and explained. The theory in this section comes from five schools of
thought that share this perspective: (1) New institutional economics; (2) Feminist economics;
(3) Social capital; (4) Collective action; and (5) Cultural materialism. While these schools of
thought present wide-ranging perspectives, they are all interested in a similar question that
guides my overarching research pursuits, which is: To what extent do social structures and
institutions impact economic outcomes?
New institutional economics
New institutional economics (NIE) is a "vast and relatively new multidisciplinary field that
includes aspects from economics, history, sociology, political science, industrial organization,
and law" (Kherallah & Kirsten, 2001, p. 1). As a field of study, NIE is concerned fundamentally
with problems of market coordination and the institutional responses devised by economic agents
to solve these problems (Poole, 1999, p. 1). Although NIE shares the neoclassical assumption of
rational, self-interested agents, it relaxes some of the more restrictive assumptions of neoclassical
perfect competition. Neoclassical assumptions such as perfect competition, homogenous
commodities, economic agents motivated only by profit, and perfect information are regarded
within NIE to be more of the exception than the rule. As Poulton et al. (1998, p. 10) argue, "the
world of perfect competition is a rarely encountered special case of a much broader set of
economic scenarios and that economic analysis and policy making should concentrate on
understanding and improving actual scenarios." NIE is appealing to many development
practitioners specifically for this reason it is grounded in reality and concerned with the
examination of how real institution play out in actual markets and the different channels through
which they assist market exchange (Fafchamps & Minton, 2002).
Key concepts in NIE are uncertainty, the costs of contracting through market exchange, the
role of institutions in reducing costs, and their influence on the organization and development of
economic activity (Poole, 1999, p. 1). Institutions have specific meaning and should not be
confused with organizations. As North (1990, p. 3) explains, "institutions are the rules of the
game in a society; more formally, they are the humanly devised constraints that shape human
interaction." Institutions guide human behavior and reduce uncertainty in human interaction.
To be effective, institutions must be enforceable through some sort of sanctions. As one of the
leading figures in NIE, North (1990, p. 54) states quite emphatically, "the inability of societies to
develop effective, low-cost enforcement of contracts is the most important source of both
historical stagnation and contemporary development in the Third World."
A distinction is generally made between internal and external institutions. Internal
institutions refer to customs, ethical norms, conventions, good manners, etc. that have evolved
over time from past human experience and proven useful in solving particular sets of problems.
These internal institutions may be formal or informal, which is delineated by whether or not
sanctions require the use of formal enforcement mechanisms. Kasper & Streit (1998, p. 106)
provide an excellent overview of these distinctions in Figure 1-1.
Internal institutions can be:
*informal, that is, not sanctioned by formal mechanisms, namely:
0 conventions, that is, rules that are of obvious, immediate benefit to
the persons whose behavior they control and whose violations harm
o internalized rules whose violations are sanctioned primarily by a bad
o customs and manners which are sanctioned informally by the
reaction of others, for example by exclusion; or
*formalized, where the sanctions are implemented in organized manner by
some members of society
Figure 1-1. Informal and formal internal institutions
In contrast to internal institutions, external institutions are "imposed and enforced from above,
having been designed and established by agents who are authorized by a political process. An
example is legislation. External institutions are coupled with explicit sanctions which are
imposed in formal ways (e.g., law courts following procedures of due process) and may be
enforced by the legitimate use of force (e.g., police)" (Kasper & Streit, 1998, p. 3 1).
Institutions and economic performance. As briefly stated earlier, institutions function to
facilitate predictability, reduce risk, and thus prevent actors from engaging in opportunistic
behavior. Within this field of economics, the focus is on "how institutions promote order in
economic interaction, that is, the patterns that emerges when individuals try to come to grips
with the scarcity of resources. Order inspires trust and confidence, as well as reducing cost of
coordination" (Kasper & Streit, 1998, p. 195). When actors have trust and confidence that others
will act predictability according to specific sets of rules, this lowers not only the cost of
coordination but also the costs of transacting. These transaction costs include those of
information, negotiation, monitoring, coordination and enforcement of contracts (Bardhan, 1989,
p. 1389). One of the central tenets of NIE is that there are substantial transaction costs involved
in most forms of economic activity and that institutions can function to lower these costs.
Transaction costs are generally classified into four categories. There are information,
negotiation, monitoring and enforcement costs. Information costs occur before an exchange
transaction and include the costs of obtaining price and product information, as well as the cost
of finding a buyer/seller. Negotiation costs deal with the actual costs of the exchange, which
could include: commission costs; drawing up contracts; or the costs and time spent bargaining
over the exchange. Monitoring and enforcement costs occur after the transaction and involve all
the costs necessary to enforce the exchange contract (Lapar et al., 2006, p. 4).
By ignoring the fact that transaction costs exist, as is done by neoclassical economists, a
key ingredient is missing on how markets function, and why so often markets fail. As Haggblade
(2003, p. 12) makes clear:
Transaction costs are unique to each market participant. The presence of transaction costs,
which are specific to each market actor, implies that there is no single effective market
price at which exchange occurs. Each agent in the market conducts transactions on the
basis of his or her specific transactions costs. The implications of transaction costs are that
markets are thin or fail if prohibitively high costs prevent exchange.
Within neoclassical economics, a great deal of energy is focused on studying how profit-
maximizing agents try to minimize the sum of the transformation (production) costs. But NIE
takes this a step further by arguing for the inclusion of transaction costs (Poulton et al., 1998).
Indeed, it is from this perspective that NIE can more adequately explain why market failures
occur so frequently in Africa. Without the appropriate institutions in place, transaction costs
remain too high, and thus make market exchange too risky, unpredictable, or simply unprofitable
NIE proponents would certainly agree with their neoclassic counterparts that it is important
to remove inefficient and costly government parastatals in order to promote competition; but, if
these initiatives are not matched with an "institutional environment" and specific "institutional
arrangements" capable of dealing with emerging transaction and coordination costs, there is little
likelihood that marketing agents will enter the market with any frequency. In such an
institutional vacuum, markets will either tend to breed monopsonistic characteristics or simply
fail periodically or completely.
One final distinction needs to be made is between "institutional environment" and
"institutional arrangements." On the one hand, macro-level analysis in NIE focuses more on the
institutional environment that is defined as "the set of fundamental political, social, and legal
ground rules that establish the basis of production, exchange and distribution. Rules governing
elections, property rights, and the right of contract are examples..." (Davis and North, 1971,
quoted in Poulton et al., 1998, p. 1 1). One the other hand, those in NIE centered at the micro-
level are more likely to study specific institutional arrangements that are defined as "an
arrangement between economic units that governs the way in which these units cooperate and/or
compete" (Davis and North, 1971, quoted in Poulton et al., 1998, p. 1 1). This distinction is
important when analyzing the type of institutional innovations proposed by policy makers and
planners who espouse a NIE perspective. This could be through the development of an
institutional environment that lowers transaction costs, such as by recommending official
systems of grades and standards, establishing dispute settlement mechanisms and legal reform, or
promoting effective governance and equitable laws. At the micro-level, certain institutional
arrangements could be promoted, such as supporting contract farming, encouraging cooperatives
and farmer organizations, or supporting trader associations (Dorward, 2003, p. 79). These
examples typify the "emerging orthodoxy" agenda for agricultural development and, as briefly
mentioned in the introduction, is exactly the type of agenda being implemented in Tanzania
under the IFAD-sponsored program.
Hypotheses derived from the new institutional economics literature include:
* Hypothesis la: Farmer groups with functioning internal institutions for guiding group
behavior and action will be better positioned to improve their marketing performance.
* Hypothesis lb: Farmer groups that establish institutional arrangements with other chain
actors (i.e., rules for guiding exchange behavior) will be better positioned to improve their
* Hypothesis 10: Farmer groups that have lower transaction costs vis-a-vis other
competitors will be better positioned to improve their marketing performance.
Feminist economics encompasses a wide-ranging field of perspectives and development
practice. As a body of theory, feminist economics provides a counter perspective to some of the
male bias found in neoclassical economics. As part of the development discourse, feminist
economics first gained notice with its refutation of mainstream economics' perspectives on the
household economy, and later gained further traction with its scathing critique of the
macroeconomic policies and their impact on women and the marginalized during the period of
structural adjustment (Elson, 1994). More recently, feminist economics has contributed to the
post-structural adjustment debate, by providing alternative perspectives and thoughtful critiques
to the positions expressed in new institutional economics and the "emerging orthodoxy."
As a field of study, feminist economics shares with feminist research the explicit goal:
[To] ask into question existing hierarchies, authorities, norms, traditions, and conventions,
to deconstruct canons and question how current (scientific) practices support the status
quo, both in society as well as more narrowly in the scientific community (Robeyns, 2000,
Feminist economics differentiates itself from the mainstream and other heterodox economics by
constantly questioning the underlying gender implications of economic phenomena, highlighting
the power differentials between men and women and how these play out within institutions and
structures of larger society. Even though it is still an emerging field of study with varied
perspectives, there are some central issues. These include studies on the "dynamic within the
household, the consequences of policy on women, and the questioning of some 'male bias' in
dominant economic concepts and generally accepted ideas and assumptions in economics"
(Robeyns, 2000, p. 4).
Most of what follows focuses on the "feminist critical economics" school of thought that
will be reviewed primarily through the work of Folbre (1988); Elson (1989; 1994); Sparr (1994),
but also finds similar perspectives in the work of other well-known feminist researchers, such as
Moser (1991); Gladwin (1991); Bakker (1994); Spring (1995), Kabeer (1999); Bruce & Dwyer
(1988); and Westermann et al. (2005). Coined by Elson (1994), "feminist critical economics"
gained prominence through its critical attack on structural adjustment policies by positing that
the "operation of economic reform at micro-, meso-, and macro-levels is male biased, serving to
perpetuate women's relative disadvantage, even though the forms of that disadvantage vary
between groups of women and are disrupted and change in the course of policy reform" (Elson,
1991, p. 38).
Sparr' s ( 1994) book entitled, Feminist Critiques of Structural Adjustment, effectively
addresses the inherent gender biases found in the neoclassical model espoused by the World
Bank and International Monetary Fund (IMF). Sparr points to the fallacy of neoclassical
economics as a "value-neutral science." She argues that neoclassical theory is culturally and
historically specific; it outlines a broad view of economic rationality not inclusive to all societies
and peoples; and that it is "historically grounded in the experience of a handful of fairly
industrialized economies at a certain point in time" (Sparr, 1994, pp. 15-16). The consequence
of such an approach is the assumption that all societies will function as a "fully monetized,
market-oriented society." As Sparr (1994, p. 16) makes abundantly clear:
Of the flaws in the monetization and market presuppositions, two are crucially important.
First, in some countries, laws and custom may restrict or prohibit women's independent
control of money; ownership of property; and paid employment.... The theory's
assumption of how people respond to changes in wages, incomes and prices as well as the
primary motivation of economic activity (self-interest) are generalizations based primarily
on male experience.... [Second] this theory considers work performed, services rendered,
and products made that do not have an explicit price to have no economic value.... Thus,
much of what society deems as women' s work (bearing and raising children, preparing and
growing food for family use, cleaning the house, gathering fuel and water, etcetera) is
rendered invisible and unimportant for understanding how economies work.
Following on Sparr's second point, the feminist critical economics school argues that the
neoclassical theorists' inability to account for women' s productive and reproductive roles as
economically valuable has left policy-makers to assume that unpaid domestic work is "infinitely
elastic." But, as Elson (1989, p. 58) suggest, "a breaking point may be reached, and women's
capacity to reproduce and maintain human resources may collapse." Even if this point is not
reached, successful macro-policies only mean increased workdays for the maj ority of female
workers. Elson (1989, p. 58) explains, "terms like 'cost' and 'productivity' and 'efficiency,'
that play a large role in discussions of economic policy, are in fact ambiguous. What is regarded
by economist as 'increased efficiency' may instead be a shifting ofcost fr~om the paid economy to
the wipaid economy" (emphasis added). Thus, when policy-markers pressure governments to
reduce social service expenditures, there is an implicit reliance, as Sparr (1994, p. 17) says, "on a
quiet army of wives, co-wives, mothers, daughters, aunts, grandmothers, sisters, female friends
and neighbors to pick up the slack."
Household economy. Another assumption made within neoclassical analysis is that a
household is generally understood to be a "harmonious unit," wherein the decision-maker,
usually considered to be the man, maximizes the entire household's utility. But, as the feminist
critical economic school argues, this assumption flies in the face of overwhelming data that
prove this position wrong. As Bruce & Dwyer (1988, p. 2) point out:
These constructs of household behavior most prominently that of New Household
Economics are deficient, not only because they fail to acknowledge intra-household
negotiation over assets and possibly severe inequalities within households, but because
they tend to separate gender dynamics at the microeconomic level from the known external
dimensions of gender differentiation and asset distribution.
The inability of neoclassical theory to account for both gender composition of households and
the power dynamics within the households stems from a quasi-mythical foundation that
presupposes, as Folbre (1988, p. 262) puts it, "that the invisible hand swept the moral economy
into the home, where an imaginary world of perfect altruism could counterbalance the imaginary
world of perfect self-interest market."
The fallacy of the "harmonious" household is most apparent when viewing the rather
complex and dynamic characteristics of farming systems. In her extensive development
experience in Africa, Spring (1995, p. 1 19) finds the neoclassical household unit of analysis
In terms of gender issues, farming households in Africa might be characterized as
overlapping but semi-autonomous production and consumption units; women were not
simply embedded within a household, but were rather autonomous or semi-autonomous
persons who often had a great deal of independence in terms of food production and
distribution. They might make their own decisions on certain plots or for a particular farm
enterprise. Men and women usually had differential access to resources, priorities and
management strategies, and income uses. This contrasted with the old model in which the
farm was seen as a single enterprise with a single-decision maker (usually male).
As the feminist critical economic school points out, the neoclassical perspective fails to account
for both gender composition of households and the power dynamics that play an inherent role in
such households. Thus, policy-makers design development proj ects that have the capacity to
change the power relations of household in favor of men but not to women (such as access to
credit and technical support), regrettably at the expense of women who may have to work with
Furthermore, the neoclassical assumption that labor is freely available underscores the
ignorance on part of policy-makers to account for institutional and customary restrictions faced
by women in many countries that impedes their "free market entrance, exit, and mobility." Elson
(1994, p. 35) explains:
The ability of women to enter into economic contracts is constrained by the way that state
legislation typically construes women as less than full citizens. A key example of this is in
the context of economic policy reform in many developing countries is the way in which
the ability of women to enter into credit contracts is constrained by women' s lack of rights
to family assets. All too often, women cannot sign contracts in their own right and require
a male guarantor (father, brother, husband).
Elson further posits, that SAPs emphasis on privatization can have the adverse effect of reducing
individual social rights in the name of removing "distortions" from the market. Individual social
rights are often reduced when privatization shifts "employment from the formal sector to the
informal sector; by erosion of customary use rights to land by commercialization of land; and by
direct legislative change to withdraw or restructure state-provided services and benefits and to
abolish employee rights, such as minimum wage legislation and the right to strike" (Elson, 1994,
p. 35). Not surprisingly, it is women before men, in most cases, who will lose public sector jobs,
be forced into the informal sector, and be denied proper use-rights to land.
Feminist economics and institutional analysis. Just as feminist economics provided a
new and much needed perspective on the impacts of SAPs on women and the poor, the field of
theory and practice has continued to contribute to the ongoing post-adjustment policy debates in
Africa. While new institutional economics is the theoretical cornerstone of the "emerging
orthodoxy," feminist economics is firmly situated in the "critical" position of this policy debate.
Much of my own critique of the development program under study in Tanzania will utilize a
feminist perspective, relying heavily on the Social Relations Approach (March et al., 1999), a
gender analysis framework developed by Kabeer (1999). The framework is particularly useful in
analyzing how institutions and the organization they embody perpetuate gender and other social
inequities. As quoted in March et al. (1999, p. 102):
This framework focuses on the institutional construction of gender relations and hence the
institutional construction of gender inequality. It is intended to direct attention to the
existence of gender inequalities in the prevailing distribution of resources, responsibilities
and power and analyzes how they are thrown up by the operations of the institutions which
Institutions are created to reduce uncertainty and risk and therefore are important in reducing
transaction costs and promoting market development; but that does not mean that all actors
benefit equally from institutional development. Many institutions are created to serve the
interest of the powerful and perpetuate and reinforce discriminatory behavior and actions toward
certain actors, whether by gender, age, class, ethnicity, religion, or otherwise (Westermann et al.,
2005). The Social Relation's Framework draws attention to these inequities by analyzing the
formal and informal rules governing social relations at key institutional sites. These are
delineated into four domains: household, community, market, and state. In regard to the planned
change initiative under study, the framework will be particularly important to ascertain the extent
that community and market institutions produce, enforce, and reproduce the rules that perpetuate
socio-political and economic inequities.
All the presented and subsequently discussed hypotheses of this study will put under the
critical lens of a feminist economics' perspective in order to determine the extent that the
underlying factors that constrain or enable a group's ability to improve their market situation
have gender and other social inequity implications. But, this study does not want to loose the
trees for the forest (i.e., looking at all inequity implications and loosing focus on specific gender
discriminatory institutions). Therefore, to keep the study grounded in a gendered framework, the
following hypothesis is put forward:
*Hypothesis 2a: The gender composition of groups, especially of the decision makers, will
influence group marketing performance.
Coleman has been widely recognized as popularizing the concept social capital within the
Hield of development first in the article, "Social Capital in the Creation of Human Capital"
(1988) and shortly thereafter developing the concept in greater detail in his seminal work, The
Foundations of Social Theory (1990). He introduced the term social capital as a conceptual tool
to bridge the theoretical chasm created by the economists, who see actors having self-interested
goals independently attained, and the sociologists, who see actors as essentially non-actors, with
actions shaped wholly by the environment by which they inhabit. As Coleman (1988, p. 96)
I have argued for and engaged in the development of a theoretical orientation in sociology
that includes components from both these intellectual streams. It accepts the principle of
rational or purposive action and attempts to show how that principle, in conjunction with
particular social contexts, can account not only for the actions of individuals in particular
but also for the development of social organization.
Social capital, then, is largely defined by its function and consists of a number of entities that
have at least two elements in common: "they all consist of some aspect of social structures, and
they facilitate certain actions of actors whether persons or corporate actors within the
structure" (Coleman, 1998, p. 97). Social capital is distinguished from other social interactions
by its productive quality, and as such, should be perceived as a resource that help actors achieve
their specified interests.
While Coleman can lay maj or claim for introducing social capital as a conceptual tool,
there is no doubt that this term gained considerable academic currency through the works of
Putnam (1993a; 1993b; 1995) in Italy and the United States. His definition is the most oft-cited
By analogy with notions of physical capital and human capital tools and training that
enhances individual productivity 'social capital' refers to features of social organization
such as networks, norms, and social trust that facilitate coordination and cooperation for
mutual benefit" (Putnam, 1995, p. 67).
In his highly influential book, Making Democracy Work: Civic Traditions in M~odern Italy
(1993a), Putnam argues that the strongest determinant in Italy for socio-economic development
is the vibrancy of what he labels as "civic involvement" or "civic traditions," that he measures by
associational life, newspaper readership, and other indicators of political participation. Much of
the recent thinking on social capital has developed from the premises and empirical research he
carried out, for as Putnam himself argues: ..working together is easier in a community blessed
with a substantial stock of social capital... The social capital embodied in norms and networks of
civic engagement seems to be a precondition for economic development as well as for effective
government. Development economics take note: Civics matters" (Putnam, 1993b, p. 35).
Differing perspectives. While the term has gained wider acceptability both by theorist
and practitioners, social capital remains theoretically and conceptually elusive. Two differing
perspectives have emerged that attempt to bring greater conceptual clarity. One school of
thought follows Coleman's original perspective by viewing social capital as endowed in
individuals; while, the other school of thought resembles Putnam's perspective where social
capital is largely the property of groups, giving it public good aspects (Harriss & Renzio, 1997).
The "group" perspective represents a great deal of the work carried out under the auspices of the
World Bank Social Capital Initiative (World Bank, 2000), including such authors as Narayan
(1999), Grootaert (1999), and Uphoff & Wijayaratna (2000). This perspective often links
structural forms of social capital (e.g., social networks, groups, associations) with more
cognitive forms of social capital (e.g., trust, solidarity, shared norms and values). The focus of
inquiry is how groups maintain social capital as a collective asset to enhance their utility. In
contrast, the "individual" perspective focuses on how individuals invest in social relations to
generate utility. This perspective focuses strictly on social structures and represents much of the
recent literature by social network researchers, such as Granovetter (1992), Portes (1998), Lin
(2001), and Burt (2005).
Structural social capital. One area of agreement between both perspectives is that certain
types of structural social capital serve different purposes, and thus yield different benefits. The
most common typological distinction is between "bonding" social capital and "bridging" social
capital. Bonding social capital refers to "ties to people who are similar in terms of their
demographic characteristics, such as family members, neighbors, close friends, colleagues" and
bridging social capital refers to "ties to people who do not share many of these characteristics"
(Grootaert & Bastelaer 2001, p. 4). A number of researchers have made similar typological
distinctions for describing difference in social capital. These include Woolcock's (1998)
"embedded" social capital, which refers to intra-community ties (i.e., bonding social capital) and
"autonomous" social capital, which refers to extra-community ties (i.e., bridging social capital).
Similarly, Maluccio et al. (2003) refer to "bound" networks (i.e., bonding social capital) and
"achieved" networks (i.e., bridging social capital).
Lin' s (2001) work on social capital captures the essence of the typological distinctions of
structural social capital, and thus, to maintain a level of conceptual clarity, this study utilizes his
definitions for bridging and bonding social capital. Bonding social capital is defined as
"expected returns of expressive actions for preserving or maintaining resources" and bridging
social capital is defined as the "expected returns of instrumental actions for searching for and
obtaining resources" (Lin, 2001, p. 18). Examples of bonding social capital include:
membership in groups within a community (i.e., producer groups; processing groups; labor
sharing groups; rotating credit and savings associations); and farmer personal networks with a
high number of intra-community ties that are more horizontal in nature. Examples of bridging
social capital include: membership in groups and farmer personal networks with more extra-
community ties, such as relationships with service providers (i.e., local government ministries
and NGOs) and agribusiness partners.
Hypotheses derived from the social capital literature include:
* Hypothesis 3a: Farmer groups with a high level of trust among members will be better
positioned to improve their marketing performance.
* Hypothesis 3b: Farmer groups that exude more altruistic rather than self-interested
behavior among members will be better positioned to improve their marketing
* Hypothesis 3c: Farmer groups with more ties to other organizations in and outside of their
community will be better positioned to improve their marketing performance.
The first two hypotheses deal with cognitive social capital and its effect on group marketing
performance. The third hypothesis is concerned with testing the effects of structural social
capital on group marketing performance.
The literature on collective action in theory and practice emerged from dissatisfaction and
failures of many of the rural development programs of the 1960s and 1970s. The development
paradigms of this period assumed that communities would willfully engage in collective
activities, but gave little scrutiny to understand the conditions or how these actions might be
sustained (Meizen-Dick et al., 2004). Beginning with the work of Olson (1965) and followed by
the works of other including Axlerod (1984), Uphoff (1986), and Ostrom (1990, 1992), a body of
theory soon developed attempting to explain the enabling conditions for successful collective
action outcomes. Much like the concept of social capital, there is no one singularly agreed upon
definition. For example, Marshall (1988) defines collective action as "action taken by a group
(either directly or on its behalf through an organization) in pursuit of members' shared interests"
(quoted in Meizen-Dick et al., 2005, p. 2). The definition has merit but puts bias on the idea that
one must be a "member" of a group for collective action to occur. Meizen-Dick et al. (2005, p.
5) does not provide a definition but does point to a number of commonalities in collective action
What most definitions have in common is that collective action requires the involvement of
a group of people, it requires a shared' interest within the group and it involves some kind
of common action which works in pursuit of that shared interest. Although not often
mentioned, this action should be voluntary, to distinguish collective action from hired or
corvee labor (authors' emphasis).
Some of the greatest gains empirically and theoretically on the subj ect of collective action has
been found in the field of natural resource management (NRM). Of particular importance has
been the works of Wade (1988), Ostrom (1990; 1992), Baland & Platteu (1996), and Agrawal
(2001, p. 1659) synthesized these works in an effort to identify a common list of enabling
conditions for successful collective action outcomes. The following conditions were supported
in the literature: (1) small group size; (2) clearly defined boundaries; (3) shared norms; (4) past
successful experiences; (5) appropriate leadership; (6) interdependence among group members;
(7) heterogeneity of endowments, homogeneity of identities and interests, and (8) low levels of
Synthesis of social capital and collective action. A review of collective action theory
parallels the social capital literature. Uphoff & Wijayaratna (2000, p. 1876) highlight how
structural forms of social capital (i.e., roles, rules, procedures, social networks) facilita~te
mutually beneficial collective action and cognitive forms of social capital (i.e., norms, values,
attitudes, trust) are conducive for mutually beneficial collective action. The authors show how
these forms of social capital brought about successful collective action measures in management
of irrigation schemes. Other studies, such as Pretty & Ward (2001) and Krishna (2001), have
similarly shown how human and social capital formation often represented in community-
based groups have been pivotal in solving many of the communities' development problems,
particularly in the areas of natural resource management.
While there is substantial evidence behind the importance of social capital to maintain and
improve natural capital, there are far fewer studies that examine how social capital is utilized for
purposes of collective action to improve groups' marketing performance. This is particularly
apparent when examining the extent that group characteristics may influence or determine
certain marketing outcomes. The studies that do emerge are often looking at higher-tier
organizations, such as cooperatives or agribusiness enterprises. For example, Jones (2004)
shows how interpersonal trust and wealth heterogeneity among cooperative members were
enabling conditions for cooperative success, especially during the first stages of cooperative
formation. Johnson et al. (2002) show how social capital, as expressed through business firm
relationships, contributed positively to firm productivity and performance.
Social capital and collective action initiatives. The body of theory and empirical evidence
encapsulated in the social capital and collective action literature is particularly useful for
examining how certain group characteristics (i.e., group size, group age, current activities,
gender composition) and selected group asset endowments (i.e., physical, financial, social,
human, and natural capital) can be determining factors in successful collective action outcomes,
which in this case means an improved market situation for farmer groups. There are a number of
collective action initiatives that will improve directly or indirectly a farmer group's market
situation. Some of these initiatives are more geared to a group's production performance, others
more directly to their marketing performance, while others can affect both. Table 1-2 shows
some of these collective action initiatives. Although this is not an exhaustive list, it does
highlight many of the collective action measures farmer groups in this study attempted to
undertake to improve their market situation. As for the benefits accrued from such actions, these
can include: increased incomes through increased sales and or higher profit margins; a more
reliable or steady income flow; and enhanced food security.
Hypotheses derived from the collective action literature include:
*Hypotheses 4a 4f: Farmers groups will be better positioned to improve their marketing
performance if the group has some or all of the following characteristics:
4a Small group size
4b Past successful experiences
4c Interdependence among group members
4d Heterogeneity of endowments
4e Homogeneity of identities and interests
4f Low levels of poverty
Table 1-2. Collective action initiatives and improved performance
Types of collective action initiatives Improving production Improving marketing
performance (input) performance (output)
Access information on farming technology and
Lower cost of production through labor sharing
Lower input costs through bulk purchasing
Access extension and financial services and
training (from government, NGOs)
Access to contract farming arrangements
Access to formal and informal sources of credit
(e.g. rotating credit schemes, micro-finance
institutions, commercial banks).
Increase bargaining power vis-a-vis processors,
traders, retailers 2
Become involved in more post-harvest value-
added activities z
Lower transportation costs
Reduce transaction costs by establishing long-
term relationships with market intermediaries
Access information on prices, quality of goods,
and consumer preferences
Adapted from Heinemann (2002, pp. 1-2)
Cultural materialism is an important final addition to the theory section because it
encourages the application of a scientific approach to find the ultimate drivers of a change
process. Marvin Harris (1979), its founder, endeavors to bring forth a scientific model to the
study of culture. He seeks "etic" rather than "emic" explanations for human behavior, being
more interested in analyzing the underlying rules and realities of societies that can be
scientifically measured, rather than studying a people's "point of view," such studies are seen
as smeared by subjectivity and unsubstantiated theoretical conclusions (Harris, 1979, p. 47).
The field of cultural ecology influenced Harris' cultural materialism perspective in terms
of his use of general systems theory and strong emphasis on analyzing techno-economic and
techno-environmental variables to explain cultural behavior and change. His work also parallels
the cultural ecologists in their refutation of cultural relativism, who argue for more cross-cultural
analysis (Mcgee & Warms, 1996).
Harris was also influenced by Marxist thought, but he relies upon a selective Marxist
analysis that rej ects the Hegelian notion of dialectics and class struggle, and instead, emphasizes
the importance of Marx' s theories on the modes of production and reproduction (Harris, 1979, p.
64). Harris insists on the primacy of modes of production and reproduction as the determining
reasons for human behavior and beliefs within a society. Harris calls this the infrastructure of a
sociocultural system. Two other specific domains exist within this paradigm: the social structure
and the superstructure. Under the mantle of modes of production and reproduction, the
infrastructure would deal with such issues as techno-environmental relationships, work patterns,
technology of subsistence, demography, ecosystems, mating patterns, fertility, and mortality.
The social structuree would include everything involved with the domestic economy (family
structuree, domestic division of labor, social organization, age and sex roles, enculturation,
education) and the political economy, which deals with such issues as political organization,
class, factions, discipline, war, and political socialization. The superstructuree consists of all the
ideologies, rituals, or symbols associated with the social structuree (Harris, 1979, pp. 52-53).
Within this paradigm, the infrastructure becomes the research priority because it is at this
level that culture interfaces with nature. His reasoning goes as follows:
Since the aim of cultural materialism, in keeping with the orientation of science in general,
is the discovery of the maximum amount of order in its field of inquiry, priority for theory
building logically settles upon those sectors under the greatest direct restraints from the
givens of nature. To endow the mental superstructure with strategic priority, as the
cultural idealist advocate, is a bad bet. Nature is indifferent to whether God is a loving
father or a bloodthirsty cannibal. But nature is not indifferent to whether the fallow period
in a swidden field is one year or ten. We know that powerful restraints exist on the
infrastructural level; hence it is a good bet that these restraints are passed on to the
structural and superstructural components (Harris, 1979, p. 57).
Thus, cultures can be studied in a scientific matter by measuring the immutable laws of nature
and seeing to what degree they affect humans. Harris' paradigm takes on a causal sequence by
his reasoning that the social structure and superstructure become dependent variables of the
infrastructure. In other words, change will occur in society on the social and then ideological
level when the modes of production and reproduction have changed, which summarily, is based
on the environment upon which it rests.
Although the infrastructure plays the determinate role, the social structure and
superstructurre are not relegated to passive domains. Harris' sociocultural system takes on a
diachronic functionalist model wherein interdependencies exist between the three domains
through system-maintaining and system-altering roles. The social structure and superstructurre
take on system-maintaining roles by resisting innovations occurring at the infrastructural level,
and thus attempt to preserve the prevailing sociocultural system. But, the social structure and
superstructurre can also take on system-altering roles by lessening resistance to infrastructural
changes and thus make adoption of new innovations occur more quickly and pervasively (Harris,
1979, p. 72). These ideas can be expressed more clearly by way of an example. Rapid
population growth in the rural areas has led to land intensification and consequently to declining
soil fertility. Population growth and declining soil fertility are infrastructural changes that will
impact the social structure and superstructurre of a sociocultural system. One on hand, the social
structure and superstructurre may attempt to resist these changes (e.g., farmers refuse to accept
new innovations appropriate to land intensification practices) in an effort to maintain the
sociocultural system they have known so well. On the other hand, the social structure and
superstructure may lesson this resistance (e.g., farmers form groups and obtain training on soil
management practices; the positive results encourage other farmers to form groups to access the
same training) and play a more system-altering role that quickens the infrastructural changes
that bring about a new sociocultural system
Putting cultural materialism into the study context. Given the fact that the primary
purpose of this study is to identify the underlying factors that enable farmers to improve their
market situation, the study needs to give due consideration to the infrastructure domain since it
should play, according to the cultural materialists, the determining role in whether farmer groups
succeed or fail to improve their market situation. The NGOs in this study are trying to create a
culture of entrepreneurship, with the fundamental message being that farmers need to produce
for markets rather than trying to market what they produce. This means that farmers and farmer
groups need to take more chances and become less risk-adverse. The NGOs hope to instill this
behavior through capacity-building training activities. Essentially, this planned change initiative
is utilizing the classic education model of social change, in which the basic premise is that the
way a person behaves can be altered by changing the way she or he thinks (Bernard, 1995, p.
24). From cultural materialist perspective, this assumes that changing the superstructure the
domain of ideology will lead to lasting culture change in terms of farmers' attitudes, values and
beliefs concerning the market. Cultural materialists would argue that unless there are changes to
the infrastructure, or at the very least, an enabling infrastructural environment within which
farmers can put into practice what they have learned, this model of social change will have little
lasting impact. Although cultural materialists are often more concerned with culture change
phenomena in the longer term, the basic fundamentals underlying this theory can still provide an
appropriate lens to research and question the utility of this planned change initiative. It also
provides an overarching and final hypothesis to this study:
*Hypothesis Sa: Farmer groups with infrastructural advantages over their competitors (i.e.,
better agro-ecological and physical market access conditions) will be better positioned to
improve their marketing performance.
By concluding the literature review section, attention now turns to the next chapter where
background information is provided on the development program under study (AMSDP) and the
specific region where the research was conducted.
PROGRAM AND STUDY AREA BACKGROUND
This chapter opens with background information on the country of Tanzania. This is
followed by a discussion of the planned change initiative under study, the Agricultural Marketing
Systems Development Program (AMSDP). After a general overview of AMSDP's project
components and the primary stakeholders, the chapter continues with a discussion of two local
NGOs tasked with implementing one of the proj ect components in the Arusha and Kilimanj aro
regions of northern Tanzania. With these two regions as the designated study areas of the
research, background information is provided on the demographics, farming systems, agro-
ecological and market conditions of the study area within the context of the farmer groups
participating in AMSDP. The chapter concludes with a discussion of the specific training
activities conducted by the two local NGOs for the farmer groups.
This study takes place in the east African country of Tanzania. From the late 1960s to the
early 1980s, the Tanzanian government promoted a socialist model of economic development.
The country's first President after independence, Julius Nyerere, called for the nationalization of
factories and plantations, emphasizing a social and economic policy of self-reliance,
egalitarianism, and communalism. This included the collectivization of the country's
agricultural system, known as Ujamaa a Swahili term meaning familyhood (Haan, 2000).
By thel980s, the country's trade deficit increased and foreign capital inflows declined.
The Ujamaa policy was not showing the anticipated results, but in fact resulted in low product
quality and an inefficient economy. The government responded with the self-guided National
Economic Survival Program in 1981, and began official structural adjustment programs in 1982.
By 1984, economic reforms had raised agricultural producer prices, reintroduced cooperative
unions for crop marketing, frozen the budget deficit, depreciated the shilling by 40 percent,
raised government wages, and eliminated consumer price subsidies on maize (Wobst, 2001).
Still, the Tanzanian government felt external pressure from international organizations to do
more. The government initiated two Economic Recovery Programs, the first in 1986 and the
second in 1989, that included such measures as devaluing its local currency, removing price
controls, and opening the economy to international trade (IFAD, 2002).
By the early 1990s, the macro-economic reforms were starting to yield some positive
results, with the economy growing 4 to 5 percent per year in the second half of the 1990s, and by
5 to 6 percent in more recent years (Minot, 2006). However, the impact of these macroeconomic
reforms on rural areas has been hotly debated, with one side arguing that market liberalization
brought new opportunities to farmers through high-value agriculture and export markets, while
the other side arguing that these reforms only widened the gap between the poor and rich,
leaving many farmers worse off in the process (Minot, 2006).
Regardless of this debate, the fact remains that Tanzania is one of the poorest countries in
the world, with over a third of Tanzania' s population of 37 million living below the poverty line.
The average life expectancy of Tanzanians is 45 years, and the adult prevalence of HIV/AIDS is
estimated at 9 percent (CIA, 2007). As with many African countries, Tanzania's agricultural
sector is the backbone of the economy, employing 80 percent of the labor force and contributing
45 percent of the gross domestic product. The sale of agricultural commodities accounts for 70
percent of rural incomes, with most agricultural production coming from the smallholder farmer
section, and over 50 percent of the farms being less than one hectare (USAID, 2007). As
discussed in chapter 1, African governments like Tanzania, with backing from the donor
agencies, are committed to reducing rural poverty through the promotion of policies and
development programs that focus on boosting the agricultural productivity of smallholder
farmers. This provides the backdrop for the formulation and implementation of the development
program under study.
Agricultural Marketing Systems Development Program (AMSDP)3
In December 2001, an agreement was reached between the International Fund for
Agricultural Development (IFAD) and the government of Tanzania to provide approximately 42
million US dollars for the Agricultural Marketing Systems Development Program (AMSDP).
IFAD is a specialized agency of the United Nations (UN) and its mission is to "enable the rural
poor to overcome poverty." IFAD's activities are guided by three strategic objectives:
* To strengthen the capacity of the rural poor and their organizations
* To improve equitable access to productive natural resources and technologies
* To increase access by the rural poor people to Einancial services and markets
In terms of the Einancial specifies of AMSDP, IFAD is loaning the Tanzanian government
approximately 16.34 million dollars and the African Development Fund, as the maj or co-
Einancier of the program, is loaning 14.46 million dollars.4 Other donor participation includes
Ireland Aid (with a grant of 1.1 million dollars), with USAID, France, and the Netherlands
providing grants either directly to the government or through non-governmental organizations of
approximately 4.49 million dollars. The government of Tanzania is contributing 4.22 million
dollars and proj ect beneficiaries are expected to contribute about half a million dollars. IFAD is
3 Unless otherwise cited, all information from the section entitled, "Agricultural Marketing Systems Development
Program," comes from the IFAD website, under the heading, 4boutlE4D [online] at the following website address:
b1llp w il it .ifad.org/governance/ifad/ifad.htm
SMembership to IFAD is open to UN member states. Most of IFAD's resources are made available to low-income
countries by providing highly concessional loans (40 year repayment term with 10 year grace period). The
resources to finance IFAD loans and grants come from three sources: 1) donor member contributions: 2) investment
income; and 3) reflows from past loans. Under most loan agreements, IFAD finds partner aid organizations, such as
bilateral agencies (USAID, DANIDA, DFID, etc.), multilateral agencies (IMF, World Bank, other UN agencies),
and international NGOs, to co-finance loans and make other program contributions. Recipient countries and project
beneficiaries are also expected to make contributions.
recognized by the other co-financiers of the program as the official appraising institution, and
thus carries additional monitoring and evaluation functions (IFAD, 2001, p. 8).
The overall goal of the AMSDP is "to increase the income and food-security situation of
the rural poor in the Northern and Southern Marketing Zones of the United Republic of
Tanzania." It proposes to do this through a comprehensive rehabilitation of Tanzania' s
agricultural marketing system by "making rural markets work better and empowering
smallholders within them." Accordingly, AMSDP expects to:
* strengthen about 1,000 producers groups (PGs) to enable them to have a better bargaining
position and more leverage on policy formulation, identification of marketing opportunities
and price negotiations for both inputs and outputs;
* assist the government of Tanzania in rationalizing the existing policies relating to
regulation, taxation, and the exchange rate so that it can contribute to improved marketing
system efficiency as a whole;
* improve market infrastructures through construction and rehabilitation of 700 km of rural
roads, 200 km of access roads and 30 market centers, and through Einancing for
* strengthen capacity of the Ministry of Cooperatives and Marketing (MCM) to collect,
compile and disseminate agricultural marketing information;
* help PGs, grass-roots institutions (GIs), traders and processors to access inventory and
capital loans from the commercial banks as required for promotion of marketing activities;
* establish and strengthen both vertical and horizontal linkages among PGs, GIs, processors,
local marketing chains and exporters (IFAD, 2001, p. vi)
These tasks will be accomplished under Hyve proj ect components: (1) Policy Development;
Support; (2) Producer Empowerment and Market Participation; (3) Financial Market Services;
(4) Rural Marketing Infrastructure Support; and (5) Program Organization and Coordination
(IFAD, 2001, p. 4).
While these proj ect components are interrelated and interact at different levels, my
research has focused primarily on the second component: Producer Empowerment and Market
Participation. This component is designed to assist producer groups, grass-roots institutions, and
small and medium-scale traders and processors to:
* strengthen their social, organizational and financial structure to enable them to participate
effectively in marketing activities;
* empower them technically through improved market information, extension and research
services to enhance their leverage and bargaining power vis-a-vis more organized agro-
processors, wholesalers, and exporters; and
* establish appropriate vertical and horizontal linkages with formal market players to
minimize the risks and uncertainties associated with the lack of assured market
opportunities and prices for both inputs and outputs (IFAD, 2001, p. 6).
Almost a third of the program funding (approximately 12 million US dollars) is being used to
achieve this component. Development-oriented national and international NGOs are carrying
out most of the tasks listed above (IFAD, 2001, p.9).
The research site for this study is located in the "Northemn Marketing Zone" situated in the
northern highlands region of Tanzania. In the initial stage of the program, eight NGOs (four in
the southern and four in the northern zone) won bids to work on the AMSDP project. The NGOs
- or partner agencies (PAs) as they are referred to in the program have been charged with two
primary responsibilities: (1) establish and/or strengthen local-level agricultural producer groups
and (2) establish marketing links between these producer groups and other market intermediaries
(i.e., traders, processors, wholesalers, etc.).
The AMSDP Stakeholders
The primary stakeholders for the proj ect component Producer Empowerment and Market
Participation are shown in Figure 2-1. This is followed by Figure 2-2, which shows the basic
communication and budgetary channels for the stakeholders in Figure 2-1.
Program Steering Committee (PSC): The PSC has overall responsibility of coordinating AMSDP.
It is chaired by the Permanent Secretary of the Prime Minister's Office (PMO), and the PSC
members include: relevant government ministries (i.e., Ministry of Cooperative and Marketing,
Ministry of Agriculture and Food Security), Tanzania Chamber of Commerce, Industry and
Agriculture (TCCIA), participating banks, and national NGOs and farmer cooperatives representing
the interests of smallholder agriculture.
Program Coordination Unit (PCU): The PCU officials are responsible for the practical day-to-day
operations of all five components of the AMSDP. The PCU has two sub-offices located in each of
the marketing zones.
District Development Committee (DDC): The DDC is composed of locally elected representatives
of village, ward, and districts councils that are responsible for implementing district-level
development programs. This is not a creation of the AMSDP program. The DDC is working in
coordination with AMSDP and shares oversight responsibilities with the sub-offices of the PCU.
The DDC, as representatives of the district offices, are also responsible for approving the budgets of
the partner agencies.
Focal Area Advisory Group (FAAG): The FAAG is the primary communication link between the
District, the partner agencies, and the producer groups. This group is chaired by the District
Executive Director (DED) and its members include: partner agency field staff, district extension and
marketing officers, producer group representatives, and district branch members of the TCCIA.
FAAG carries the specific responsibility of monitoring the progress of program activities, including
budget oversight. All PA proposed expenditures must be approved by the FAAG on a quarterly
basis before further program activities within the district can continue. FAAG can also provide
technical assistance to producer groups as requested by PAs.
Partner Agencies (PAs): Eight PAs were originally selected by AMSDP to carly out the project
component involved with producer empowerment and market linkages. In the second year, another
twelve PAs were selected to cover the other regions of Tanzania. My research involves work with
two of these PAs situated in the northern highlands of Tanzania.
Producer Groups (PGs): These producer groups are composed of smallholder farmers and they are
the targeted beneficiaries of the AMSDP program. More detail will be given later on the
composition and the characteristics of these groups.
Figure 2-1. Primary AMSDP stakeholders (adapted from IFAD, 2001, pp. 11-12)
Figure 2-2. Communication and budgetary flows of primary stakeholders
Apart from the primary stakeholders, there is also a wide-range of market chain actors who
are key stakeholders in the context of this proj ect. They are not the target audience but some of
these chain actors have benefited from this intervention due to the linkages forged between them
and producer groups. Therefore, it is useful to point out some of the key market participants.
Mendoza (1995, pp. 260-261) provides a classification of the participants in the marketing
process, which is represented in Figure 2-3.
Producer: Enters the marketing channel the moment she or he decides what, when, and how much
Rural Assembler: The trader/transporter is the first link between the producer and other
intermediaries. Most important function is transportation and assembly of scattered rural produce
into a single load for further distribution.
Wholesaler: Concentrates the various, intermediate size loads and puts the product into large,
uniform units. Plays a significant role in price formation and accepts certain risks associated with
the transfer of property rights attached to the goods and services being bought and sold. May also be
involved specialized storage operations, transportation, and in general, the subsequent distribution
operations involving retailers.
Retailers: This includes supermarkets and other large-scale retailers who divide up shipments of
produce and sell it to consumers in small units. The basic function that they provide is bulk-
Other notable and influential market chain actors
Exporters and Importers
Government Institutions or Agencies
Producer and Consumer Associations
Other Organizations: Includes parastatals, marketing boards, and private companies with semi-
Figure 2-3. Market chain actors as other key stakeholders
Stakeholder partnerships. The main tasks of the partner agencies (PAs) are to train
existing or newly formed producer groups in a number of capacity building and marketing skills
measures, and where possible, to establish sustainable market linkages with other market chain
actors. The PA relationship with the district-level stakeholders (i.e., district development
committee and focal area advisory group) was crucial from the beginning given the management
and fiscal responsibilities afforded to the districts. These responsibilities included: selecting the
focal areas where the program should be implemented with new producer groups (with input
from the PA); overseeing the general progress of the proj ect; and most importantly, approving all
Primary market chain actors
budgets and allocating required funds to the PAs. From the beginning, this "partnership"
between the PA and district officials was tenuous at best and culminated in much confusion and
conflict between these two stakeholders for reasons that will be highlighted later.
Under the program contract agreement, each PA was expected to train 20 producer groups
and establish sustainable market linkages between these groups and other chain actors on an
annual basis. This was to be done with a new cohort of groups annually over a three-year period,
bringing the total to 60 capacitated producer groups linked to markets for each PA. Furthermore,
if judged to fulfill their contractual agreement, the PAs would be hired on for another round to
continue the process, with a grand total of 120 participating producer groups per PA.
Research Sites and Partner Agencies
AMSDP started in January 2004 with eight PAs implementing the program with producer
groups in eight regions located in the northern and southern parts of Tanzania. The partners for
this study were two local NGOs participating in the role of PAs in the Arusha and Kilimanj aro
regions. They are the Traditional Irrigation and Environmental Development Organization (TIP)
and FAIDA Market Link (FAIDA).6 During the research period, both PAs worked in one
specific district in each region. TIP worked with producer groups in the district of Armeru, in
the Arusha region, and FAIDA worked with producer groups in the district of Hai, in the
Kilimanj aro region. Although working in two different regions, the two districts actually border
each other. Ar-umer-u district surrounds the slopes of Mt. Mer-u Africa' s fifth highest mountain
(4,533 meters) and Hai district is situated on the plains and lower slopes between Mt. Mer-u and
5 AMSDP did not specify the exact number of farmers that should be in each group, but generally speaking, the
farmer groups ranged in size from 20 to 40 members. If we assume an average of 30 members per group, the project
expected each PA to work with approximately 3,600 smallholder farmers. AMSDP ended up contracting twenty
partner agencies to implement the project. Maintaining a similar estimate of 30 members per group, AMSDP
expects approximately 72,000 smallholder farmers to participate in this AMSDP project component.
6 Faida means profit in Swahili.
Mt. Kilimanj aro Africa' s highest mountain (5,895 meters). The districts, and my research
sites, are pointed out in Figure 2-4.
Figure 2-4. Map of Tanzania with designated research sites (adapted from IFAD, 2001, p.3)
The soils on the slopes of Mt. Meru and Mt. Kilimanj aro are volcanic and have high-to-
medium levels of fertility. The regions have a bimodal rain pattern, with the "long" rains
occurring between March and May and "short" rains between October and December (Haan,
2000, p.76). On average, Hai district gets between 500 to 800 millimeters (mm) of rain per year,
and Arumeru gets somewhat more with averages for the district over 1000 mm of rain per year
(Mansoor & Piters, 1999, 21). Smallholders traditionally plant food crops during both seasons,
but in recent years, the short rains have been unreliable, starting later than usual and often
without the same level of rainfall as in the past. This is especially true for groups in lower
elevation areas, that includes many of the groups in Hai district living on the plains between the
The total area of Ar-umer-u is 2,896 square kilometers (sq. km), with a population of
516,814 and a population density of 181 people per sq. km. Hai is a smaller district covering a
total area of 2, 186 sq. km, with a population of 259,958 and a population density of 1 17 people
per sq. km. (R & AWG, 2005, p.104). Out of the 199 districts in Tanzania, both Hai and
Ar-umer-u are in the top twenty districts with the lowest levels of poverty. In Ar-umer-u district, 18
percent of the households are below the poverty line and 22 percent of the households are below
the poverty line in Hai district (R & AWG, 2005, p. 75).7 Further information on demographics
and farming systems within the context of the study's sample of farmer groups will be discussed
later in this chapter.
Traditional Irrigation and Environmental Development Organization (TIP)
TIP has been an officially registered NGO in Tanzania since 1999. Before becoming its
own entity, TIP was a proj ect funded by the Netherlands Development Organization (SNV). The
proj ect was undertaken to improve traditional irrigation systems in various regions of Tanzania.
SNV emphasized the inclusion of participatory approaches and gender awareness in dealing with
the many water user groups established under its watershed management schemes. After
becoming a NGO, TIP continued to build a reputation as one of the more experienced
organizations dealing with participatory development and gender within the country, particularly
in the field of natural resource management. TIP also began to build partnerships with other
donors and research institutes. Of particular importance to this study, TIP developed a
relationship with the International Center for Tropical Agriculture (CIAT), that provided
SIn Tanzania, the basic needs poverty line for 2000/01 was set at 262 Tanzanian Shillings (TShs) per adult
equivalent per day, which is the equivalent of US$ 0.33 (R & AWG, 2005, p.54).
technical assistance in the Hield of agro-enterprise development a programmatic component that
TIP added to its organization in 2002. By the start of the AMSDP proj ect in 2004, TIP had only
implemented an agro-enterprise approach with a few farmer groups, and with mixed success.
Thus, it started the AMSDP with a great deal of technical capacity thanks in large measure to
CIAT but not with a great deal of practical experience.
FAIDA MaLi (FAIDA)
FAIDA shares a somewhat similar history to TIP. From 1994 to 2003, FAIDA was also a
SNV-funded proj ect. Since its inception, it has been involved in facilitating linkages between
smallholder farmers and agricultural companies. After becoming a Tanzanian-owned and
operated non-profit entity in 2004, FAIDA has continued to play a leading role in developing
market linkages between local growers and agricultural business at both the regional and
international level. At the start of AMSDP, FAIDA brought in a large amount of marketing
experience, as well as a reputation that was recognized by agricultural businesses. However, as
will be explained later, FAIDA' s particular strengths could not be utilized fully under the
confines of AMSDP.
As with any NGO, TIP and FAIDA have their own development approaches or processes.
TIP's entry point for agro-enterprise development starts at the farm level by either working with
existing farmer groups or establishing new groups. Utilizing a bottom-up participatory approach,
TIP works closely with groups to develop their capacities through skills training and then to
strategize together on how best to exploit possible market opportunities. Through TIP-facilitated
market research and analysis, groups may decide to develop an agro-enterprise based on existing
products (crops they already grow), or try new agricultural products that show greater market
potential but may be riskier.
The entry point for FAIDA is different and starts by engaging agricultural companies
interested in finding farmers to grow particular products for them. Once FAIDA has clarified the
needs of the company, it finds the most appropriate agro-ecological and infrastructural
conditions suitable for the crop in question. In this top-down oriented approach, only when these
conditions have been met are farmers who can satisfy the demands of the agricultural company
found. FAIDA then brings the parties together and provides technical assistance that may
include: contract negotiations; skills training (i.e., group dynamics, crop husbandry practices,
business skills training, etc.); conflict management; as well as other areas to ensure that both
company and farmers maintain a good working relationship.
In terms of exit strategies, both TIP and FAIDA claim that ideally they remove themselves
from the process once the market linkage is well established and sustainable. However, they
admit that in practice, their exit strategy is usually tied more to diminished funds to support the
linkage than to issues of sustainability.
Group Selection Criteria
Based on IFAD's mandate to target the rural poor, AMSDP provided a list of criteria to be
used by the PAs and District Office in their selection of proj ect areas and farmer groups. The
criteria list included the following:
* The proj ect participants should already be involved in farming and/or trading.
* The average cropped area of smallholders within producer groups should not exceed 2
* About 50 percent of group members should be from the poorer households with an income
below the poverty line.
* At least 40 percent of the beneficiaries of the group should be women and female-headed
* At least one crop with high market potential has been identified for cultivation and local
* Access roads to areas of production exist and within reasonable reach from the main access
* There should exist a minimum critical mass related to the proposed marketing activity with
similar marketing aspirations.
* The group members live and operate in walking distance of one another and are willing to
collaborate in specific activities (such as training).
Even though the PAs are expected to use these guidelines in their selection process, both FAIDA
and TIP admit it was impossible to meet all these criteria because of the reality facing
smallholder farmers. The program aims to work with poor farmers who have crops with market
potential and good access to roads. The problem is that in reality most poor farmers do not have
crops with market potential and generally live in areas with poor road access. Another reason
the PAs had difficulty implementing these criteria were due to local politics. Each District
Office (DO) had its own agenda when deciding which areas and groups should be chosen for the
proj ect, and it had more to do with pleasing the DO and its local constituents than in meeting
requirements of a criteria list.
In Ar-umer-u district where TIP was operating, the District Office chose the most politically
expedient path by selecting a producer group from each ward.8 In some cases, an existing
producer group was chosen to represent the ward. In other cases, smaller groups were told to
provide representatives to establish a ward level group; in cases where few groups existed,
individual smallholders were chosen to represent their village. The decision by the District
Office to spread the proj ect over the entire district may have pleased a wide array of constituents,
especially during a presidential election year, but it provided a logistic and financial nightmare
for TIP. The final set of groups that were selected met some of the criteria in terms of crops and
areas with market potential, but as a consequence, TIP had only a few groups with members who
fit the criteria of coming from poorer households.
SPolitical divisions in Tanzania, starting from the regional level (i.e., Kilimanjaro or Arusha region) to smaller
political units, include: District Ward Village Hamlet Ten Cell (i.e., blocks of ten households in the hamlet).
FAIDA also had to struggle over the selection process with the government authorities in
Hai district. In this case, the District Office decided that AMSDP would be a good follow up
proj ect to the Participatory Agricultural Development Proj ect (PADEP), another government-led
proj ect that was currently operating in the district. From 2002 to 2006, PADEP established a
number of village-level producer groups and provided production-oriented training, along with
two years of agricultural input loans. PADEP also carried out some infrastructural improvement
proj ects at the village level, dealing mostly with repairing traditional irrigation schemes. Given
the fact that PADEP was production-related and AMSDP was market-oriented, it seemed to be
an excellent match. Logistically, it would also be easier to conduct training activities since most
of the farmer groups were located in just four villages. The problem for FAIDA was that these
groups had little potential in terms of marketable crops, and most of these villages were not
located in areas with the appropriate agro-ecological conditions or proper infrastructure that
would attract agricultural companies. These circumstances made it difficult for FAIDA to
exploit its greatest strength its wide network of agricultural companies willing to do business
with smallholder farmers.
Selected Farmer Groups
For the first year of the proj ect, TIP worked with 18 groups and FAIDA worked with 20 groups. 9
In TIP's case, only two groups were located in the same ward, whereas FAIDA had 16 out of its
20 groups located in four villages. Table 2-1 shows group size and composition that FAIDA and
TIP worked with at the beginning of the project. These groups provided the interview sample for
9 My Study sample includes only 16 of the 20 FAIDA groups participating in AMSDP. Two groups were omitted
because they were trader groups (i.e., businessmen and businesswomen engaged in trading, and not farming, as their
primary livelihood) and did not share enough similarities with the other farmer groups to make intra- and inter-group
comparisons valid. The two other groups were not included because they declined to participate in the study.
Table 2-1. Size and composition of TIP and FAIDA groups
FAIDA TIP Sample total
Sample size 16 18 34
Newly formed groups 5 9 14
Existing groups 11 9 20
Newly formed 5 6 11
Existing 4 1 5
Village 16 9 25
Ward 0 9 9
Age of groups (years)
Mean 4.7 3.5 4
Range 2 -16 2- 8
Due to the PADEP proj ect, FAIDA was working with groups that had already been established.
But, five of the PADEP groups were formed only months before FAIDA entered. Thus, even
though PADEP and FAIDA were working at the same time, almost all the group capacity
training responsibilities were relegated to FAIDA. For TIP, nine of the groups were newly
formed, while the other nine groups existed prior to program and consisted mostly of village
level groups chosen to represent the ward. There were also a number of farmer groups that were
combinations of smaller groups, as shown under the "combined groups" heading in Table 2-2.
The "existing groups" under this heading were smaller groups that came together on their own
accord to increase scale and access to service providers. There were also a number of combined
groups that were specifically brought together by the PAs, which is represented in the table by
"newly formed" groups. Over half of the FAIDA groups (56%) were a combination of smaller
groups, whereas this represented seven of the TIP groups (39%). Also, both FAIDA and TIP
worked with relatively "young" groups, with an average age of 4.7 and 3.5 years, respectively.
Table 2-2 shows the membership of these groups disaggregated by the sex of the members.
Within the study sample, 16 of the FAIDA groups had a total of 507 members and TIP worked
with 18 groups that totaled 678 members.
Table 2-2. Group membership by partner agency and sex
Members by sex FAIDA TIP Total
Females 277 272 542
Males 230 406 636
Total 507 678 1,185
In terms of gender, more than half of the members in the FAIDA groups are female. The higher
number of female membership can be explained by the fact that three of the FAIDA groups were
female-only groups. As for TIP, it worked with one female-only group and males tended to
dominate more of the membership in the other groups. This is reflected in the table where 406
out 678 (60%) of the members are men. Neither FAIDA nor TIP worked with any male-only
Demographic Characteristics and Farming Systems of Farmer Group Members
Out of the 1,185 farmer group members, a sample of 3 88 members was given self-
administered questionnaires, with 189 members from FAIDA and 199 members from TIP. The
sample represents 33 percent of group members (see sample methodology in chapter 3). The
questionnaire included demographic information in such areas as education, ethnicity, age, and
religion. The demographic data collected from the 388 group members who filled out the
questionnaire are represented in the following tables. Table 2-3 shows members by years of
schooling for each of the PAs. To achieve a basic level of math skills and literacy in Swahili,
most school children need to complete at least 4 years of schooling. Literacy is high in the
regions of Arusha and Kilimanj aro and within the districts of Hai and Arumeru. In Arumeru
district, adult literacy stands at 79 percent (84% for males and 74% for females) and the rate in
Hai district is 86 percent (89% for males and 83% for females). The study sample reflects an
even higher rate, with roughly 92 percent literacy among the group members working with
FAIDA and TIP.
Table 2-3. Years of schooling of producer group members
Years in school FAIDA TIP Total
0 16 6 22
1-3 4 4 8
4 -7 143 144 287
8 -11 22 42 64
12 -13 4 3 7
There are some differences when comparing FAIDA and TIP group members. Members from
FAIDA groups have more illiterate members and fewer members with secondary education
(eight years of schooling or more). Approximately 11 percent of FAIDA group members are
illiterate compared to only 5 percent for the TIP group members. Likewise, TIP members have
achieved higher levels of education, with 23 percent of TIP members achieving at least 8 years
of schooling and a few members completing advanced secondary school levels. FAIDA also had
a few members complete advanced secondary school, but overall had significantly less members
with secondary education, with only 14 percent of its members with 8 years of schooling or
Table 2-4 shows the religious affiliations of the members of the FAIDA and TIP groups.
The majority of the members of both groups are Christian. FAIDA is also working with the
maj ority of Muslim members represented in the study sample. Neither TIP nor FAIDA worked
with groups that had exclusively Muslim members, but 16 out of the 18 TIP groups were
exclusively Christian. FAIDA worked with primarily mixed groups, which included five
Muslim-maj ority groups and eight Christian-maj ority groups. In Hai District where FAIDA is
working, it is common for villages to have a mix of both Christians and Muslims and there
appears to be little animosity or conflict occurring along religious lines. As this study sample
shows, it is also very common for groups with mixed-religious backgrounds to cooperate and
engage in collective action initiatives.
Table 2-4. Maj or religious affiliations of producer group members
Religious affiliation FAIDA TIP Total
Religion by members
Traditional 5 0 5
Muslim 67 9 76
Christian 117 190 307
Religion by group
Muslim-maj ority 5 1 6
Christian-majority 8 1 9
Evenly-mixed 1 01
Christian-only 2 16 18
Muslim-only 0 0 0
Table 2-5 shows the age of the FAIDA and TIP group members. The majority of members
for both PAs are in the age range of 40 to 49 years, which makes up 34 percent of the total
membership sampled in this study.
Table 2-5. Age range of producer group members
Age range FAIDA TIP Total
20 29 18 13 31
30 39 41 62 103
40 -49 60 72 132
50 -59 38 34 72
60 -69 22 14 36
70 and older 10 4 14
Both PA groups also compose a fair portion of "younger" members, with 3 1 percent of FAIDA
group members and 38 percent of TIP group members under the age of 40.
Table 2-6 shows the maj or ethnic affiliations by PA groups. The overwhelming maj ority
(70%) of FAIDA group members are Chaga, whereas the Meru peoples make up a slight
majority of TIP group members at 53 percent. The Chaga are the dominant ethnic group in the
Kilimanj aro region and the Meru and Arusha peoples make up the dominant ethnic groups in the
Table 2-6. Maj or ethnicity of producer group members
Ethnic group FAIDA TIP Total
Arusha-Masai' o 17 63 80
Meru 3 106 109
Chaga 132 14 146
Pare 23 4 27
Others 14 12 26
The greater ethnic diversity within the TIP groups can also be attributed to the fact that Arusha
city is near the proj ect area and it has for sometime attracted people from all over Tanzania
seeking work opportunities, particularly in the tourist service sector, tanzanite mining, and agro-
industries (e.g., sisal production in the past and cut flower production more recently).
Table 2-7 provides a breakdown of the commodity types by PA farmer groups.
Commodity types refer to the crops or livestock that group members grow and were put forward
as possible commodities to promote as a viable agro-enterprise. The commodity types also
represent most of the existing farming systems located in the study area. The cereals/legumes
category typifies a mixed-farming system with corn intercropped with legumes, such as beans or
pigeon peas. Many of the farmers in this system also plant sunflower as a windbreak, as well as
for cooking oil for home consumption. This farming system is generally located on the lower
slopes and plains surrounding Mt. Mer-u and Mt. Kilimanj aro.
'0 As shown in Table 3-6, I have placed the Arusha and Masai under one category. Even though these are in fact two
distinct ethnic groups, their histories and cultural heritage share a great deal in common. The Arusha peoples are
farmers who have been cultivating the slops of Mount Meru for over two centuries; the Masai peoples are cattle
pastoralists living on the plains of northern Tanzania and western Kenya. Over the generations, these ethnic groups
have developed an interdependent relationship that has helped them cope in times of crises (i.e., drought periods,
disease epidemics, civil strife). Many Arusha have depended on the Masai for cattle and protection from marauders,
and the Masai have often counted on the Arusha for foodstuffs during times of drought and livestock disease. These
economic and social interdependencies, as well as the exchange of cultural customs and rituals (mixed marriages,
age groups, rites of passage, etc.) have blurred the lines of ethnic distinction. This is particularly true when it comes
to "settled" Masai that now make their livelihood primarily from farming. There is little difference between these
Masai and the Arusha peoples. These are the peoples in my study and the reason that Arusha-Masai is one ethnic
category only. A more detailed account of the relationship between the Masai and Arusha peoples can be found in
11 In most cases, farmers take the sunflower seeds directly to a processor, who will charge them to process the seeds
into cooking oil for home consumption. Any surplus seeds not needed for home consumption are usually sold to
traders or wholesalers at farm-gate prices.
Table 2-7. Commodity types by PA farmer groups
Commodity types FAIDA groups TIP groups Total
Cereals/Legumes 9 5 14
Vegetables/Fruits 5 5 10
Coffee 0 3 3
Dairy (cows) 13 4
Chickens 0 1 1
The vegetable/fruit category typifies a mono-crop horticultural farming system wherein
these crops (e.g., tomato, cabbage, carrot, onion, lettuce, cucumber) are grown in individual plots
that often utilizes canal irrigation s, and in a few cases spray irrigation. This farming system is
generally located on lower to mid-level slopes of the mountain ranges. The coffee category is
another mixed farming system where coffee is intercropped with "cooking" bananas, which
provide shade for the coffee and also are used for income and home consumption purposes. The
coffee/banana farming system is located on the upper slopes of Mt. Meru and Mt. Kilimanj aro.
The rice category represents a mono-crop farming system of paddy-rice that utilizes canal
irrigation schemes. This farming system is located on the lower slopes of the mountain ranges,
but few farmers in this study sample or study area are involved in this farming system. The
livestock category complements the other farming systems as a way to supplement income and
home consumption needs.
Returning to Table 2-7, the maj ority of the FAIDA groups belong the cereal/legume
farming system. Both TIP and FAIDA have five groups each that are involved in vegetable
production, as well as one group each involved paddy-rice production. Also, three TIP groups
belong the coffee/banana farming system. The PA group differences in commodity types and
farming systems are largely due to the selected proj ect areas. On one hand, Hai District where
FAIDA works is located on the plains and lower slopes between Mt. Meru and Mt. Kilimanj aro,
which explains the greater number of cereal/legume farming systems. On the other hand,
Arumeru district where TIP works is situated along the slopes of Mt. Meru, which explains the
greater diversity of farming systems (i.e., rice and cereals/legumes on the lower slopes;
vegetables/fruits on the mid-level; and coffee/banana on the higher slopes).
The PA Training Activities for Farmer Groups
When the PAs were first contracted to train farmer groups, there were few guidelines from
AMSDP. Thus, each PA was given a certain level of flexibility to decide the subject content and
how to conduct the different training activities. Nonetheless, FAIDA and TIP had more
similarities than differences in their training content and style. There were a number of training
activities that can be classified as group strengthening training. Although the timing or grouping
of these training activities may have differed, essentially both PAs covered the following subj ect
areas with their farmer groups:
* Group registration
* Communication and group dynamics
* Group record keeping
* Administrative and Einancial management
* Setting up bank account
* Establishing action plans
* Establishing savings and credit cooperatives
Each PA also provided business awareness and marketing skills training that included such topic
* Farm record keeping
* Cost-benefit analysis
* Accessing market information
* Negotiating prices
* Choosing an enterprise and Einding potential markets/buyers.
* Contract farming
Beyond these training activities, each PA also arranged study tours for group representatives to
visit farmer-formed saving and credit cooperatives (SACCOs). Some of the groups also had the
opportunity to go on production-related study tours to visit other farmers growing a particular
crop with market potential or to visit agriculture research institutes to get training on modern
There were also a number of training activities that each PA did that were different from
each other. For example, TIP incorporated training activities that covered important social
issues, such as gender, HIV/AIDS awareness, and environmental conservation. TIP also
included a two-day training activity on participatory market research visits where group
representatives had the opportunity to visit area food markets and supermarkets and talk to a
range of other chain actors, such as wholesalers, brokers and retailers. Contacts were established
with these chain actors and TIP used these meetings to encourage the farmer groups to seek
market information through these new contacts.
FAIDA also had a few training activities that differed from TIP. Notably, FAIDA
allocated more training time to developing higher-tier, or apex, organizations among the groups
located in the same village. This was done to get groups to form a savings and credit cooperative
to serve the groups, as well as other village members. FAIDA also had a different strategy for
improving the farmer groups' access to market information. Due to the concentration of groups
in Hai district, AMSDP instructed FAIDA to set up market information boards in each village by
which weekly information of the maj or commodities from the local markets could be posted.
FAIDA encouraged all groups to fund the construction of these boards, and after a few months of
collecting the requisite money, market information boards were erected in the four villages
where the maj ority of FAIDA groups were located.
Study Entry Point
Field research for this study began in April 2005, a year after the start of AMSDP.
According to the program plans, each PA should have finished the training activities and market
linkages with the first round of 20 producer groups and started the selection process with the next
20 groups. This was far from the case as FAIDA and TIP were still conducting group capacity
training and had made only a few market linkages. There are a variety of reasons for this but a
notable one involved conflicts with the District Offices over budgetary issues. On a quarterly
basis, the District Office approved the PA' s budget for its next quarter of work with the producer
groups. There was much disagreement between the PA and District Offices over what activities
should be done and on how much these activities should cost. On several occasions, both PAs
stopped working with their groups because the District Offices would not approve their budgets.
One such occasion occurred at the beginning of this research in January 2005. The issue was not
resolved for two months, and PA activities only resumed the middle of April.
Once the program resumed, the author began to accompany the PAs to the field to observe
the implementation of the training activities. FAIDA had finished most of its group
strengthening and marketing skills training and was focusing on establishing saving and credit
cooperatives (SACCOs). This included training the group members on organizational and
financial management skills, as well as holding an additional training for group leaders on how
to formulate an apex organization for the purposes of setting up village-based SACCOs.
TIP had also finished most of its group strengthening and marketing skills training and was
proceeding with the participatory farmer market research activity. It was also undertaking
training and study tours of SACCOs. Observing the training activities was an opportunity to see
how each PA operated in regard to its training and management style, as well as a means of
being introduced to many of the groups whom would be interviewed later.
The next chapter introduces the research methods of this study. This includes discussions
on the overall research design, study time-frame, data collection methods, sample representation,
operational definition of the primary dependent variable group marketing performance as
well as definitions and values for independent variables that can affect a group's ability to
improve its market situation.
This chapter examines the Sustainable Livelihoods Approach (SLA) as a basis for the
development of a conceptual model specific to this study to understand and analyze the
planned change initiative. After presenting this conceptual model, the chapter continues with a
discussion of the research design. A description of the data collection methods and study sample
follows. The rest of the chapter is concerned with developing the operational definitions and
values for the dependent variable group marketing performance followed by the description
and values of the independent, or explanatory, variables. The chapter concludes by revisiting the
research obj ectives and matching the hypotheses with the appropriate explanatory variables.
Sustainable Livelihoods Approach
The Sustainable Livelihoods Approach (SLA) provides an analytical framework to study a
planned change process, such as in the case of this study where AMSDP is attempting to
improve the market situation for smallholder farmer groups. The foundation of SLA is built
upon the insights and years of experience gained in the field of participatory development
(Chambers & Conway, 1991; Chambers, 1997; Ashley & Carney, 1999). The approach is based
on a core set of principles; development should be people-centered, holistic, dynamic, built on
people's strengths, and sustainable (Ashley & Carney, 1999, p. 7). As defined by Chambers &
Conway (1991, p. 6), "a livelihood comprises the capabilities, assets (stores, resources, claims
and access) and activities required for means of living." The SLA framework provides a wide
enough scope of analysis to capture both micro-level change processes and the larger structural,
environmental and institutional conditions but it never loses focus on one thing how change
processes impact local peoples' livelihood systems. As a framework of analysis, SLA highlights
five elements for understanding rural people's livelihoods by investigating the:
Figure 3-1. Sustainable livelihoods framework
Natural capital is defined as the resource stocks from which resource flows and services
useful for livelihoods are derived. Examples of natural capital include: amount of cultivatable
12 Except where otherwise noted, the definitions and examples of livelihood assets were adapted from Scoones
(1998, pp. 7-8).
* structural and environment conditions and trends (i.e., demography, history, macro-
economic conditions, agro-ecology, politics, etc.);
* livelihood assets available (i.e., social, human, financial, physical and natural capital);
* institutional processes and organizational structures that influence access to livelihood
* livelihood strategies employed given the structural and institutional context and
* likely sustainable livelihood outcomes and tradeoffs based on the strategies employed
(Scoones, 1998, p. 6).
Figure 3-1 illustrates the key elements of the SLA framework (Ashley & Carney, 1999, p. 47).
Livelihood assets.12 Within the context of this study, farmers are in possession of certain
livelihood assets that help them meet their livelihood obj ectives. These are conceptually broken
down into five "capitals" that taken as a whole constitute the configuration of livelihood assets
available to the farm family to meet their practical and strategic needs.
H = Human Capital S = Social Capital
N = N~atural Capital P =Phy~sical Capital
F = Financial Capital
CONTEXT H< >
use of NR base
land; land use and tenure (i.e., owned, shared, communal, rented, usufruct rights); land
characteristics (i.e., soil fertility, erosion, topography, and other agro-ecological features);
ownership/access to other natural resources (i.e., forests, water, wildlife, etc.).
Physical capital is the basic infrastructure and producer goods needed to support
livelihoods and include: shelter and housing; farm implements and tools (e.g., hoes, plows,
tractors, etc.); processing machines; storage facilities; pumps and latrines; transportation (e.g.,
bicycles, carts, trucks, etc.); and other "wealth" items (radio, television, gas stoves, etc.).
Financial capital is defined as financial resources that people use to achieve their
livelihood obj ectives, including flows and stocks that contribute to consumption as well as
production. This includes: savings in the form of liquid assets (e.g., livestock, jewelry, etc.);
bank deposits; cash; pensions and other state transfers; and remittances.
Human capital refers to the skills, knowledge, ability to labor and good health that together
enable people to pursue different livelihood strategies and achieve their livelihood obj ectives.
This includes: household composition and life cycles that determine producer to consumer ratios
of available labor; indigenous knowledge and farming expertise; formal education, including
literacy, numeracy; marketing skills and market knowledge (e.g., access to marketing
information on quantity and quality standards, price information, and existing and emerging
Social capital is defined as an "investment in social relations by individuals through which
they gain access to embedded resources to enhance expected returns of instrumental and
expressive actions" (Lin, 2001, p. 18). Social capital is distinguished between the cognitive and
structural forms. Cognitive social capital deals with group attributes of trust, norms, attitudes,
and shared values (Uphoff & Wijayaratna, 2000, p. 1876). The structural forms of social capital
are distinguished further between bonding and bridging social capital. Bonding social capital is
defined as "expected returns of expressive actions for preserving or maintaining resources" and
bridging social capital is defined as the "expected returns of instrumental actions for searching
for and obtaining resources" (Lin, 2001, p. 18). Examples of bonding social capital include:
membership in groups within a community (i.e., producer groups; processing groups; labor
sharing groups; rotating credit and savings associations); and farmer personal networks with a
high number of intra-community ties that are more horizontal in nature. Examples of bridging
social capital include: membership in groups and farmer personal networks with more extra-
community ties, such as relationships with service providers (i.e., local government ministries
and NGOs) and agribusiness partners.
Impact of institutions, policies, and processes. These livelihood assets interact with
policies, institutions, and processes in influencing livelihood strategies and alternatives. The
market is one of the critical institutional site that influences and directly impacts livelihood
alternatives. Thus, is particularly important to study both the institutional environment (e.g.,
government policies on land tenure) under which farmers must operate and the specific
institutional arrangements (e.g., contract farming) that impact farmers' access to key livelihood
Context of vulnerability. Vulnerability refers to things outside people's own control.
SLA highlights several broad areas under the vulnerability context which must be given due
consideration. The vulnerability context includes:
* trends in population, resources, economic indicators like prices, governance, or even
* shocks like changes in human or animal health, natural disasters, sudden economic
changes, or conflict; and
* seasonality in prices, agricultural production, employment opportunities, resource
availability, or health (Scoones 1998, pp. 6-7).
Livelihood outcomes and strategies. With specific mention to the project under study,
there are two obj ectives (livelihood outcomes) that should be achieved through greater market
access for smallholder farmers. These are increased incomes and food security. While these are
the explicit obj ectives, the proj ect also expects that these livelihood outcomes to be sustainable.
Sustainability is a fundamental component of the SLA framework and takes on a specific
meaning. As defined by Chambers & Conway (1991, p. 6), "a livelihood is sustainable when it
can cope with and recover from stresses and shocks, maintain or enhance its capabilities and
assets, and provide sustainable livelihood opportunities for the next generation."
Understanding SLA within the study context. Through the lens of a livelihoods
approach, AMSDP is attempting to improve the marketing performance of smallholder farmers
by enhancing their stock of human and social capital. Providing skills training in marketing and
entrepreneurship increases farmers' human capital, whereas linking these groups to others in the
market chain forges new business partnerships thereby enhances their social capital. AMSDP
expects that these asset-enhanced farmer groups will carry out collective action initiatives to
improve their marketing performance. This, in tumn, should lead to improved livelihood
outcomes, such as higher sales from existing or new products, more reliable markets, more
steady and regular income, better access to affordable inputs, as well as in the longer term,
greater food security. Of course, the success or failure of any collective action initiative is never
entirely determined by the farmers groups' actions alone. Therefore, due attention must be given
to the policies, institutions, and process, as well as to the external infrastructural environment -
from a cultural materialist perspective to discern the extent that these factors will impede or
enable successful collective action outcomes.
Conceptual Model for Study
A conceptual model was developed to understand and explain the flow of the planned change
process under study. The model borrows from the SLA framework and utilizes the terminology
and perspectives of cultural materialism. By including certain factors under the infrastr-ucture
and social structure, the model tries to separate and analyze the determining factors from the
enabling or constraining factors that affect group marketing performance. As represented in
Figure 3-2, the wider and determinate infrastructure encapsulates this planned changed initiative
and includes such factors as the smallholder groups' farming systems, the agro-ecological
conditions under which they must work, and their physical access to markets (e.g., distance to
markets, access to feeder roads, conditions of roads, etc.).
Physical access to markets
Farmer Groups s Collective LSIPerformance Outcomes
Group composition Action Improved market situation
andrc aracterstics A ntatvsnd positive livelihood
SPA Intervention Marktet Chain Actors
Figure 3-2. Conceptual framework for research study
Farmer groups are represented under the social structure and this includes a number of
factors that will affect a group's ability to enact successful collective action initiatives (i.e., the
group's asset configurations, composition and characteristics). PAs are intervening to enhance
human capital in form of marketing skills, business acumen, and other group capacity training,
which is represented by the solid line going from the PA directly to the social structure. Along
with these training activities, the PAs are also providing some groups with market linkages to
other chain actors and this is represented by the dotted line going to the collective action
initiative, as well as the lines connecting PA intervention to market chain actors. Farmer groups
are also carrying out collective action initiatives without direct linkages from the PA, which is
represented by the lines connecting the collective action initiatives to the market chain actors.
The performance outcomes represent the extent that groups have improved their market situation
and affected positive livelihood outcomes to the members of their groups. Using this conceptual
model as a guide for understanding the change process, we can now turn our attention to the
overall research design of this study.
This study is similar in the methodological approach of action-oriented research. Bogdan
& Biklen (1982) define action research as "the systematic collection of information that is
designed to bring about social change" (quoted in Merrigan & Huston, 2004, p. 26). The
research design was constructed on the premise that the Eindings would prove useful to other
study stakeholders, particularly for FAIDA and TIP as the two implementing partners of
AMSDP. Through the course of this study, particular efforts were taken to provide FAIDA and
TIP feedback on emerging findings. The programmatic recommendations in the last chapter
were also specifically tailored to NGOs, like FAIDA and TIP, engaged in agro-enterprise
development initiatives. As such, this study most resembles an impact assessment, which is one
of the more comprehensive types of action research utilized by researchers and practitioners in
the field of development (Lilja & Johnson, 1999). Impact assessments are results-oriented and
typically divided into short-term outputs, intermediate-term outcomes, and long-term impacts of
the proj ect in question. Observing over a year of the proj ect work cycle allowed reporting on
some of the short-term outputs and intermediate outcomes of the project. If the expected long-
term impact of AMSDP is sustained increases in incomes and food security, it is proposed that
the primary outcome (intermediate result) of training and market linkages is an improvement in
farmer group marketing performance.
To assess the effect of the intervention on producer group outcomes, pretest-posttest
research design was utilized (Johnson, 1998, p. 149). The pretest observations (i.e., the first
round of interviews) were undertaken as the groups were undergoing training from the PAs. Six
to eight months later at the conclusion of the intervention, the same groups underwent posttest
assessment (i.e., the second round of interviews) to ascertain any changes in their market
situation. Because the study did not start before the intervention, considerable time was spent
during the first interview to assess a group's baseline position on marketing and to allow
sufficient room for the group to comment on how the intervention had changed or had not
changed its market situation.
Data Collection Methods
A mixed-methods approach (Marsland et al., 2000) was utilized for this study by
employing the following qualitative and quantitative data collection methods:
* Document analysis of the project materials that included participatory rural appraisals
(PRAs) on the farmer groups and villages/wards; market analysis studies of the study area;
and quarterly reports submitted to AMSDP on the progress of the interventions.
* Key informant interviews with PA management and Hield staff, District officials and field
officers (e.g., agricultural/livestock extension and marketing/community development
workers), and with elected and traditional village leaders.
* Participant observation of PA training activities with their respective farmer groups
* Semi-structured interviews with 34 producer groups. Two rounds of interviews were
conducted six to eight months apart. The content of the jirst group interview included:
background history and baseline information on the group; recent activities within the
group and with other service providers; participatory evaluation of AMSDP; marketing
obstacles and opportunities; and future plans for the group. A participatory market chain
analysis activity was also carried out with several of groups to identify the marketing
channels of the major commodities in the study area. An experimental economics game
was also played with all the groups. 13 The content of the second group interview focused
primarily on questions and conversations to assess each group's progress in accomplishing
their short-term goals and improving their market situation. The second interview also
provided an opportunity to fill in the gaps missed in the first interview.
* Self-administered questionnaires were also included with the first round of interviews in
order to collect quantitative and qualitative data on individual members. Individual data
were collected on 381 members 182 participants from FAIDA groups and 199
participants from TIP groups. Much of these individual-level data were later aggregated to
the group-level in order to assess group characteristics that may play a role in improving
their market situation. 14
Study Time Frame
TIP and FAIDA officially started their component of the intervention in January of 2004.
This study began in April of 2005, roughly a year after the farmer groups had been selected to
participate in AMSDP. In the year prior to the commencement of this study, FAIDA and TIP
had already conducted most of their group strengthening and marketing skills training. During
this same time period, both TIP and FAIDA had also linked several groups to other market chain
actors, a practice they continued while this study was underway. The study started by joining
TIP and FAIDA to the field to observe some of their remaining group training activities. After
observing these training, a first round of group interviews was conducted, which were then
followed up with a second round of group interviews roughly six months later. The last of the
13 The experimental economic game played with the groups is called the Public Goods Game (PGG) and it is
explained in greater detail later in this chapter.
14 The questionnaire was translated into Swahili, which is the language of instruction in primary schools in
Tanzania. Illiterate members, whom there were only a few, were assisted either by my research assistant or by
another group member. For the English version of the questionnaire, see Appendix A.
group interviews concluded in late May of 2005. Thus, from start to finish this study recorded
roughly a year of the proj ect intervention.
As illustrated in Table 3-1, the sample for this study is comprised of 34 groups with a
mean group size of 35 members. FAIDA groups have more females than TIP groups, as three
out of the four female-only groups in this study are working with FAIDA.
Table 3-1. Group study sample
FAIDA TIP Sample total
N % N % N %
Sample size 16 18 34
Size of group
Mean 32 38
Range 15 40 20 150 35
Membership by sex
Female 277 55% 272 40% 549 46%
Male 230 45% 406 60% 636 54%
Total membership 507 678 1185
Members sampled by sex
Female membership sampled 105 56% 68 34% 173
Male membership sampled 84 44% 131 66% 215
Total membership sampled 189 37% 199 29% 388 33%
In terms of membership sampled, 37 percent of all the group members working with FAIDA and
29 percent of all the group members working with TIP participated in the first round of
interviews. Thus, the total study sample represents 33 percent of all group members. These
group members attended the group interviews and filled out questionnaires about themselves and
their households. In terms of gender representation, FAIDA groups were well represented with
female members representing 56 percent of the total sampled, but TIP group were slightly under
represented with the total membership of women at 40 percent and the female membership
sampled at 34 percent.
The study relied on a convenience sample of the group members. When requesting the
interview, each group leadership was asked to invite roughly 10 to 12 members and to make sure
that both male and female members were invited. The method of invitation employed by group
leaders was inquired during the meeting. One of three answers was usually given. In the most
common pattern (46% of groups), all members in the group were invited, with those in
attendance at the meeting dependent on individual availability. The second most frequent
response (31%) was that active members those who can be counted on to show up at meetings
- were invited to come to the meeting. The third reason (23%) given was that representatives
were chosen to attend the meeting. This was most often the case with groups formed at ward
level that consisted of smaller groups, or with larger groups that had sub-committees or sub-
groups that were chosen to represent their particular constituency.
Dependent Variable: Marketing Performance Rating
As the primary dependent variable of this study, a marketing performance rating (MPR)
was constructed to ascertain the extent to which the groups showed any concrete signs that their
market situation had improved through the proj ect intervention. Each group was given a
marketing performance rating ranging from 0 to 2, with 0 signifying no improvement, 1
signifying some improvement, and 2 signifying large improvement in their market situation. The
MPR scale was established in the following manner:
No improvement. Groups were given this rating if, from their own admission, they felt
that the proj ect intervention had led to little meaningful improvement in their market situation.
Some groups did feel that the training activities were beneficial to them and that given the right
circumstances, they could use these training to improve their situation. However, by the end of
the study and proj ect cycle, none of these groups could provide concrete examples of such
training making a difference in their market situation.
Some improvement. Groups were given this rating by showing that they had the ability to
take the training activities and successfully put them into practice. Such groups were able to
provide concrete examples of how their market situation had improved from participation in the
project. Examples included: increased sales and higher prices for existing products; more
reliable sales markets through enhanced relational or formal contract farming arrangements;
diversification into higher value crops, increased access to market information and bargaining
power; getting involved in post-harvest value adding activities; participation and increased
profits by involving themselves in more chain management activities.
Large improvement. A few groups showed the ability to improve their market situation
and did this at a level of success that separated them from the other groups. In most cases, these
groups showed striking market improvements by initiating several collective action initiatives.
As an example, one group shifted their production to more profit-making enterprises, shortened
the market chain between the group and retailers by taking over transportation activities, and
entered into farming contract arrangements with an agribusiness all of which has led to
increased incomes and a more reliable market for their crops. A complete list of the collective
action initiatives enacted by groups to improve their marketing performance will be provided in
the study findings.
Improved marketing performance is an outcome of a number of endogenous and
exogenous factors. The following covariates of MPR were identified a priori based on the
literature reviewed previously. The following tables show a list of explanatory variables
delineated into five headings: group assets; group composition; group characteristics; group
heterogeneity; and market access. In most cases, the following tables of explanatory variables
provide adequate description and values, but there are some variables that require further
Group Asset Variables
Table 3-2 provides a description and value for a number of explanatory variables dealing with
livelihood assets that may affect a group's ability to improve its market situation. The physical
and natural assets include the variables: group wealth ranking; reliable water source; and land.
The social capital assets are delineated by cognitive and structural forms, and include: group
trust; group altruism; community leaders; and providers/partners.
Group wealth ranking. Wealth ranking techniques have been widely used in social
science research for assessing household and individual wealth, especially when it is difficult to
obtain income and expenditure data (Spring et al., 1996; Brandt & Spring, 1998; Grandin, 1998).
A number of studies comparing wealth ranking techniques to income, expenditure, and other
wealth-related data have proven these techniques to be a reliable and valid way to measure
wealth (Adam et al., 1997; Temu & Due, 2000). This ranking was established by aggregating
individual's wealth ranking to the group level. Individual wealth rankings were modeled after
those established in a joint World Bank and Government of Tanzania study (Narayan, 1997) that
utilized a number of participatory approaches for establishing regional wealth rankings for all of
Tanzania, which included the regions of Arusha and Kilimanj aro. This provided the foundation
by which further details were added based on discussions with test-pilot groups and
crosschecked with key informants.
Table 3-2. Group asset variables by description and value
Explanatory variables Description
Physical and natural assets
Group wealth ranking
Aggregated score taken from
individual group members
based on household
ownership of physical, natural
and financial assets
Maj ority of group members
having access to reliable water
Aggregated and average score
taken from the total amount of
land (owned and rented) by
Three questions concerning
trust aggregated to provide
three group scores
Group score based on playing
the Public Goods Game
Group members that are also
elected or traditional leaders
within their community
Service providers and other
partners that group has
worked with and/or currently
Membership in other
agriculture and development-
1= very poor
2 = poor
3 = average
4 = rich
5 very rich
Reliable water source
0 = No
1 = Yes
Social capital assets
Cognitive social capital
Structural social capital
Number of acres
Scores between 1- 3
Higher the score, higher
Score between 0.46 1
Higher the score, higher
the altruistic behavior
Percentage of members
interviewed that are
The total number of
providers divided by
number of years of group s
Score between 0.44 2
Total number of ties to
One particular area of depth that was added was the development of a material asset index. The
first step was asking the participants to circle all items owned by their household from a pre-
defined list. Items were chosen for the questionnaire that could be used to improve farm
productivity (i.e., hand hoe, oxen plough, processing machine, and tractor) and marketing (i.e.,
bicycle, cart, motorcycle, car, truck, radio, mobile phone, and television). Using Anthropac
(Borgatti, 1992) for the analysis of unidimensionality, a Guttman scale was established to assess
and assign material wealth (Guest, 2000). Eight household items (hand hoe, radio, bicycle, cart,
oxen plough, mobile phone, television, processing machine) produced a high unidimensionality
score with a coefficient of reproducibility (CR) of 0.95 and coefficient of scalability (CS) of
0.62. These scores strongly indicate that all items are a composite measure of one underlying
concept (Bernard, 1995).1 A material asset index score (1 to 8) was then assigned to each group
member. The material asset index score along with further considerations, which is shown in
Appendix B, were used to assign each group member a final wealth ranking score from 1 to 5.
Reliable water source. Each group was assigned a score of 0 or 1 based on group's
access to a reliable water source. This was ascertained by individual members' answers to the
self-administered questionnaire, field observations, and discussions with group members and key
informants. In most cases, a reliable water source meant that maj ority of group members had
access to irrigation for their crops, or where groups felt that they could rely on the rainfall, as
was the case with two groups in high elevation and precipitation areas.
15The coefficient of reproducibility (CR) is a measure of the unidimensionality of the items in a scale. By
convention, a CR of 0.90 or higher is accepted as evidence that a set of items have scaled unidimensionality (Guest,
2000, p. 350; Bernard, 1995, p. 296). There is, however, one problem with the CR. A high CR is sneet
extreme marginal distributions in terms of both items and individuals so that a high CR can be achieved even in
random data" (Guest, 2000, p. 351). In order to deal with extreme responses, Menzel (1953) developed the
coefficient of scalability (CS), which can be interpreted as a proportion reduction in error (PRE) statistic. To the
extent that the scale has fewer errors than expected by chance, the CS moves toward 1.0. Although there is no
definitive score, CS of 0.60 or higher is generally considered an acceptable level of error (Guest, 2000, p. 351).
Group altruism. Each group member was assessed on whether she or he exuded more
self-interested or more altruistic behavior toward the rest of the group. This individual
assessment was then aggregated to the group level to get a group altruism score. To assess
various levels of individual altruistic behavior, each group played a one-shot Public Goods Game
(PGG). The PGG is well known within the field of experimental economics and has been
applied in several cross-cultural studies (Gurven, 2004; Henrich, 2000). In this adaptation of the
PPG, each group member is given ten tokens (square pieces of paper) each worth 50 Tanzanian
Shillings (TShs), thus totaling TShs 500 (or the equivalent of US$0.40). Each member is given
the option of contributing none, some, or all her/his money to the group fund. Any money that is
contributed to the group fund will be doubled and then shared equally among all members. To
clarify the possible outcomes, three scenarios are explained to the group: (1) all contribute
everything to the group fund, and thus double their money with each member getting back TShs
1000 (roughly half a day's wage); (2) no one contributes anything to the group fund, thus
holding onto their original sum of TShs 500; and finally (3) most members contribute everything
and a few members contribute nothing. This final option is further explained to the group
through an example. If seven players contribute everything and three players contribute nothing,
the total amount in the group fund will be TShs 3,500. Once doubled it becomes TShs 7,000,
which divided equally leaves each member with TShs 700, except for the three players that
contributed nothing. Each of these players will end up with their original sum of 500 that they
did not contribute plus the 700 from the group fund, thus totaling TShs 1200.
Faced with these scenarios, the PGG game shows how free-riding and self-interested
behavior can be rewarded, but not without negative consequences to the rest of the group. Once
all the options have been explained, each member makes her/his contributions in private to the
group fund. The tokens the member chooses not to contribute are then turned over on the spot
for monetary remuneration. All this insures confidentiality with only the final tally of group
fund contributions made to the group, and the final amount in the group fund divided equally.
Group altruism scores are based on the group's total contribution to the fund, and ranged from
0.58 to 1, with lower scores revealing groups with more self-interested members and higher
scores revealing groups with more altruistic-minded members.
Group trust. Three statements concerning group trust were presented on the
questionnaire using a three-point Likert scale, with participants responding that they: (1) agreed
with the statement; (2) felt neutral or "middle" about the statement; or (3) disagreed with the
statement. They responded to the following three statements:
* Most members in your group can be trusted.
* Most members in your group are willing to help if you need it.
* In your group, members generally do not trust each other in matters of lending and
These answers were then coded from 1 to 3 and aggregated to the group level to provide three
general measurements of group tr-ust. 16 Trust questions represent qualities of cognitive social
capital and were adapted from the World Bank' s Social Capital Assessment Tools (SOCAT). 1
Structural social capital. The three variables community leaders, providers/partners,
and other groups were selected as indicators of structural social capital.ls The first two
variables capture the aspect of "bridging" social capital, and thus deal with the extent that groups
16 The answers to the third trust question concerning matters of money were recorded to fit the scale, with lower
scores representing low trust and higher scores representing high trust among the members.
17 SOCAT can be accessed at World Bank (2007) [online]
http://web. worldbank.org/WB SITE/EXTERNAL/TOPIC S/EXTSO CIALDEVELOPMENT/EXTT SOCIALCAPITA
1s These are recognized indicators of structural social capital in the World Bank's Social Capital Assessment Tools
(SOCAT) see above footnote to access the SOCAT online.
have ties or relations with other actors or organizations outside of their community (i.e., village
or ward). These ties or relations are more vertical in nature and offer opportunities for the group
to access and obtain new resources. The third variable other groups captures the aspect of
"bonding" social capital, and thus deals with ties and relationships within their community.
These ties and relationships are more localized and horizontal in nature and offer opportunities to
maximize existing resources. But, the conceptual divide between bonding and bridging capital is
not always so straightforward, since many of these other groups may also have the ability to
access new resources and thus make it possible that belonging to other groups will serve both the
purposes of enhancing bonding and bridging social capital. For purposes of this study, the most
important point to make when considering these sets of social capital variables is to distinguish
between structural social capital and cognitive social capital. These are divided in Table 4-3
with structural social capital indicators being community leaders, providers/partners, and other
groups; and the cognitive social capital indicators being the three trust variables and group
Group Composition Variables
Table 3-3 provides a description and value for a number of explanatory variables dealing
with group composition that may affect a group's ability to improve its market situation. These
variables include the following: age of members; education; religion; ethnicity; and gender.
There are two variables for gender, with the first delineating groups by gender category and the
second considering leadership by sex.
Table 3-3. Group composition variables by description and value
The average age of the
members for each group
Age of members
Number of years
The average number of years
of schooling for each group
Religion of the majority of
Ethnicity of the majority of
Number of years of
0 = Christian
1 = Muslim
1) Female-only groups
2) Female-dominated groups
3) Male-dominated groups
4) Gender-balanced groups
Score between 0-1
0 All female group
0.5 -Balanced leadership
1 All male group
< 0.5 Female majority
> 0.5 Male majority
Leadership by sex
Each group classified into one
of four gender categories
based on membership,
leadership, and decision-
Proportion of male to female
elected leaders in the group
Groups by gender categories. Groups were classified into four basic gender categories.
These designations were based on group membership, the leadership structure, and the decision-
making dynamics of the leaders.
* Female-only groups
* Female-dominated groups: Maj ority of the members are women; maj ority of the leaders
are women, and they guide the group or make the important decisions themselves.
* Gender-balanced groups: There is a balance of female and male members in the group;
there is a balance of male and female leaders in the group and decisions involving the
group are not dominated by a particular sex.
* Male-dominated groups: Maj ority of the members are men; maj ority of the leaders are
men, and they guide the group or make the important decisions themselves.
Group Characteristics Variables
Table 3-4 provides a description and value for a number of explanatory variables dealing
with group characteristics that may affect a group's ability to improve its market situation. The
activity level variable refers to the number of effectively operating internal activities that a group
runs. The type of activities that these groups were involved in, prior to and after the intervention,
included: rotating credit schemes; collective marketing; bulk input purchasing; labor sharing
activities; and group farms for purposes of consumption or collective marketing, as well as
demonstration plots for experiential learning and seedling nurseries.
Table 3-4. Group characteristics variables by description and value
Explanatory variables Description Value
Group age Number of years the group has Interval
been in existence Number of years
Group size Number of active members based Interval
on data collected during 2nd Number of group members
Combined groups Groups that are a combination of Nominal
smaller groups 0 = No
1 = Yes
Group level Groups that have been formed at Nominal
the village level or at the ward 0 = Village
level 1 = Ward
Activity level The number of internal activities Interval
that groups are currently doing Total number of activities
apart from the project per group
Maturity Newly formed group versus Nominal
groups that have been in existence 0 = Groups in existence for
for a longer period of time two years or less
1 = Groups in existence for
three years or more
Group Heterogeneity Variables
Table 3-5 provides a description and value for a number of explanatory variables dealing
with group heterogeneity that may affect a group's ability to improve its market situation. All
these variables will be tested to ascertain the extent that group homogeneity or heterogeneity
affects group marketing performance. Three variables ethnicity, education, and religion -
represent "cultural" attributes and are tested with the wealth and gender to see how such shared,
or divergent, views, beliefs, attitudes, and assets produce positive or negative effects on group
marketing performance. The coefficient of variation was used to measure educational and
wealth heterogeneity. The coefficient of variation is one of the most commonly used statistics by
organizational researchers for assessing the effects of group-based demographic diversity
(Bedeian & Mossholder, 2000). Ethnic, religious, and gender heterogeneity were measured
using a proportions statistic. Again, organizational researchers commonly use this statistic to
assess the effects of heterogeneity on group behavior and actions (Williams & Mean, 2004).
Table 3-5. Group heterogeneity variables by description and value
Explanatory variables Description Value
Ethnic Proportion of the members of Interval
other ethnic groups by the Score between 0 0.73
members of the majority ethnic Higher the score, the more
group ethnic heterogeneity
Religious Proportion of the members of the Interval
minority religions divided by Score between 0 0.5
members of the major religion Higher the score, the more
Educational Coefficient of variation, which is Interval
the standard deviation divided by Score between 0 1.13
the mean score of the years of Higher the score, the more
schooling of members educational heterogeneity
Gender Proportion of male members to Interval
female members Score between 0 0.76
Higher the score, the more
Wealth Coefficient ofvariation, which is Interval
the standard deviation divided by Scores range from 0.13 0.46
the mean score of member's Higher the score, the more
wealth ranking wealth heterogeneity
Market Access Variables
Table 3-6 provides a description and value for a number of explanatory variables dealing
with market access factors and how they may affect a group's ability to improve its market
Commodity types. Each group has certain crops that their members grow that were put
forward as possible crops to promote as a viable agro-enterprise. It was important to make some
demarcation between these crops since some have more marketable qualities. The crop headings
also highlight the different farming systems available to the groups. The cereals/legumes
category represents traditional staple food crops, and includes: corn, beans, millet, pigeon peas,
and sunflower. The other categories encompass a number of higher value and non-staple food
crops, and includes: fruits/vegetables; livestock (i.e., raising dairy cows and chickens); coffee;
Table 3-6. Market access variables by description and value
Explanatory variables Description Value
Distance to market Distance from group meeting Interval
location to the major regional Distance in kilometers
Road conditions Road conditions from group Ordinal
meeting place to major markets 0 = Bad
1 = Average
2 = Good
Commodity types The primary crops grown and Nominal
selected by groups to improve their 1) Cereals/Legumes
market situation 2) Fruits/Vegetables
Partner agency (PA) The two partner agencies working Nominal
with the farmer groups 1) FAIDA
PA linkages Whether or not PAs actively linked Nominal
producer groups to other chain 0 = No
actors 1 = Yes
Distance to markets and road conditions. These two variables represent physical access
to markets, and thus consider this variable from an exclusively geographical and infrastructural
perspective. The distance to markets variable measures the distance from group meeting place to
the maj or market in the region, which for the FAIDA groups is the town of Moshi in the
Kilimanj aro region, and for the TIP groups it is the town of Arusha in the Arumeru Region. The
road condition variable was built on an ordinal scale with the following delineations:
* Good road conditions: Group meeting place is 2 kilometers or less from a paved road
that connects to the maj or regional market.
* Average road conditions: Group meeting place and surrounding area connected by
gravel road. Road is fairly flat and accessible most of the year.
* Bad road conditions: Group meeting place and surrounding area connected by dirt road
only. Road is uneven, difficult to maneuver, and may not be passable during the rainy
Partner agency (PA) and PA linkages. The partner agency variable is categorized under
market access since the PAs are attempting to enhance the marketing skills of the farmer groups
in the hope that groups will undertake collective action initiatives to improve their market
situations. Given some of the differences in the training activities, there is the expectation that
each PA will have varying results in improving the marketing performance of their respective
farmer groups. Likewise, PAs are also active in linking farmer groups to other chain market
actors, and thus both their market linkage success and failures will impact groups' marketing
Matching Research Objectives with Hypotheses
This final section restates the principal research obj ectives of this study and matches these
obj ectives to the relevant hypotheses. This is followed by a number of explanatory variables that
will be used to test each set of hypotheses.
Objective 1: Farmer Group Assets
The first obj ective is to assess the extent that certain livelihood asset configurations (i.e., natural,
physical, financial, human, and social) will affect the group's ability to improve their market
situation. Hypotheses relating to the group asset objective include:
* Hypothesis 4f: Farmer groups with lower levels of poverty among members will be better
positioned to improve their marketing performance.
It has been clear from the onset of this study that "poverty" is broadly defined to include the five
livelihood capital assets (i.e., physical, natural, financial, human, and social). The first
hypothesis will be tested using the following variables: Wealth ranking; Reliable water source;
Land; Education; Commodity types. There are also more specific hypotheses dealing with
different aspects of social capital. The first three hypotheses are concerned with cognitive social
* Hypothesis 3a: Farmer groups with a high level of trust among members will be better
positioned to improve their marketing performance.
* Hypothesis 3b: Farmer groups that exude more altruistic rather than self-interested
behavior among members will be better positioned to improve their marketing
* Hypothesis 4c: Farmer groups with higher level of interdependence among members will
be better positioned to improve their marketing performance.
These hypotheses will be tested using the following variables: General trust; Help trust; Money
trust; Altruism. The last hypotheses under the group asset objective deals with structural social
* Hypothesis 3c: Farmer groups with more ties to other organizations in and outside of their
community will be better positioned to improve their marketing performance.
This hypothesis will be tested using the following variables: Community leaders;
Providers/Partners; and Other groups.
Objective 2: Farmer Group Composition
The second obj ective is to assess the extent that certain group composition attributes will affect
their ability to improve their market situation. Hypotheses relating to the group composition
obj ective include:
* Hypothesis 4d: Farmer groups with heterogeneity of endowments will be better positioned
to improve their marketing performance.
* Hypothesis 4e: Farmer groups with homogeneity of identities will be better positioned to
improve their marketing performance.
* Hypothesis 2a: The gender composition of groups, especially of the decision makers, will
influence group marketing performance.
These hypotheses will be tested using the following variables: Leadership by sex; Gender
categories; and the five heterogeneity variables (i.e., ethnic, religious, educational, gender,
Objective 3: Farmer Group Characteristics
The third obj ective is to assess the extent that certain group characteristics, which include the
group's institutional capacity, will affect their ability to improve their market situation.
Hypotheses relating to the group characteristics obj ective include:
* Hypothesis la: Farmer groups with functioning internal institutions for guiding group
behavior and action will be better positioned to improve their marketing performance.
* Hypothesis 4b: Farmer groups with past successful activities will be better positioned to
improve their marketing performance
* Hypothesis 4a: Smaller farmer groups will be better positioned to improve their marketing
These three hypotheses will be tested using the following variables: Maturity; Age of group;
Current activities, and Size of group. The first three variables act as proxy measures for