IMPLICATIONS OF INCENTIVE-BASED CONSERVATION PROGRAMS FOR
GOVERNANCE, GENDER AND COLLECTIVE ACTION IN THE ULUGURU
MOUNTAINS, TANZANIA
By
Mamta Vardhan

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Community, Agriculture, Recreation, and Resource Studies
2011

ABSTRACT
IMPLICATIONS OF INCENTIVE-BASED CONSERVATION PROGRAMS FOR
GOVERNANCE, GENDER AND COLLECTIVE ACTION IN THE ULUGURU
MOUNTAINS, TANZANIA
By
Mamta Vardhan

Payment for environmental services (PES) is often promoted as a win-win option for
conservation and development, but its potential social impacts remain underexplored. PES
literature has not focused much attention to the important issues of institutions and governance
underlying PES projects, the potential of PES to alter existing rights to resources which may
work against marginal groups, and the possible perverse impacts of introducing money on
collective action for provision of ecosystem services. This research explores these issues in the
context of tree planting programs and forest management arrangements in the Uluguru
Mountains in Tanzania. The first essay, “Forest Governance in Participatory Forest Management
Regimes in Tanzania: Lessons for REDD,” indicates that although participatory forestry
programs appear to offer a good institutional base for launching REDD, in practice they are
plagued by poor governance, corrupt practices and low accountability. Participatory forestry
institutions in Tanzania need to be strengthened in order to ensure effective and equitable
implementation of REDD. The second essay, “Payments for Environmental Services and
Women’s Land Rights in the Uluguru Mountains, Tanzania,” presents a case where men in the
study villages have used tree planting projects to claim women’s lands. The paper cautions that
contestations over land may become pronounced if PES projects continue to ignore the social
relationships underlying the production of ecosystem services. The third essay, “Pro-Social
Behavior and Incentives: Experimental Evidence from the Uluguru Mountains, Tanzania,”

explores the impact of payments on collective action through village-level field experiments. It
shows that participation in communal tasks is high irrespective of the incentive; high payments
can increase participation but low payments can reduce participation relative to when no
payment is made.

ACKNOWLEDGEMENTS

This dissertation would not have been possible without the long and untiring support of my
advisor Dr. John Kerr. He gave me the “researchable idea” that became this dissertation. He
stood by me when the going was rough and slow (which was a lot many times than either one of
us would have liked it to be). My sincere thanks to you John for helping me pull this through!
I also thank Dr. Brent Swallow of the World Agroforestry Center (now at University of Alberta)
my supervisor during fieldwork in Tanzania for giving me the generous funding and freedom to
pursue this work and being patient with me as I juggled my student and mom self to stay on top
of my fieldwork.
I also acknowledge the constant stream of kind words and encouragement that I received from
my graduate committee members, Dr. Anne Ferguson, Dr. Tracy Dobson, Dr. Kim Chung and
Dr. Joe Arvai. My dissertation was a challenging project with all kinds of methods and
theoretical streams, but not for a moment did I feel that I do not have your faith in my ability to
weave all of this together. I extend a big thanks to all of you for your unwavering support during
all stages of this project.
I acknowledge all my research participants in the villages in the Uluguru mountains and my
friends in Morogoro who made my field stay most memorable! And thanks to Jimmy, my driver
and field brother for running around and tying all loose ends (of which there were many!).

iv

I am deeply grateful to my husband and fellow colleague for his unfathomable belief in me, and
my abilities to carry this project to term. Sometimes, it was just like the eighteen hour long labor
that we endured together (when our dear Kartik was born). He guided me through the personal
end of the PhD labor through his sometimes gentle, sometimes forceful (and yet filled with
belief) coaxing and come ons. I did it Rohit! Thanks for being there with me all the time-in
graduate school and out of it.
Kartik who was born just when I started my PhD, has been through all of it, all the pains and
joys, from conception to fruition. He is the best thing that happened to me at grad school and he
helped immensely in both fraying my nerves but best by calming them down. Thanks Kartik for
carving out those Mommy moments in my life as a grad student!
My parents and family in India deserve a big chunk of appreciation. They endured my long
absence and never complained. Through their silent support, smiles and cheers, I derived
strength to persevere. Mummy, Papa, Mona, Bhaiya, Rupa, Ruchir, Chacha and Chachi, thanks
for being there. LOVE you all! Thanks to Mariah for creating those oases of calm and relaxation
throughout my journey as a student mom.
Looking back, I do not think I leave with just a PhD, I leave with a fuller and richer sense of lifetouched by so many intellectual, emotional and spiritual moments. I will truly cherish my Grad
school days.

v

TABLE OF CONTENTS

List of Tables

vii

Chapter 1: Introduction
References

1
10

Chapter 2: Forest governance in participatory forest management regimes in
Tanzania: Lessons for REDD
2.1 Introduction
2.2 Decentralization of forest management: Justification, process and
Outcomes
2.3 Decentralized forest management in Tanzania
2.4 Conceptual framework: Actors, powers and accountability
2.5 Study area
2.6 Methods
2.7 Findings: Assessing PFM in practice according to Actors,
Powers and accountability framework
2.8 Relevance for REDD
2.9 Conclusion
References
Chapter 3: Payments for Environmental Services and Women’s land rights in
the Uluguru Mountains, Tanzania
3.1 Introduction
3.2 Gender and intra-household effects of environmental stabilization
Projects
3.3 Methods
3.4 Study area
3.5 Findings
3.6 Implications for PES
References
Chapter 4: Pro-social behavior and incentives: Experimental evidence from
the Uluguru Mountains, Tanzania
4.1 Introduction
4.2 Human motivation, incentives and collective action
4.3 Study context
4.4 Methods
4.5 Results and discussion
4.6 Conclusion: Implications for PES
References

vi

11
11
14
15
20
24
26
31
44
46
59

63
63
67
72
78
81
89
99

103
103
105
109
110
116
119
129

LIST OF TABLES

Table 2.1 Actors, powers and accountability according to the legal framework
of decentralized forest management in Tanzania

50

Table 2.2 Criteria for selection and composition of focus groups

51

Table 2.3 Focus group sessions

51

Table 2.4 Focus group guides

52

Table 2.5 Code Book: Forest governance in participatory forest management in Tanzania 53
Table 2.6 Matrix: Accountability mechanisms in JFM and CBFM in Tanzania

58

Table 3.1 Code Book: Matriliny, tree tenure and women’s land rights

92

Table 3.2 Matrix: Women’s land rights

97

Table 4.1 Village wise preferences for group versus individual rewards for
hypothetical work to patrol village forests

124

Table 4.2 Results of stated choice experiment about participating in communal
work to maintain the village school yard under various types and levels of reward

124

Table 4.3 Results of field experiment about participation in communal work to plant
trees in village school yard under various types and levels of reward

125

Table 4.4 Summary of exit interview responses under various treatments in
the field experiment

125

Table 4.5 Record form for participants of field experiments

126

Table 4.6 Script for the experiment

127

vii

CHAPTER 1: INTRODUCTION

Tropical forests and ecosystems in developing countries provide a number of valuable
environmental services such as biodiversity conservation, carbon storage and watershed
protection. However, forest environmental services are being lost because people who live in
forests are not compensated for providing environmental services and thus often have
insufficient incentive to protect forests. In view of the failure of traditional command and control
approaches to conservation, there is growing interest in designing more direct institutional
approaches whereby those who receive environmental services compensate those who provide
them (Pagiola et al., 2005). The resulting approach is known as payment for environmental
services (PES). PES thus involves compensating land users through economic incentives so that
they benefit directly from conserving landscapes and providing environmental services. Thus,
PES projects are akin to creation of a market for environmental service through which buyers
and providers of environmental services interact. In the past decade many PES projects have
been established in developing countries. For example, PES is being explored as a way to
compensate communities all across the tropics under the new Reduced Emissions from
Deforestation and Forest Degradation (REDD) regimes being promoted to reduce green house
emissions by conserving forests in developing countries.
Reactions to PES in conservation and development circles have been mixed. Proponents
argue that by targeting conditional payments to those who constitute a potential threat to
environmental service provision, PES represents a more cost-efficient way to achieve landscape
conservation (Wunder et al 2008). This vision of PES is promoted since environmental financing
is limited and greater efficiency in spending is the overriding concern. However, this narrow
1

vision of PES is deemed problematic on several counts. PES literature has not focused much
attention to the crucial issues of 1) institutions and governance underlying PES projects, 2) the
potential of PES to alter existing rights to resources and create enclosures which may work
against the marginal groups, and 3) the possible perverse impacts of introducing money
influences collective action that underlies provision of ecosystem services. The slant towards
technical issues in design of REDD (such as scale, funding designs, monitoring systems) with
too little attention on governance dimensions such as the social impacts, fairness and distribution
has been flagged as an issue by scholars (Hufty and Haakenstad, 2011). Others critique the
commoditization entailed in PES and emphasize that as monetary values for ecosystem services
arise, other values (held by other people) may be left behind, resulting in enclosures of newly
valuable land and dispossession of the poor and powerless (Sullivan, 2009). There are concerns
that the use of cash as incentive may erode culturally-rooted, intrinsic motives for cooperation
and trigger people to behave in self-interested ways (Deci, 1971) with the result that this
“motivational crowding out” may destroy existing pro-environmental ethics (Fehr and Falk,
2002).
The need for a sharper understanding of the problematic complexities of the relationship
between payment-based conservation instruments and local social processes, including impacts
on governance, gender equity and norm-based collective action is important as there is much
discussion but little empirical evidence about these aspects of PES. This research explores each
of these three big issues in pilot PES tree planting projects and ongoing forest management
institutions in the villages in the Uluguru Mountains in Tanzania. The research attempts to locate
answers to the following three research questions:

2

RQ 1. What are the institutional and governance outcomes of decentralized forest management
in Tanzania? What do these outcomes imply for the implementation of upcoming village
level PES/REDD projects?
RQ1. How does introduction of payment based landscape conservation and tree planting
projects mediate and impact gendered access to and control over land resources?
RQ 2. In what ways does the introduction of direct financial incentives interact with social
norm-based collective action for management of common pool forest resources?
Research on these issues remains important in the study villages because the region is a global
biodiversity hotspot and the provider of many ecosystem services of global and local importance.
The Uluguru mountain ecosystem faces conservation threats from human activities such as
conversion of forests into farm, heavy soil erosion and declining agriculture productivity. The
mountains are home to the Waluguru people, who are a matrilineal community where
traditionally women own land. The forests in the region are being managed under several
decentralized, community based, participatory regimes. There is a tradition of undertaking social
norm mediated voluntary work and collective action on issues of village development. PES
projects are being introduced in the area as a means for landscape conservation and poverty
alleviation. However, many of the projects are being overlaid on existing social, economic
institutions and ecological systems without being deliberate about the likely interactions of these
new projects with pre-existing institutions, resource use and conservation practices. This
research is an attempt to study the likely interactions of the payment-based ideology with preexisting norms of access and control over land, the day-to-day governance of forests and the
potential impacts of PES projects on norm-based collective action.

3

This research work draws from a variety of theoretical and methodological domains, from
behavioral economics, social psychology, environmental anthropology to political science. Thus,
methods that are representative of research in these various disciples were purposefully selected
and applied to collect data under this study. In the tradition of qualitative research, an
ethnographic approach was followed, which included unstructured interviews and conversations,
participant observations, participatory research such as rural appraisals and more structured
approaches such as topical interviews, focus group discussions. Following the quantitative
tradition, household survey and field experiments were conducted. The research participants
included government officials at the district, ward and village level, community members from
six villages.
The three research questions are pursued in separate essays:
Essay 1: Forest Governance in Participatory Forest Management regimes in Tanzania: lessons
for REDD
The Tanzanian national REDD framework identifies participatory forest management (PFM) as a
strategy to achieve forest conservation. Lessons about the performance of PFM are important for
assessing the social feasibility of REDD. Two main approaches to promote PFM in Tanzania are
Joint Forest Management (JFM) on government-owned forests where the communities are not
entitled to any benefits and Community Forest Management (CFM) on village-owned forests
where community members have rights to proceeds from forests. This paper is based on a
comparative study of forest governance in two JFM and two CFM sites in the Uluguru
Mountains of Tanzania. Findings indicate that people responsible for managing village
institutions under JFM feel challenged by the responsibility of protection and expect more

4

support from the state. In contrast, even with its apparently more favorable tenure and benefitsharing arrangements, CFM in the study villages suffers from weak governance in particular
inequitable benefit-sharing, corruption and elite capture. With the reluctance of Tanzanian state
to share benefits under JFM, it is likely that the REDD revenues may remain centralized and the
communities continue to bear the costs of protection. Given the poor governance outcomes of
CFM, questions about its use as a means to deliver equitable forest conservation under REDD
remain unresolved.
Essay 2: Payments for Environmental Services and Women’s Land Rights in the Uluguru
Mountains, Tanzania
Payment for Environmental Services (PES) is often promoted as a win-win option for
conservation and development. At the same time, there is little focus on the how placing new
values on land through PES might reshape rights to land and alter social relationships that
underlie production of environmental services. This paper examines these processes in the
Uluguru Mountains, Tanzania, focusing on the shift in gendered rights to land occurring in a
context of environmental and socio-cultural change. While the weakening of women’s land
rights is happening endogenously, they risk coming under further contestation with the spread of
incentive-based tree planting programs. This paper presents a case in which men have used tree
planting as a means to claim women’s lands, and raises the concern that this process could
become more pronounced as PES projects scale up without sufficient understanding of gendered
property relationships around land.

5

Essay 3: Pro-Social Behavior and Incentives: Experimental Evidence from the Uluguru
Mountains, Tanzania
Payment for environmental services (PES) for common pool resources is based on the premise
that offering payments to groups of land users will motivate them to organize collectively to
provide environmental services. In contrast, research from behavioral economics shows that
introducing monetary incentives can undermine collective action that is motivated by social
norms. In such a case PES could have perverse impacts. To resolve this dichotomy and better
understand the relationship between monetary incentives and collective action, we conducted a
set of choice and field experiments in rural Tanzania. Findings indicate that participation in
communal tasks is high irrespective of the incentive; high payments can increase participation
but low payments can reduce participation relative to when no payment is made; payments can
reduce the satisfaction people derive from participating; social approval incentives play a strong
role in motivating participation; and people express preference for payment in the form of
infrastructure as opposed to either cash payments to individuals or groups.
The essays in this dissertation make a contribution to the scholarly and policy literature by
adding much needed empirical evidence about issues that remain crucial to the successful
implementation of PES and REDD projects.
The first research paper, by providing evidence about the governance outcomes of participatory
forest management programs in Tanzania, indicates that even when the forest management
policies are robust in principle, with appropriate powers vested with appropriate local level
institutions, in practice the governance outcomes are not all positive. These research findings are
important because participatory forest management is being explored as a means to propel the

6

UN-REDD program in Tanzania. The research indicates that the design of REDD projects needs
to account for the multiple actors at local level so as to account for the interests of the powerful
and not-so-powerful. This is especially important because REDD implementation entails huge
amount of funds to be used as incentives to scale up conservation. However, if these power
hierarchies and poor governance outcomes are ignored, the funds under REDD may create
perverse incentives and enormous potential for misappropriation.
The second essay adds to and furthers the scholarly work on gendered access to resources. It
supports previous research experience that concluded that projects aiming to commercialize use
and management of natural resources may displace the rights of women and marginal groups.
This essay furthers this literature by providing empirical evidence about the potential adverse
social equity aspects of PES projects. These findings are important especially because market-led
environmentalism is being touted as the panacea to the problems of landscape degradation and
further conservation in the developing world. This paper uses the evidence from emerging
contestations around land and displacement of women’s land rights to caution policy makers that
these dispossessions might worsen if the complex social context of the area remains invisible
within the PES discourse. It demonstrates the need for project designers to understand and
account for complex relationships around land rights if PES is to help rather than harm the poor.
Interaction between payments and social norms are critical areas for research with strong
implications for PES and REDD. Many REDD projects in Tanzania are expected to be
implemented on communally managed forestlands. The provision of environmental services (in
this case reduced emissions) would require that these community groups manage forests
collectively. In the case of collective contracts, it is important to understand the interaction
between payments and collective action that would underlie provision of ecosystem services.
7

The third and final essay in this dissertation explores these issues and suggests that during design
of incentives under REDD it is important to refrain from offering very low payments as they may
reduce participation, and not to rely on money alone but also build on the positive role that social
norms play to motivate people to undertake collective tasks and consider offering non-cash
incentives at the group level. Much discussion is ongoing on this issue of what types of
incentives to offer under REDD. The focus of these discussions is on cash incentives. However,
as this research shows there is a need to consider other forms of incentives.
To summarize the overall message of this dissertation, the concerns that it raises about
governance, gender equity and erosion of norm-based collective action call for a reformulation of
the present idealistic but simplistic vision of PES with its focus on efficiency to a new and more
realistic vision that acknowledges potential tradeoffs between efficiency and equity in the supply
of environmental services and articulates the possibility of synergy between them.

8

References

9

REFERENCES

Deci, Edward L. 1971. Effects of externally mediated rewards on intrinsic motivation. Journal of
Personality and Social Psychology, Vol 18(1), Apr 1971, pp 105-115
Fehr, Ernst and Armin Falk. 2002. Psychological foundations of incentives. European Economic
Review, Volume 46, Issues 4-5, pp 687-724
Hufty, M., and A. Haakenstad. 2011. Reduced Emissions for Deforestation and Degradation: A
Critical Review. Consilience: The Journal of Sustainable Development, Vol. 5, Iss 1, pp
1-24
Pagiola, Stefano., A. Arcenas and G. Platais. 2005. Can payments for environmental services
help reduce poverty? An exploration of issues and the evidence to date from Latin America.
World Development Volume 33, No. 2, pp 237-253
Sullivan, Sian. 2009. Green Capitalism and the Cultural Poverty of Constructing Nature as
Service Provider, Radical Anthropology 3, pp 18-27
Wunder, Sven, Stefanie Engel and Stephano Pagiola. 2008. Taking stock: A comparative
analysis of payments for environmental services programs in developed and developing
countries. Ecological Economics, Vol. 65, Issue 4, pp 834-852

10

CHAPTER 2: FOREST GOVERNANCE IN PARTICIPATORY FOREST
MANAGEMENT REGIMES IN TANZANIA: LESSONS FOR REDD

2.1. Introduction
Deforestation and degradation of the world’s forests produces 17% of global carbon dioxide
emissions, thereby contributing significantly to climate change (IPCC 2007). UN-REDD, the
United Nations program on Reducing Emissions from Deforestation and Forest Degradation in
Developing Countries, offers a mechanism to mitigate these emissions by offering financial
incentives to preserve existing forests (Miles and Kapos, 2008). REDD+ goes beyond REDD by
also offering incentives for enhancing forest stocks (UN-REDD, 2011). Wealthy countries have
pledged $4.5 billion for the implementation of REDD+ activities in developing countries (Astill,
2010). Decentralized forest management through communities is being viewed as a means to
channel REDD+ activities at the local level (Angelsen, 2009).

Over the past two decades, many countries have introduced decentralized forest governance
aiming to improve forest management. The rationale for these reforms is that decentralization
may result in fair and effective local forums for cooperative resource management (Kaimowitz
and Ribot, 2002). However, evaluations of some decentralization efforts point to challenges
ranging from weak governance including local corruption, poor accountability, inequitable
benefit sharing and contested forest ownership (Ribot and Larson, 2005). These are important
concerns for REDD+ because of the likely use of community forest management as a means for
its implementation (Angelsen, 2009). At the same time there are concerns about the emphasis on
technical issues in design of REDD+ (such as scale, funding designs, monitoring systems) with

11

too little attention to governance dimensions such as the social impacts, fairness and distribution
(Hufty and Haakenstad, 2011).

The purpose of this paper is to analyze the process of forest governance by drawing upon
empirical research on decentralized forest management in Tanzania and to draw lessons for
implementation of REDD+ projects. Mainland Tanzania has gained the reputation of having one
of the most progressive community forestry management regimes in Africa as reflected in policy,
law and practice (Wily, 2000). The Tanzanian forest policy aims to create an enabling legal
environment for decentralized forest management under which local communities, alone or in
conjunction with government authorities, can participate in the management of forest resources
(Blomley and Ramadhani, 2006). The three stated objectives of Tanzania’s decentralized
participatory forest management policy are: (1) improved forest quality through sustainable
management practices, (2) improved livelihoods through increased forest revenues and secure
supply of subsistence products, and (3) improved forest governance at village and district levels
through effective and accountable natural resource management institutions (URT, 2003).

Tanzania is one of the nine pilot countries for the United Nations REDD program (Burgess et al.,
2010). Substantial funds have been pledged for REDD interventions in Tanzania that conserve
forests and improve local livelihoods. In Tanzania, many pilot REDD+ interventions are being
linked to the existing participatory forest management (PFM) programs. However, considerable
challenges to REDD implementation at the local level, especially with regards to governance
deficiencies including occasional corruption (Burgess et al., 2010). It is important that the
development of the UN REDD program in Tanzania learns from these challenges in PFM to

12

design REDD+ interventions. The few evaluations of Tanzania’s PFM programs comment
mostly on the outcomes in terms of forest quality (Persha and Blomley, 2009; Blomley et al.,
2008) and livelihoods (Lund and Treue, 2008), and not so much on forest governance. Forest
governance outcomes can be defined in terms of accountability of local institutions, effects on
equity and inclusion, elite capture and corruption (Agrawal and Ribot, 1999). In light of the
likely reliance of REDD+ implementation on participatory forest management institutions in
Tanzania, it is important to examine the governance outcomes decentralized forest management
on the ground.

This paper assesses PFM in practice according to the actors, power and accountability
framework (Agrawal and Ribot, 1999), relative to what is the formal policy. The paper is based
on comparative case study of three villages engaged in PFM activities in Morogoro district in
Tanzania. In this paper, PFM is used as a generic term to describe a number of different
participatory arrangements, including Joint Forest Management (JFM), in which communities
participate in the management of state owned forest, and Community-based Forest Management
(CBFM), in which village communities have the rights to declare and manage parts of village
lands as village land forest reserves. The study villages include one with a local forest managed
under JFM, a second village with forests being managed under CBFM, and a third in which one
patch of forest is under JFM and another under CBFM. By examining forest management under
these participatory regimes, this paper aims to show the complexity of actors, their interests
involved in decentralized forest management and underscores the need to be aware of this
institutional context in order to develop policies and projects that satisfy the objectives of

13

REDD+ implementation while engaging local communities as active partners in forest
management.

2.2. Decentralization of Forest Management: Justification, Process and Outcomes
Decentralization occurs when a central government formally transfers powers to actors and
institutions at lower levels in a political-administrative hierarchy (Ribot, 2002). Over the past
two decades, a majority of national governments in Asia, Africa and Latin America have
launched decentralization initiatives in natural resources management (Ribot & Larson, 2005).
The most far-reaching policies of decentralized natural resource management in developing
countries around the world have been in the case of management of forestlands (Agrawal et al.,
2008). In developing countries, community-owned and managed forests are estimated to be
around 22 %of all forests (White and Martin, 2002). Around 10-12% percent of forests in 21
sub-Saharan Africa countries are being managed under decentralized arrangements that include
popular participation (Sunderlin et al., 2008).

Political decentralization or democratic decentralization is said to occur when the central
government formally transfers powers to actors or institutions that are downwardly accountable
to the population in their jurisdiction (Ribot and Agrawal, 1999).
In most cases decentralization of natural resource management to local institutions is done on
grounds that it may result in improved efficiency, equity and improved local participation and in
turn better and more sustainable resource management (Manor, 1999; Crook and Manor, 1998;
Ribot, 2002). It is argued that the equity and efficiency benefits of decentralization come from
the presence of democratic local institutions that can better discern and are more likely to

14

respond to local needs and aspirations because they have better access to information and are
more easily accountable to local people (Ribot, 2002).

While decentralization reforms are promoted on the grounds of equity and inclusive participation
that results in better management of local resources, these desired outcomes are seldom achieved
according to research on decentralized natural resource management. Ribot (2003) through a
comparative study of innovative decentralization reforms in six countries of Asia, Latin America
and Africa, show that the purported goals of equity and efficiency have not come about. Ribot
(2003) explain that the fundamental elements of democratic decentralization, including transfer
of discretionary powers and downwardly accountable local institutions, are missing in these
cases. Studies of forest decentralization in Uganda, Mali and Senegal (Oyono, 2004; Ribot,
2004) report that these decentralization efforts have been attenuated by subsequent de facto and
de jure limitations on exercise of local powers (Ribot, et al., 2008).

2.3. Decentralized Forest Management in Tanzania
Tanzania has about 34.6 million ha of forest and woodland habitats. The most important
administrative categorization of a Tanzanian forest or woodland is “reserved” and “unreserved”.
About 14.3 million ha of forests and woodland are found within forests declared as reserves
(URT, 2001). These forest reserves are under the legal authority of the either the central
government (national forest reserves) or local government (local authority forest reserves) and
are managed either for production (timber and other uses) or protection (of biodiversity, water
catchment). The remaining 19 million hectares of unreserved forests and woodland areas form a

15

larger proportion (56%) of the total forest area in Tanzania and fall mostly on village or general
land (URT, 2001).

While forest reserves traditionally were characterized by an exclusionary management policy,
since the 1990s forest management in Tanzania has undergone a reform where key responsibility
of forest management was transferred from central government to village government. The 1998
Tanzanian national forest policy and the 2002 Tanzania forest legislation allow for advancement
of participatory forest management and entitles local communities, either alone or along with
government authorities, to manage forest for sustainable use and conservation (Blomley and
Ramadhani, 2006). Currently more than 3.6 million hectares of forest are being managed under
some form of PFM arrangement (Blomley et al., 2008)

Two main approaches for implementing PFM exist in Tanzania: Joint Forest Management (JFM)
and Community Based Forest Management (CBFM). Both of these programs transfer some
forest management rights to elected local institutions called village environment committees
(VEC), which report to the village council, the governing body of the village elected by village
residents. While VECs are the lowest institution responsible for forest management in both JFM
and CBFM, there are important differences in the rights and decision-making authority under the
two regimes.

Joint Forest Management: JFM is a collaborative management strategy that divides the
responsibility of forest management between the state forest department (usually the central or
district government) and the communities living adjacent to the forest. JFM is usually undertaken

16

on reserved forests such as the national or local authority forest reserves. JFM arrangements are
formalized through the signing of a joint management agreement between village representatives
and either the District Council or the Director of Forestry and Beekeeping (Blomley and
Ramadhani, 2006). JFM agreements specify that local village communities participate as comanagers but under no circumstances can they become owners of reserved forests (Wily and
Dewees, undated). In principle, the joint management agreement does not allow the sharing of
any benefits (in terms of harvest and sale) from forests under JFM (Blomley and Ramadhani,
2004).

Community Based Forest Management: CBFM takes place on village land, i.e., land that is under
the jurisdiction of the village council. The objective of CBFM is to bring unreserved forests and
woodlands under management and protection. Under CBFM, village councils are given the right
to declare, enclose and manage parts of village lands as village land forest reserves, which are
registered with the district councils. The villagers take full ownership and management
responsibility of these village forest reserves, including the rights to harvest timber, collect and
retain royalties, and undertake patrols (including arresting and fining offenders). They are
exempted from local government taxes on forest products and are not obliged to remit any part of
their royalties to either central or district government (Blomley and Ramadhani, 2006). The
implementation of CBFM gets further support by the provisions of two key legal documents.
Firstly, the Village Land Act (1999) which recognizes customary tenure rights for village and
communal land and allows for its registration as “village land”, and secondly, the Local
Government Act (1982) that empowers village councils as executive agencies charged with the

17

supervision of adjudication of village land within their boundaries, including forests and
woodlands (Blomley and Ramadhani, 2006).

To summarize the two forms of PFM, JFM reserves all rights and benefits for the state while
encouraging local government institutions to help enforce access restrictions. Local government
can keep the proceeds of fines that it chooses to impose on intruders. CBFM, on the other hand,
is an effort to introduce management arrangements to forests that presumably were previously
ungoverned, and thus local government is given much stronger authority over their management,
including the rights to all economic benefits. Theoretically, the returns from forest conservation
under CBFM are vested with the local community thus providing the community with right
incentives to conserve as compared to JFM where there are limited or no returns to conservation.

Studies that have monitored forest conditions over time comment that decentralized forest
management results in improved forest conditions. Blomley et al., (2009) offer a comparative
evaluation of forest conditions in 13 forest sites under JFM and CBFM and show that forest
condition under JFM is better than CBFM in terms of variables such as stems per hectare, basal
1

area and volume increment. While the data from the two forest types are not directly
comparable owing to different base condition across forests, nevertheless they raise some
interesting questions about the returns to forest conservation and incentives to conserve under
JFM and CBFM regimes. Blomley et al. (2009) provide an explanation in the form of elite
capture in JFM forests; they argue that by concentrating benefits within a small group of people,

1

Basal area is the term used in forest management that defines the area of a given section of land
that is occupied by the cross-section of tree trunks and stems at their base. A higher basal area
signifies a better forest condition.
18

the incentives under JFM may become sufficient to maintain active patrols that exclude other
villagers. In an evaluation of coastal forests Persha and Blomley (2009) report that forest
conditions under JFM are worse than forests under CBFM. While comparing a CBFM, JFM and
an open access forest, Persha and Blomley (2009) conclude that greater tenure security and
institutional autonomy under CBFM resulted in maintenance of good forest conditions as
compared to forest managed under JFM. While underscoring the importance of a well-designed
institutional structure for forest management, these results are based on special circumstances
wherein the CBFM forests included in the study were privately owned and managed by a
cooperative group for a long time before the start of CBFM.

The outcomes of decentralized forest management in terms of enhancing livelihoods of
communities engaged in PFM are ambiguous. Persha and Blomley (2009) suggest that
generating livelihood benefits has remained a challenge under decentralized forest management.
Lund and Treue (2008) discuss that CBFM has been successful in generating forest revenues that
cover the costs of management and even fund local public services; at the same time, the poor
households who depend on forests for livelihoods are negatively affected.

Regarding the effect of decentralized forest management on governance, the few studies that
exist point to corruption at the level of the village council. Brockington (2007) questions the
remarkable successes of decentralized forest management in Tanzania. In a study from Iringa
region, he juxtaposes the claim that CBFM brings about better democracy and conservation, on
one hand, with detailed accounts of corrupt and violent practices of the village government on
the other. He suggests a need to reevaluate the purported success of CBFM. Lund and Treue

19

(2008) remark on the limitations of CBFM in achieving the goals of good governance. Lund and
Treue (2008) present evidence about corrupt practices and the tendencies to promote personal
economic goals prevalent within the village council, and suggest devising additional
accountability measures to ensure better governance outcomes.

This paper analyses the outcomes of forest governance in JFM and CBFM regimes in
Tanzania. The actors, power and accountability (explained in the next section) framework is used
to assess the practice of JFM and CBFM and compare it to the formal policy with a goal to
understand the gaps in implementation of decentralized forest management and what they imply
for the upcoming REDD regimes.

2.4. Conceptual Framework: Actors, Powers and Accountability
The actors, power and accountability framework (Agrawal and Ribot, 1999) is an analytic
framework to assess the degree to which the conditions necessary for meaningful
decentralization are represented in the framework of PFM in Tanzania. This framework suggests
three aspects that underlie all dimensions of decentralization.

Actors in decentralization: Actors in decentralization in the local arena who exercise powers
over public resources may include appointed or elected officials, NGOs, chiefs, powerful
individuals or corporate bodies such as communities, committees or cooperatives. Each of these
actors is typically located in particular relations of accountability and has certain types of
powers. These relationships depend on the historical, social and political constitution of powers
of each actor, which may be based on ideology, wealth, election or other means. Consequently,

20

the nature of decentralization depends to a significant degree on who gets to exercise power and
the accountability relations to which they are subject.

Types of powers: The framework distinguishes between four types of powers of decision making
that are crucial to understanding decentralization. These four types of powers correspond to
legislative (creation of rules), executive (implementing and enforcing decisions), and judicial
(adjudication of disputes). More specifically they are:
(1) The power to make rules or modify old ones is held in some domain sof decision making
over which governments seek to decentralize control, and in relation to some group of actors.
This set of powers allows actors to legislate principles that determine who can benefit from
given resources, how, and to what extent.
(2) The power to make decisions in some domain of action that influences others increases the
autonomy of the actor who gains these powers. Such powers enhance the discretionary
authority of local bodies. For instance, if a local body has powers of raising revenue, a degree
of decentralization has been achieved.
(3) The power to implement and enforce compliance to the new or altered rules implies the
power to execute, monitor whether actors are carrying out the roles that they are supposed to
perform and also includes the power to impose and enforce sanctions on those who do not
subscribe to the tasks they are supposed to perform.
(4) The power to adjudicate disputes that arise in the effort to create rules or modify existing
ones. It is important that such adjudication is done independently and without bias at a level
that is accessible to local constituents.

21

Accountability in decentralization: The effectiveness of decentralization hinges on
accountability. If powers are decentralized to actors who are not accountable to their
constituents, or who are accountable only to themselves or their superior authorities, then
decentralization is not likely to attain its stated aims. In order for decentralization to be effective,
the constituents need to possess countervailing power. Downward accountability of power
holders to their constituents broadens participation and ensures that decentralization is effective.
A number of ways exist through which actors can be made downwardly accountable to their
local constituents. These may include elections, procedures for recall, referenda, legal recourse
through courts, third party monitoring, political pressures and lobbying by associations, auditing
and evaluation, belief system of leaders and their communities, embeddedness of leaders in their
community, pride of leaders, social movements, threats of social unrest and resistance.

2.4.1 Characterizing the legal framework of PFM in Tanzania according to the Actors,
Powers and Accountability Framework
This section analyses the Tanzanian policy on JFM and CBFM on the three dimensions of
decentralization as per the actors, powers and accountability framework. The analysis of these
three dimensions of decentralization follows from literature on Tanzanian PFM policy (Blomley
and Ramadhani, 2006).
Actors:
As per this framework, the decentralized forest management reforms under JFM and CBFM
entitle actors at the lowest level, the village councils and village environment committees (VEC)
to participate in forest management. Both these institutions are bodies elected by the village
assembly, which comprises village residents 18 and older. While the actors in both PFM

22

arrangements are the VEC and village council, there is an important difference between the two
PFM regimes in terms of the powers that have been decentralized.
Powers:
Under CBFM the powes of the VEC and village council are clearly defined and vested with
them. The village council has been given the power to own village lands and the power to
declare part of this village land as village forest and mange it under CBFM. This implies that the
village council in consultation with the VEC has the power to make, enforce and ensure
compliance to rules for forest management, decisions as regards commercial exploitation of
forests and the power to impose and collect tax revenues from sale of forest products. The VEC
reports to the village council and depends on its approval in matters of apprehending and
sanctioning offenders.

In contrast, although JFM is termed as a participatory forest management arrangement, the
powers and autonomy of local actors is limited. Unlike CBFM, the VEC and village council do
not have the powers to decide which forest areas can be brought under JFM and do not have the
tenure rights over forests under JFM. The VEC participates in the decision to make rules of
protection however the approval of these rules is vested with the district or central government
authority. The VEC remains responsible for day-to-day management in the form of patrolling
forests, apprehending offenders, sanctioning them with support of the village council, and the
district forest offices. Although the VEC has the responsibility of patrolling the forests and the
power to enforce rules, it does not enjoy any share in benefits from the co-managed forests.

23

Accountability:
In terms of accountability mechanisms, both CBFM and JFM have downward accountability
mechanisms wherein the VEC reports to the village council, which in turn is answerable to the
village assembly through bi-annual meetings of the village council. However, in JFM along with
reporting to the village council, the VEC is also answerable to the district forest offices about
developments in the JFM areas and expects support in sanctioning offenders.

CBFM can be characterized as democratic decentralization as it entails the devolution of
authority of forest management to local level actors, who have appropriate legislative, executive
and judicial powers and are downwardly accountable to their constituents. Analyzing the JFM
regime through the actor, powers accountability framework reveals that under JFM the local
actors enjoy rather limited powers and autonomy, while the state continues to centralize
decisions as regards designation of JFM areas and control over benefit streams. These
centralizing tendencies do not qualify JFM as a form of democratic decentralization.

2.5. Study Area
The field research for this paper was conducted in the Uluguru Mountains in Tanzania. The
Uluguru Mountains are among Africa’s ten most important conservation sites (Burgess et al.,
2002). The research was conducted during 2008-2009 in three villages in the eastern part of the
Ulugurus. Another round of follow-up field work to triangulate and validate findings was
conducted in late 2010 and early 2011. The three villages in the study area are located in the
wards (administrative sub-unit of the district) Mikese, Gwata and Matombo of Morogoro district.
Before selecting the study villages a list of villages in Morogoro district under different kinds of

24

PFM arrangements was compiled. The compilation of this list revealed that some of these
villages were studied more frequently, while others were not. Based on this knowledge, three
villages were purposefully selected to represent the variation of PFM arrangements and the fact
that they were not researched previously. Three villages in Morogoro district were thus selected.
The village Muhungamkola represented the CBFM arrangement, Kibangile represented JFM,
and Gwata represented both CBFM and JFM.

Muhungamkola is located at a distance of 25 kms from Morogoro town along the Morogoro-Dar
highway on a dirt road about 10kms off from the highway. The village is heterogeneous with
members of Mluguru, Mkamii, Mkwere, Mzigua tribes. These groups are settled in the three subvillages of Muhungamkola, Kikulungundu and Nyagwambe. There are 283 households in the
village (Source: Village records). The village has been protecting two patches of CBFM forests
since 2004. They are called the forest of Kila towards the north-east and the forest of Ngongolo
towards the south. The forests represent the dry woodlands of the miombo type.

Kibangile is located on a dirt road about 45 kms away from Morogoro on the Matombo road.
The village has four sub villages, Gulioni, Kibangile Juu, Kinasimba and Vizi. The village has
298 households (Source: village records) and is predominantly inhabited by the Mluguru. The
catchment forest reserve Kimboza surrounds the village on three sides. The Kimboza forest
reserve holds a significant conservation value as the source of the tributaries that make up the
river Ruvu which is the lifeline of the city of Dar es Salaam. The forests represent low montane
forests and harbor many historic and sacred sites. The VEC was formed in the year 2004 and

25

JFM was initiated in partnership with the state and the four villages surrounding Kimboza,
including Kibangile.

Gwata is located on the Morogoro-Dar highway, about 40 kms from Morogoro town. The
village has five sub-villages, Lukwambe, Geza Ulole, Kaalia, CCM and Shuleni. The village is
heterogeneous and inhabited by members of Mzigua, Mkwere, Mluguru, Mkamii tribes. There
are 382 households in the village (Source: Village office records). The village has a JFM forest
towards west and a CBFM forest to north. While JFM forest is part of the central government
owned catchment reserve called Kitulanghalo, the CBFM forest is called Kimunyu. The VEC
was formed in the year 2002 and protection of both JFM and CBFM areas ensued thereafter. A
part of the JFM forest is used and managed as a research forest by the local agricultural
university in Morogoro. The university has paid two guards to patrol this forest. In 2007, the
university also initiated a pilot project that paid village community for good protection and
upkeep of its JFM and CBFM forests. These forests are representative of the dry miombo
woodland that characterize much of central Tanzania.

2.6. Methods
To address the research question about the governance outcomes of the two decentralized forest
management regimes, JFM and CBFM in Tanzania, this study relied on using qualitative
methods which included focus group discussions. Qualitative data are non-numeric, textual and
visual data derived from interviews, observations, documents or records, gathered from a small
number of informants (Chung, 2000). Qualitative methods are useful when research topics
involve exploring processes and meanings of phenomena under study (Reinharz, 1992).

26

A focus group consists of a small group of individuals from a well-defined target population in
order to generate discussion focused on pre-selected topics of interest to the researcher (Knodel,
et. al., 1990). Morgan (1988) suggests that focus groups be used in situations where the
researcher is new to the area and the research topic requires an exploration of topics. The main
feature of the focus groups is the explicit use of the group interaction to produce data and
insights that would be less accessible without the interaction found in the group (Morgan, 1988).

In each of the three study villages focus groups were organized to understand the current and
past use of forests by various social groups, current management rules, and functioning of village
environment committees and village councils. As the villages in the study area differed from
each other in terms of the type of institution involved in forest management as well as the
different types of communities who use forests for particular purposes, a study design to
compare and capture this heterogeneity was developed. Study villages were categorized as per
break and control variables (Knodel, 1990), to determine the number and composition of focus
groups to be organized. Break variables define how study villages are differentiated from each
other. Two sets of break variables were identified (table 2.2). These are: (i) the type of forest
management institution (JFM/CBFM), (ii) type of forest users: poor women, men and charcoal
makers. Together these break variables define different subsets for which separate focus group
sessions were held. While the groups were differentiated along various break variables, it was
also ensured that they share some common characteristics. Such characteristics that are shared by
all members of each group are referred to as control characteristics (Knodel, 1990). Membership

27

in village environment committee was identified as the control characteristic for identification of
focus groups.

Four focus groups per village were thus identified as per a combination of these break and
control variables. However, since all combinations of break variables were not found in the study
area, a few of these were eliminated. For instance, during fieldwork it was found that charcoal
makers do not exist in Kibangile so this category was eliminated from the list of focus groups to
be conducted in that village. Thus, focus groups were purposively selected to ensure a match
between break characteristics and actual situations on the ground. In addition, some
combinations of focus groups were replicated in each village to get responses from a larger
cross-section of the village residents. For instance, in village Gwata, two focus groups were
organized with charcoal makers as charcoal production is a major activity in the village.
Similarly, focus groups with women were replicated in Kibangile, Gwata and Muhungamkola. In
all, fifteen focus groups spread over these three villages were conducted (Refer table 2.3). The
final composition based on the combination of break and control variables of the fifteen focus
groups is listed in table 2.3.

The participants of the focus group were purposively recruited after conducting key informant
interviews with village chairpersons across selected villages. Each focus group discussion had
about ten members and followed a question guide which consisted of a number of open-ended
questions around management of forests, roles and responsibilities of various stakeholders,
problems and challenges, ownership of resources, past and present management practices, use of
forest, and people’s perceptions about the village council and VEC. All focus group discussions

28

were conducted in the local language Swahili, audio recorded and transcribed in verbatim. The
transcribed data were translated into English. This resulted in a rich set of textual data that were
analyzed using Carney’s ladder of analytical abstraction (Miles and Huberman, 1994). Carney’s
ladder of analytical abstraction is an analytical framework used to analyze qualitative data. The
ladder has three steps:

i. Summarizing and packing the data: This step of analysis involved transcription of recorded
data to create a text to work upon. These transcripts were read with an eye to look for codes
using key ideas from the actors, powers and accountability framework on decentralization. Codes
are tags or labels used for assigning units of meaning to the descriptive information gathered
during a research (Miles & Huberman, 1994). These first rounds of codes were descriptive in
nature (Miles & Huberman, 1994) and entailed little interpretation, and related directly to the
topics covered in the group discussions. A provisional list of codes thus emerged from this early
coding. These descriptive codes were later applied to the data to look for relationships among
codes and to arrive at a revised and enlarged code list. These second order codes were recorded
in a code book memo (Refer table 2.5 for code book)

ii. Repackaging and aggregating the data: The aim of this step in analysis is to search for
relationships in the data and find out the areas of emphases and gaps. At this stage, the
transcriptions from all sets of data were searched for relationships and patterns among codes.
Definitions and operational rules for application of codes were developed for each of these
codes, to help in the process of standardization and cross-comparison of the data (refer to table
2.5 for themes and definitions). Upon completion of coding a final family of inferential codes

29

emerged, termed as a structure. The final structure thus contained “larger” (more conceptually
inclusive) and “smaller” (more differentiated instances) codes (Miles & Huberman, 1994). This
final structure thus tied related codes or themes into a pattern or a meta-code. These meta-codes
elaborate upon explanations for the research problem (Miles & Huberman, 1994).

After the entire data had been coded, case summaries were prepared for each theme or concept
that emerged from the list of codes. For instance, case summaries were prepared for the theme
financial management, as it emerged as a theme central to assess the performance of JFM and
CBFM regimes. These case summaries were then combined in order to have an overall summary
statement for each major theme or concept. This step resulted in reduction of the entire
transcribed data, into three-four line summaries around major themes and concepts in the data.

iii. Developing and testing hypotheses to construct an exploratory framework: The aim of this
level of analysis is to cross-check findings through matrix-analysis of major themes in data and
finally integration of data into an exploratory framework. At this stage, the summary statements
developed in the previous step are sorted and organized as per important concepts that emerge
from the data. These summary statements are organized in form of matrices or grids, and can
take many forms, depending upon the nature of relationships and patterns among themes. A
matrices for the theme accountability mechanisms was prepared to examine the pervasiveness of
the poor information flow across JFM and CBFM villages (Refer table 2.6 for matrix on
accountability mechanisms)

30

2.7. Findings: Assessing PFM in practice according to the Actors, Powers and
Accountability Framework
The analysis of the Tanzanian PFM regimes through the lens of the actors, powers and
accountability framework shows that in principle, forest management has moved from being an
exclusionary management approach to a decentralized one where people have been handed the
responsibility of forest management. However, the extent to which these decentralized regimes
are also democratic remains in question. Clearly, JFM with its tendencies to centralize revenues
and tenure over forests and upward accountability structures cannot be termed as democratic
decentralization. On the other hand, CBFM with a high degree of devolution of powers to local
institutions is clearly democratic decentralization. At the same time, it remains important to
analyze the de facto performance of Tanzanian PFM to understand whether decentralized forest
management is performing as expected. Analysis of the village data assesses the reality of
democratic decentralization in Tanzania and shows the tensions and multiple interests that come
up repeatedly may potentially undermine governance outcomes even in a progressive
decentralized regime.

2.7.1 Actors and their interests in practice
There are multiple actors in the study villages, each having distinct interests that impinge on
forest use and management. While the village council and the VEC are the actors formally
entrusted with responsibility of forest protection and management, there are other actors such as
the farmers, graziers and charcoal producers whose interactions with the forest are mediated by
their livelihood choices, with implications on how forests are managed and powers are exercised.

31

VEC and forest management
The use of forests in JFM and CBFM areas is banned. This includes a ban on felling trees for
timber, firewood, poles. The rules of protection are strict and fines are stiff across JFM and
CBFM areas. In group discussions in the JFM and CBFM villages, the community members
mentioned refraining from use of JFM and CBFM areas. Women in particular said that the
closure of forests has not had any adverse consequences as there are plenty of wooded patches on
their private lands that they use to access fuel wood. There is no commercial timber logging in
any of the CBFM and JFM forests, there are incidents of illicit timber logging by powerful
members in the three study villages. The members in CBFM village Muhungamkola report better
outcomes for forest conditions in terms of return of some wildlife and stream flows from the
forest. Forest patrols were frequent and rules were enforced when offenders were apprehended.
During one of the forest patrols in which the author accompanied the VEC to the CBFM area the
VEC caught a person who was sawing timber for charcoal. The offender was brought to the VEC
who decided upon the sanction. The tools were confiscated and a stiff fine of Tshs 50,000 was
levied. Upon inspection, the record books of the VEC revealed that patrols were frequent in 2009
and fines were levied if offenders were apprehended. This incident confers confidence in the
ability of the VEC to protect the CBFM forests. Confidence in the ability of VEC is also
reflected in the responses of VEC members who sense a feeling of belonging and ownership to
the forests. Considering that these lands were under state control until a few years ago and face
heavy pressure from the local users given an extensive demand for charcoal, the ability of VEC
to continue with protection is remarkable.

32

Similar enthusiasm prevailed in the JFM village Kibangile as observed during the first round of
fieldwork in 2008-09. The VEC patrolled the JFM forests and earned revenue from visitors to the
forest and from offenders. It participated in fire fighting and received considerable support from
the district forest office in the form of training sessions and donation of bicycles and boots.
Farmers and forest management
Farmers are actors with interests that impinge on forest management. The farmers are interested
in gaining farmland. The village council can grant framers rights to new farmland if it is
available. The farmers slash trees on their newly allotted lands and carve out fields for
agricultural use. As of now there is no shortage of farmland, and most residents across the study
villages report not having to experience land shortages. However, much of the public land has
been allocated to non-resident outsiders who, having experienced land shortages in Morogoro
town, come the village to acquire land. Large contiguous patches of public lands, many with
wooded areas, have been allocated to outsiders who only farm parts of this land. Thus, across
villages, almost all public land has been allocated to big private farms, some spanning almost
500 acres. The only public land remaining in the village of Gwata and Muhungamkola is what is
now the CBFM forest. Since none of the farmers report that they currently face land shortages
this is not an important issue for the time being. However, as demand for land will likely
increase in the coming decades and since all lands have already been apportioned, CBFM forests
being the last patches of open public land could come under immense pressure. With the village
council vested with authority to convert village lands and earn revenue through these land
allocations, it can be expected that some lands from CBFM forests might be diverted for other
uses. The closure of public land is compounded by the absence of maps that document the extent
and location of village lands and the location of forests vis-à-vis village lands. As a result of this

33

ambiguous state where boundaries of private lands and forests are known to residents but not
documented, it is possible that parts of public lands that are currently forests are allocated as
private lands.

A case illustrative of this came to light during the author’s field visit in early 2011. In village
Gwata, in a follow-up focus group discussion with the VEC it was discovered that a substantial
part (5000 acres) of a CBFM forest called Kimunyu was under private ownership. The VEC
members forcefully asserted that in all probability the CBFM forests had been sold by the village
council without the knowledge of the VEC or the village assembly. The purported owner of the
CBFM forest is from Dar es Salaam, and plans to use the forest to log timber and mine for
limestone. An interview with the village council members revealed that the land transfer to the
owner predates the establishment of the CBFM forests. However, the VEC members harbor
mistrust in the word of the village council and maintain that the village council might be
involved in this deal. In fact, the VEC members are quite disgruntled with the audacity of this
private owner who wants to mine the forests that the VEC has protected since 2004.s They also
express frustration in this state of affairs wherein as the entity responsible for management of
CBFM areas, they neither have information about the transfer of CBFM forestlands nor do they
have the power to oust the occupant or ask the village council to explain things. When the issue
was brought to light of the district forest office, the nodal agency for CBFM, the officer in
charge did not comment upon the issue. At the time of the close of fieldwork, the matter
remained unresolved but the private owner had initiated fencing and mining in the patch of
CBFM that he claimed to own.

34

While not representative of all CBFM villages, the issue does raise some interesting questions
that will have implications for the implementation of REDD+ interventions. First is the mosaic
of land tenure arrangements on wooded areas that lie outside of protected forests that makes it
potentially impossible to control land and its use. Second is the absence of village level mapping
that documents and delineates of village forest lands from other public and private lands. This
kind of mapping and documentation is an absolute must for the ground level implementation of
REDD+, with implications not only for the design of projects but also to avoid potential land
grabs by the powerful who in absence of any maps may fence off forest areas and claim public
lands as private property.
Charcoal makers and forest management
Charcoal making although illegal in Morgoro district is still practiced, especially in villages that
fall along the Morogoro to Dar es Salaam road. Two of the study villages, Gwata and
Muhungamkola have charcoal producers. Until the mid 1990s, the district authorities issued
permits to make charcoal to groups of outsiders from Mbeya and Iringa. While these licensed
traders exploited village public land, the village council had no authority to stop them. However,
after the enactment of Tanzania Village land Act (1999) the public lands being exploited by
outsiders became village property. At around the same time, charcoal making was banned in the
district. In early 2000, parts of village public lands in the villages of Gwata and Muhungamkola
were enclosed to create the CBFM forests. While these outsiders made charcoal under
commercial licenses, local villagers also produced charcoal from the public lands but mainly as a
subsistence activity to supplement agriculture incomes. The charcoal makers in the two study
villages report using public lands that are now CBFM areas. Charcoal makers report a decline in
the number of bags and quality of charcoal that they produce annually. This is because they have

35

to scout for trees on private lands as public lands in these two villages are non-existent.
However, charcoal makers do not seem to complain about the decline in their productivity
because charcoal making is illegal and the general tone within the village is laden with the
rhetoric of conservation of CBFM areas. In interviews with the village council and the VEC, the
idea that all forests should be strictly protected and that there are no obvious losers of protection
of CBFM areas was pronounced across the two study villages. As a result, during 2008 in
Muhungamkola the patrols were frequent; in one instance a charcoal kiln that was in the process
of being built was destroyed by the VEC members in the presence of the author.

While the charcoal makers do not complain publically about this loss in production due to forest
closures, a conflict over forests is visible in the tension between charcoal makers, the village
council and the VEC. This conflict stems from the fact that the VEC is entrusted with protection
of CBFM areas. At the same time, charcoal making used to be a source of revenue for the
village council. However, with the protection of CBFM areas and the district ban on charcoal
making, the village council uses its authority to penalize charcoal makers through seeking bribes.
Of the 16 charcoal makers who were included in this study, a majority report paying bribes to the
village council members and specifically to the village executive officer at some point. The
charcoal makers cannot resort to any official means to address this conflict as what they are
doing is illicit. In fact the village council in Gwata is complicit with some charcoal makers who
are family members; the village council allows them to make charcoal with impunity in CBFM
forests. This puts the VEC in a state of quandary when they try to stop charcoal makers who are
aligned with powerful people in the village. There are spillovers of this stalemate to the other
residents of the village, who question the authority of VEC in enforcing rules in CBFM areas

36

when it cannot stop these charcoal makers with powerful connections. As a result, the state of
affairs in CBFM in Gwata is that stealing is common both by villagers and several charcoal
makers report to chopping trees in CBFM areas to produce charcoal. At the time of fieldwork
during 2009, the author witnessed a charcoal kiln right inside the boundary of CBFM forest in
Gwata.

Charcoal making in the study villages has implications for REDD interventions, especially as it
2

impacts leakage . If village forests constitute unit of project implementation under REDD, then
clearly activities such as charcoal making get displaced to other patches of wooded areas inside
the village. So while protection of CBFM areas may be in good shape, the other destructive uses
of forests may continue.
Graziers and issues of authority
Graziers are a third set of actors that impinge on use and management of forests and on the
authority of the village council and the relationship between VEC and village council. Migrant
graziers en route to the coast from the central parts of Tanzania come around each year and graze
their herds on public lands along the way. While this practice was not problematic in the past it
has become so now. With the decline of public lands and enclosure of CBFM forests in both
Muhungamkola and Gwata and several nearby villages, there is a conflict between farmers and
graziers and between graziers and members of the VEC. On the one hand, the graziers’ herds
frequently graze on private farmlands, sparking conflicts between these two groups. On the
other, the graziers take their herds into CBFM areas and when VEC members try to stop them,
the graziers reply that they have the permission from the village council. In addition, the village
2

In an avoided deforestation program like REDD, leakage refers to deforestation outside of the
program area that takes the place of deforestation inside the program area.
37

residents question the authority of VEC in stopping the residents from accessing CBFM areas
when it is unable to keep the graziers out. In general, there prevails a situation of mistrust, with
villagers assuming that the village council has accepted bribes from the graziers.

As a result during 2009, the residents of Muhungamkola took complaints to the district
authorities. The village executive officer was expelled from office in early 2010. However, at the
time of the author’s second round of fieldwork in late 2010, the issue of graziers remained
unresolved. The district forest officer of Morogoro suspected graziers’ connections to senior
people in the administrative hierarchy. Meanwhile, the village residents continue to push the new
village executive officer to address the issue with little knowledge of the fact that it might be too
difficult to address it.

In light of this discussion, it can be concluded that while on one hand decentralized forest
management, especially in CBFM has devolved considerable rights in Tanzania and has made
transparent the conflicts and divergent interests around forests which were so far neglected. As
evident from these examples, the marginal groups such as charcoal producers, village residents
feel empowered to challenge the village council and the VEC and to voice concerns about
addressing trade-offs associated with protection and the use of forest fines and revenues. At the
same time it has also given the local elite an opportunity to gain privileges as illustrated by cases
of rent seeking from charcoal makers and grazers.

38

2.7.2 Powers and their negotiation in practice
While decentralization efforts under PFM regimes grant powers to local communities and
institutional structures, the issues of who represents local people and how these representatives
use these powers remain unresolved but important questions.

The CBFM regime transfers significant powers to the village council over decisions to enclose
village lands as forests, levy duties, make and enforce rules, and sanction offenders. These
powers make CBFM a very forward-looking policy that has instilled a sense of ownership and
belonging amongst community members towards the CBFM forests. In all the study villages,
despite the multiple and often conflicting interests, community members express great pride over
their ability to participate in forest management. In the two CBFM sites of Gwata and
Muhungamkola, a sense of ownership can be discerned from the group discussions. In a group
discussion with elder men, there emerged a consensus “It is our village and our forests and we,
not the government, can oust the outsiders if they do not comply to our rules.” This feeling
conveys a sense that given the right institutional support, communities may be able to partake in
CBFM.

At the same time the complete vesting of powers with the village council has created avenues for
their misuse. For instance, the village council in Gwata has been involved in seeking bribes from
the charcoal makers routinely. Interviews with charcoal makers in this village revealed that they
were routinely apprehended by village council members who demand bribes. These bribes were
explained as tax that the charcoal makers were required to pay when charcoal making was not
prohibited. No receipts were issued for any of these transactions. In the same village, the family

39

members of the village council were complicit in charcoal making. Two sons from the extended
family of sub-village chairperson made charcoal in the CBFM forests with impunity.

Discussions with village residents and also members of VEC in Gwata revealed that there are
more incidents of offences within CBFM areas as compared to JFM forests. The reasons for this
is that people feel that since CBFM is their forest it is easy to negotiate lighter penalties because
one knows the people who enforce the rules. On the other hand, the village residents express that
the incidence of offences is quite low in JFM forests. The reasons that came out of discussions
were that JFM forests, although under co-management, are still considered state property for
which there is higher degree of regard as compared to CBFM areas.

The narratives of charcoal makers raises a concern about the issue where the village council
holds the legislative, executive and judicial powers over all matters. While vesting powers over
all domains of a village with a local level institution represents the strength of the decentralized
forest management in Tanzania, at the same time it also represents its weakness in terms of
achieving good governance. In fact such concentration of powers breeds corrupt and violent
practices as evidenced from the narratives of charcoal makers and graziers. At the same time,
this fusion of powers in one body undermines exercise of powers by other local institutions such
as the VEC, which is an important actor in decentralization. Increasingly, in the cases of charcoal
making or graziers the village council has made decisions that undermine forest management.
The VEC officially is the body in charge of managing forests, but it has little say to overrule the
village council’s decisions. The evidence brought forth by this research shows that the vesting of
VEC within village council has eroded the motivation of the VEC in upholding the system of

40

patrols. For instance, the VEC members in Gwata express a sense of disempowerment that stems
from their lack of authority to take action against offenders on CBFM forests (which are
technically part of the decision-making domain of the VEC) some of which are complicit with
the village council. There are spillover effects of this erosion of the powers of the VEC to the
other residents of the village, who question the authority of VEC in enforcing rules in CBFM
areas when it cannot stop these charcoal makers with powerful connections.

In JFM areas, owing to the very limited power that the VEC can exercise and the nearly nonexistent returns from forest patrols that the VEC members undertake has decreased the morale of
VEC in patrolling the forest. During the author’s first round of fieldwork in 2008-09, the VEC
members across the three villages were pretty enthusiastic about taking on the roles of protection
and patrol. They cited their reasons for undertaking these roles, without any direct monetary
incentives as doing the job with heart or “kwa moyo”. They equated the job of patrolling to other
voluntary activities that they are expected to perform in the village. In the initial years of PFM,
the district forest office provided support in the form of training, bicycles and boots. In addition,
the money collected from fines from offenders and fees from visitors to the forests was used to
compensate the VEC members. However, during the author’s recent field visit during 2010-11,
the VEC members were not enthusiastic about continuing with forest patrols. The VEC members
in the three villages feel fatigued as expressed by the members of VEC Kibangile who say, “We
have lost heart,” (Tumekosa moyo) and, “Why should we meet? We have no authority and we
have made no patrols in this year.” Among the reasons that were mentioned by VEC Kibangile
were that they do not receive support from the district authorities, they lack powers and authority
to decide over JFM forests and, and they do not gain any returns from patrolling JFM areas. As a

41

result, the patrols into JFM areas have become infrequent and the VEC members responded with
rd

delay in a recent fire outbreak that burned down a about 2/3 of JFM forests in 2010.

What appears on paper to be an ideal situation with regards to democratic decentralization, in
practice is mired with concentration of powers and their misuse by one group in the village to the
detriment of governance goals of equity and transparency. It has led to the erosion of powers of
other actors. This discussion points to the need for having additional mechanisms to ensure
firstly that powers are not concentrated at one level and second that there are checks and
balances to monitor use of powers.

2.7.3 Accountability in practice
While decentralization policies privilege actors at the local level as sources genuine of authority
they allow insufficient consideration for whether those actors are accountable to local people. As
pointed in the actors, powers, accountability framework, decentralization is incomplete if actors
are not downwardly accountable to their constituents. As analyzed in the previous section, in
principle the mechanisms for downward accountability are in place in the JFM and CBFM
arrangements. However, the embeddedness of the VEC and the village council in hierarchical
power relations challenges the exercise of downward accountability. In principle, the village
council is the authority vested with powers over forests in CBFM. Bi-annual meetings of the
village council are an accountability mechanism where village council shares information about
village matters with its constituents, the village assembly.

42

The de facto situation on the ground reveals that downward accountability is much more difficult
in practice. The accountability mechanisms are compromised on two accounts: Firstly, the
concentration of all powers with the village council leaves little space and authority to negotiate
with other actors in the decentralization process. Thus, the village council, at least in the cases of
Gwata and Muhungamkola, misused its powers. A second reason why accountability is difficult
in practice is that the actors who have the powers and their constituents are entrenched in
hierarchical structures of power that are difficult to overcome.

The inability of the VEC to demand accountability from the village council stems from the fact
that the VEC is a body within the village council and finds itself without powers to make the
village council answerable. At the same time, some information about the use of funds by the
village council is known to the VEC through members of VEC who also sit on the village
council. This ambiguous channel of communication presents various avenues of mistrust among
the VEC members. So for instance, in the village Gwata, funds were transferred from a local
university for a pilot project offering payment for avoided deforestation. This money was banked
by the village council, which planned to use it for village development. However, because a few
members of the VEC were represented on the village council, they came to know about this
money. At the same time, the other members of VEC remained unaware. The issue of the money
came up in a group discussion wherein a part of the VEC already had misgivings about misuse of
this money by VEC members who sat on the village council.

While annual meetings of village council are organized regularly, in discussions the village
residents pointed out repeatedly that matters related to forests and forest management revenues

43

are not disclosed in the meetings of the village council. Illustrative of the power divide is this
discussion with charcoal makers in village Gwata: in response to the question of whether they
demand (from the village council) financial records of taxes earned from the forest, the charcoal
makers responded, “We get air (as an answer).” They were of the opinion that they cannot ask
for the records since they are complicit in the illicit charcoal business and have no moral grounds
to subject the village council to interrogation.

To conclude, while downward accountability is difficult to achieve in practice, decentralization
of forest management has made local actors aware of their rights and the need to ask questions
about corrupt practices and use of monies from their leaders. Thus, while unsupportive of the
corruption within leadership, the village residents support forest conservation and expect their
leaders to use of funds for the benefit of the village and share information about it.

2.8. Relevance for REDD
One of the proposed outcomes identified for the UN-REDD pilot program in Tanzania includes
building the capacity of the institutions in decentralized forest management to implement
REDD+. Work on achieving this outcome would primarily be undertaken in two ways: first, by
implementing REDD+ at the village level on lands that fall outside the formal government
managed reserved forests; and second, albeit on a more limited level, by implementing REDD+
on the forest reserves co-managed by the state and the local communities (Burgess et al., 2010).
Implementation of REDD+ aims to make use of the PFM arrangements in two ways: by building
upon lessons from PFM implementation into the design of REDD projects, and by making use of
REDD funds as an incentive to instill conservation in PFM (Burgess et al., 2010).

44

Given the evidence from this paper, REDD implementation on lands that fall outside the formal
reserves i.e. lands that are under the jurisdiction of the village would be comparable to the
CBFM scenario discussed in this village. The existing institutional infrastructure of CBFM and
the ability of communities in the two study villages to successfully take charge of conservation
of village forests is indicative of the level of commitment that a REDD regime can expect. At the
same time, the issues of conflicting interests of actors and the fusion of legislative, executive and
judicial powers at the level of village council might undermine the equity, fairness goals of
REDD projects. Additionally, the availability of considerable REDD+ funds might create
opportunities of misappropriation given the track record of coercion and corruption in some
CBFM sites. These issues would be significant for REDD+ implementation and if remain
unaddressed, they can threaten the functioning, especially permanence of REDD conservation.
Thus, a major lesson that REDD+ projects need to take from the decentralized forest
management on village lands is the need to strengthen accountability mechanisms and establish
checks and balances to the powers of the village council in order to make REDD implementation
fair, equitable and permanent.

The implementation of REDD activities along with use of REDD funds to provide incentives to
conserve reserve would be similar to JFM in so far as conservation is on reserve forests, but
unlike JFM where there are no incentives to participation, REDD on reserve forests plans to
disburse funds as incentives to undertake conservation. The plan to use REDD+ funds to provide
incentives to local communities to participate in conservation on this category of land would
make for a good strategy. As the discussion on JFM implementation in this paper points out, the

45

VECs under JFM are demotivated and face challenges to continue with protection without any
tangible returns from JFM. The potential of REDD+ funds to support VEC and incentivize their
efforts may give a boost to the lethargy prevalent in JFM. However, there is a strong concern that
the reluctance of Tanzanian state to share revenues and returns from JFM might carry into
disbursement of REDD+ funds. Given the past record of the government on benefit sharing and
the protracted delays on coming to any conclusion on this issue, it is quite likely that REDD+
funds on co-managed lands might be centralized while the communities continue to bear the
costs of conservation. JFM and REDD+ implementation on co-managed forests should draw
lessons from CBFM, which indicates that when meaningful powers to manage and rights to
revenues and management are devolved, local communities are ready to take up the task of
managing natural resources. With REDD+ the availability of funds provide the added attraction
and incentive for communities to partake in forest management. Only cautionary note is that,
there needs to be mechanisms for close monitoring of exercise of powers and funds so as to
avoid the pitfalls of corruption and coercion that have surfaced in some CBFM sites.

2.9. Conclusion
This paper has analyzed the institutional context of decentralized forest management in Tanzania
with respect to governance outcomes and contrasted the de jure status of decentralization to the
de facto situation on the ground. There are a number of positive effects of decentralization, in
terms of ownership of forests and ability to take local level decisions. At the same time, some
negative outcomes in terms of equity, misuse of powers and poor accountability are lessons that
call for improvements in the way in which decentralized forest management operates in practice.

46

While not being able to generalize to the entire country, one thing that comes out clearly from
this paper is that where powers have been meaningfully devolved (such as in the case of CBFM),
decentralization has aided the ability of local communities to shoulder responsibility to conserve
village forest lands which were under state control until about ten years ago. This development
of independent local capacity and enthusiasm in conservation constitutes a strength of the
decentralization efforts. However, the concentration of power with the village government has
also bred corrupt and coercive practices. But, despite the prevalence of these practices in the
study villages, the village residents in certain instances have felt empowered to question and
challenge their leaders. While important these outcomes are presented here not as an argument
against CBFM but as a caution that future policies need to have additional accountability
mechanisms in place. Clearly, CBFM has lessons for JFM. JFM can also become a program with
broad base support based not on fear but on negotiated respect if meaningful powers are
transferred at the local level. The current state of apathy that VECs in JFM villages face could be
reversed if significant rights were devolved to these institutions.

Finally, the lessons about forest governance from the implementation of PFM have practical
relevance for the design and implementation of REDD+ projects in Tanzania and elsewhere.
Firstly, the design of REDD+ projects needs to account for the multiple actors at local level so as
to account for the interests of the powerful and not-so-powerful. As pointed out in this paper,
forest governance is affected by the socio-political structures. Therefore, the design of REDD
projects should be tailor made to fit these realities. This is especially important because REDD+
implementation entails a huge amount of funds to be used as incentives to scale up conservation.

47

However, if these power hierarchies and existing relationships are ignored, the funds under
REDD may create perverse incentives and enormous potential for misappropriation.

In conclusion, is Tanzania ready for REDD+? Since Tanzania has a favorable legal and policy
environment for REDD+ implementation on the majority of its forest land, it provides a good
institutional base for REDD. However, the implementation of these good policies requires
strengthening of accountability mechanisms in the existing participatory forest management
policy.

48

Appendices

49

Table 2.1: Actors, powers and accountability according to the legal framework of decentralized forest management in
Tanzania
Decentralization
variable
Actors in the local
arena

Types of powers
Power to make
rules/modify old ones

Power to make decision
to use resources

Power to ensure
compliance and
implement/enforce rules

JFM

CBFM

Village environment committee along with the
village council has joint authority to mange
forests along with district forest office;
ownership of forests vested with the state

Village environment committee along with the
village council has the full authority to manage
forests; ownership of CBFM forests lie with the
village council

de jure: rules made, altered with joint
participation of village environment committee
along with the district forest office, rules
approved by the district forest office.
de jure: made by the district office, no power to
use, harvest, sale any forest products

de jure: rules made, modified and approved by
VEC, Village council and village assembly

de jure: powers to harvest, sell, tax forest
products lie with Village council, VEC
participates in these decisions, rules shared with
village assembly.
de jure: responsibility of VEC to patrol,
de jure: responsibility of VEC to patrol,
monitor rule enforcement/compliance,
monitor rule enforcement and compliance,
offenders are apprehended by VEC and
offenders apprehended by VEC but sanctioned
sanctioned by village council with help from
by village council in consultation with VEC.
district office. Fines and penalties are stiff.
Fines and penalties are graduated and specified
in byelaws.
de jure: with district authority; VEC and village de jure: power to adjudicate disputes with
council to adjudicate on minor infractions only. village government.

Power to adjudicate
disputes
Accountability
mechanisms
Downward accountability de jure: VEC reports to village council, Recall
mechanisms
mechanisms for VEC, village council exists.

50

de jure: VEC reports to village council, Recall
mechanisms for VEC, village council exists

Table 2.2: Criteria for selection and composition of focus groups
Village

Control
VEC

Kibangile
√
Gwata
√
Muhungamkola √

Charcoal
producers
√
√

Break Variables
Women Men
Type of PFM
√*
√*
√*

√
√*
√

JFM
JFM&CBFM
CBFM

* signifies a replicate; √ signifies a focus group

Table 2.3: Focus group sessions
1. Village environment committee JFM village Kibangile
2. Village environment committee JFM and CBFM village Gwata
3. Village environment committee village CBFM Muhungamkola
4. Charcoal producers JFM and CBFM village Gwata
5. Charcoal producers CBFM village Muhungamkola
6. Women forest users JFM village Kibangile I
7. Women forest users JFM village Kibangile II
8. Women forest users JFM and CBFM village Gwata I
9. Women forest users JFM and CBFM village Gwata II
10. Women forest users CBFM village Muhungamkola I
11. Women forest users CBFM village Muhungamkola II
12. Men forest users JFM village Kibangile
13. Men forest users JFM and CBFM village Gwata I
14. Men forest users JFM and CBFM village Gwata II
15. Men forest users CBFM village Muhungamkola

51

Table 2.4: Focus group Guides
Village Environment Committee
1. What is the forest area under JFM/CBFM/sacred forests in your village? What is the name of
the forest? Which other villages have rights over it?
2. Are there any areas with trees outside of PFM forests? What are these used for? What are the
major threats to these areas?
3. Why was PFM initiated in the village? When did PFM activities start in the village? From
where did the village get the idea to initiate PFM?
4. When was the VEC formed? How many members? How was the VEC members selected?
Probe criteria for selection as per sub-village, women etc.
5. How was the area to be protected under PFM decided/demarcated? Who decided upon the
selection of this area? Did village people apart from the VEC and village council participate
in deciding the area? If yes, what was their role?
6. What was the area currently under PFM used for before the initiation of PFM? What was the
status of the area (in terms of vegetation, trees density, species diversity) before the start of
PFM? Record types of uses of PFM area and types of users (eg. Graziers, charcoal producers,
fuel wood collection).
7. What uses are currently allowed in the PFM area? Can fuelwood, timber,charcoal wood,
poles, medicines be collected?
8. What are the rules for management of the PFM area? Are these management rules strict?
Why did you decide on such strict rules? When will harvest be permitted? How were these
rules formed? Were village people apart from VEC and village council consulted?
9. What are the roles of the VEC in protection? What activities does the VEC do to protect the
poorest? Are their any guards employed to patrol the PFM area?
10. If someone from the village or outside is found engaging in illegal activities in the protected
forest what actions are taken? How is fine decided? Who collects it? What is fine money
used for?
11. If someone on the VEC engages in illegal activities related to protected forest what action is
taken?
12. What are the sources of income for the VEC? Are VEC members paid?
13. What motivates the VEC members to engage in patrol and protection?
14. In what ways does the district/catchment office assist you in PFM activities? Did you receive
money, training, fish ponds or other income activities? Who got these?
15. In your opinion, has the forest changed under PFM? In what ways?
16. In your opinion, have any forest areas outside PFM/public lands changed after the start of
PFM? In what ways have they changed- better or worse? What tree related uses from the
current PFM area have been displaced since the start of protection? To which lands/ other
forest area? What are the major threats to protection of forests outside of PFM?
17. In what ways have VEC members benefited from PFM? In what ways have the village
people benefited from PFM?
18. Have any community members who previously used the PFM area lost out because of PFM?
Have the VEC members discussed in what ways these people can be accommodated?
19. Have there been any conflicts between VEC and village community over use of PFM area?
Please discuss what was the issue in detail?
52

Table 2.5: Code Book: Forest governance in Participatory Forest Management in Tanzania
Theme
Knowledge of
program
(CFM/JFM)

Definition
Evidence of know
how about forest
protection programs
amongst village
communities
Guidelines that
specify how to
protect of PFM
forests

Rule
Direct or indirect
reference by
community members
in the village about the
functioning of PFM
projects
Direct reference to
rules that exist to
facilitate protection of
PFM lands

Rules for
protection of
general lands

RULGEN

Process of rule
formation

Rules of forest
protection

Code
Code name
Knowledge
KNOWPROG
about existence,
working of forest
protection
programs in
village
Rules for
RULPFM
protection of
PFM forests

PROCRUL

Guidelines that
specify how to
protect general lands
in the village
Processes and
negotiations through
which rules for
protection of forest
are made

Direct reference to
rules that exist to
facilitate protection of
general lands
Direct reference to
deliberations and
negotiations involved
in making of rules of
protection

Evidence that forest
offences are
committed

Direct reference to
instances of illegal use
of protected forests

Offences in
Stealing/theft in
forest protection forest areas

OFFFORST

53

Example
About five years ago they
reserved the two forests of
Kila and Ngongolo

So since it has been declared
you will find that you cannot
take any tools such as axe
inside there. If you are found
then you will be charged

We had our meeting where
this issue of reserve was
discussed. The ministry of
natural resources helped us to
make rules. We could not
have done by ourselves, we
could have failed. We are
lacking skills.
Yes, people do steal from the
reserved forests in Kimboza.
People steal wood and make
timber. You cannot tell who
does it.

Table 2.5 (cont’d)
Scarcity of
Absence of open
general lands
areas to extend
farming

SCRLND

Evidence that the
general lands in
village are getting
scarce

Direct reference to the
fact that general lands
are on the decline

RSNSCRLND

Ways in which
general lands are
diminishing

Direct reference to
ways in which lands
are getting scarce

Use of PFM
lands

Evidence about the
use of PFM lands

Use of general
lands

Evidence about the
use of general lands

Direct reference to
ways in which PFM
forests are used
Direct reference to
ways in which general
lands are used
Direct reference to
instances of discord
between herders and
village government or
between herders and
village environment
committee
Direct reference to
instances of discord
between village
government and
village environment
committee

Reasons for land
scarcity

Use of village
lands

Conflicts
around forest
management

Conflict with
herders over
forest protection

CONHERD

Evidence of
conflicts that exist
with herders as
regards forest
management

Conflicts within
Village forest
committee and
village
government

CONVEC

Evidence of
conflicts that exist
between village
environment
committee and
village government

54

I think if those areas were
open forests, even me I
would own a part of it. That
is why I have a small farm as
land is scarce
Because we have been
invaded by newcomers. They
take big areas of land. Now
then you see there is no open
areas in Gwata.
Legally we are not allowed
as this is a reserve. So no
harvesting will be done.
We get firewood and wood
for building from open lands
These pastoralists are here
for a long time. We had
serious problems.

The village forest committee
is responsible for evaluating
income from forest. But the
government has separated us.

Table 2.5 (cont’d)
Motivation of
Voluntary
Village forest
committee

MOTVOL

Support by
government

Financial
management

Perceptions of
fear about PFM

MOTGOVT

Use of fines and
monies from
offenders in
PFM
Corruption
Charcoal
regulation and
fines

USEFINE

Fear/respect for
JFM

FEARJFM

CHARCRPT

Descriptions of
voluntary ways in
which the village
environment
committee is
motivated to engage
in forest protection
Description of how
support from the
government
motivates the village
environment
committee to engage
in forest protection
Descriptions of use
of money collected
as fine
Descriptions of how
illicit charcoal
activity is regulated
through fines and
corrupt rent seeking
Descriptions of how
feelings of fear or
respect for
government property
act as deterrents to
forest offences in
JFM

55

Direct references to
feelings of voluntary
behavior, duty being
the reason for
engaging in protection

VEC is chosen by the village
and national government.
Their work is volunteer
work. There is no pay. This
is government work.

Direct references to
incentives from the
government/forest
department as being
the reason for
engaging in protection

Government gives them
incentives such as shoes and
bicycles

Direct reference to
knowledge of use of
money collected from
fine
Direct and indirect
references to instances
where charcoal makers
were subjected to rents

That is (money) is their
benefit. That amount is in the
stomach, it goes into the
crocodile mouth.
Yeah (for those who are
doing charcoal business) they
are paying something. No not
to serikali, mtendaji.

Direct reference to
instances where
people refrain from
using a JFM forest out
of fear or respect for
state authority

People are afraid to cut trees
so thy will get punished. It is
government forest for so
long.

Table 2.5 (cont’d)
Fear/respect for
PFM

FEARCFM

Accountability
in PFM

Information
sharing between
Village
government and
village
environment
committee

INFOVG

Power in
institutions of
forest
management

Power of village
government

POWVG

Power of village
forest committee

POWVEC

Descriptions of how
feelings of fear or
respect for
government property
act as deterrents to
forest offences in
CFM
Descriptions of how
information about
use of funds, rule
making is shared at
various levels within
the village

Descriptions of how
power to
control/regulate
forest management
is vested with the
village government
Descriptions of how
power to
control/regulate
forest management
is vested with the
village environment
committee

56

Direct reference to
instances where
people refrain from
using a PFM forest out
of fear or respect for
village authority

The village forests belong to
the village. You know the
committee and the people.
Serikali ya kijiji na kitu
kidogo.

Direct or indirect
reference to channels
that exist for the flow
of information
between village
government and
village environment
committee
Direct or indirect
reference to how the
village government
wields influence to
control/ regulate forest
management
Direct or indirect
reference to how the
village environment
committee wields
influence to control/
regulate forest
management

Now if they (village
government) are collecting
taxes, we in the committee
are not informed.

If you have little charcoal,
when you reach the road to
sell your charcoal is taken
away by the leaders. We
don’t know their procedures,
they use power.
As a committee we are not
free. I mean the
environmental committee has
problems. The village
council decides for us. When
we are needed they put us
there like furniture

Table 2.5 (cont’d)
Power of herders

POWHERD

Descriptions of how
power to break rules
is vested with certain
actors such as the
herders

Direct or indirect
references to instances
where herders have
gotten away with
breaking the rules of
forest management
and the perceived
reasons behind it

BENPPL

Descriptions of how
PFM benefits people

Benefits for
community
development

BENCDEV

Descriptions of how
PFM benefits
community
development

Direct references to
improvements in
quality of life as result
of PFM
Direct references to
improvement in
opportunities created
for community
development through
PFM

Livelihoods
displaced

OUTLVLH

Descriptions of how
initiation of PFM
has resulted into
displacement of
forest based
livelihoods of
certain groups

Benefits of PFM Benefits to
people

Outcomes of
PFM

57

Direct reference to
forest based
livelihoods (charcoal
making) being
negatively affected as
a result of PFM

If you look at the herder he
has 1000 cattle and he is not
a normal person. His cattle
may belong to political
leaders. That is why you see
they have more power. The
village government could
have more power, and it has
failed.
We are benefiting in terms of
rain, soil erosion

When guests come to pay a
visit to the forests, they
cannot enter without a
permit. They will enter with
a permit. So our village can
earn something and when it
goes to the village it is also
for our development
The charcoal maker and
farmer is affected. Because
when the farmer wants to
extend his farm, no more
areas. Charcoal maker cannot
even get 50 bags or even a
quarter of it. Here are no
more trees on open lands.
Forest is closed.

Table 2.5 (cont’d)
Forest condition

OUTFOREST

Descriptions of how
there is an
improvement in
forest condition as a
result of PFM

Direct reference to
cases where forest
condition has become
better after initiation
of PFM

It is better now. It has
improved a great deal since
protection.

Table 2.6: Matrix: Accountability mechanisms in JFM and CBFM in Tanzania
Summary Gwata
There does not seem to be formal avenues
for sharing of information among the VEC
and village government. This is almost
notable in instances where sensitive matters
such fines, rents or allocation of forest land
are concerned. The CBFM forest has been
sold and even though the VEC is supposed
to know about these details it is unaware.
The VEC has come to know of this only
because one of the member of VEC is also
on the village government council. In
another case, while charcoal making is
illegal the village government allows it and
seeks rent. The VEC feels disempowered to
deal with forest offences because the village
government overrides decisions in domains
that fall under the purview of VEC

Summary Muhungamkola
The VEC of Muhungamkola is skeptical about
the role of village government in safeguarding
forests. The VEC complains that the village
government has not shared matters of revenue
from forests and the matter of negotiations
with graziers. Two incidents point this out. On
the issue of revenue from forest, the VEC
found illegal logging in village lands that was
approved by the village council However, die
to the absence of any formal information
sharing between the two, the VEC found out
about this only after the forests on general
lands were logged. On the issue of graziers,
the VEC feels that the village council did not
share the negotiations that transpired between
the graziers and village council. As a result,
they are unable to challenge the graziers and
demand eviction.

58

Grand summary
The lack of any formal channel of
information flow between the VEC
and village government is an
important issue. The issue has three
dimensions: (a). overlap of the
domains of decision making of VEC
and village government; (b). VEC
reports to the village council on
matters that are under its domain of
decision making and (c). lack of
downward accountability of the
village council.
This has resulted into opportunities
for corrupt practices for the village
government and the resultant erosion
in the morale and authority of the
VEC.

References

59

REFERENCES

Agrawal, A. and A. Chatre, and R. Hardin. 2008. Changing governance of the world’s forests.
Science 320, pp 1460-1462.
Agrawal, A. and J. Ribot. 1999. Accountability in Decentralization: A Framework with South
Asian and West African Cases. The Journal of Developing Areas, Vol. 33, pp 473-502.
Angelson, A. with M. Broackhaus, M. Kanninen, E. Sills, W.D. Sunderlin, and S. WertzKanounnikoff. 2009. Realising REDD+: National Strategies and policy options, , CIFOR,
Bogor, Indonesia
Astill, James. 2010. Better REDD than Dead: a special report on forests. Economist, September
23, 2010.
Blomley, T. and H. Ramadhani. 2006. Going to scale with participatory forest management:
early lessons from Tanzania. International Forestry Review 8, pp 93-100.
Blomley, T. K. Pflienger, J. Isango, E. Zahabu, A. Ahrends and N. Burgess. 2008. Seeing the
wood for the trees: an assessment of the impact of participatory forest management on
forest condition in Tanzania. Fauna and Flora International, Oryx, Vol. 42, No. 3, pp 380391.
Burgess, N., Doggart., N., Lovett, J.C. 2002. The Uluguru Mountains of eastern Tanzania: the
effect of forest loss on biodiversity. Oryx Vol 36, No. 2. pp. 140-152.
Burgess, N., B. Bahane, T.Claris, F. Daneilsen, S. Dasgaard, M. Funder, N. Hagelberg, P.
Harrison, C. Haule, K. Kabalimu, F. Kilahama, E. Kilawe, S.L. Lewis, J.C.Lovett, G.
Lyatuu, A.R. Marshall, C. Meshack, L. Miles, S.A.H. Lilledge, P.K.T.Munishi, E.
Nashanda, D. Shirimia, R.D. Swetnam, S. Willcock, A. Williams and E. Zahabu. 2010.
Getting ready for REDD+ in Tanzania: a case study of progress and challenges. Fauna
and Flora International, Oryx, 44(3), 339-351.
Chung, K. 2000. Using qualitative Methods to Improve Data Collection and Analysis of the
Living Standards Measurement Surveys, In: Designing and Implementing the Living
Standards Measurement Surveys: Lessons from Ten Years Of Experience, M. Grosh and
P. Glewwe, eds. Washington DC: The World Bank.
Crook, R.C. and J. Manor.1998. Democracy and decentralissation in South-East Asia and west
Africa: participation, Accountability and Performance, Cambridge: Cambridge
University Press.
Intergovernmental Panel on Climate Change 2007. Climate Change 2007: Synthesis Report.
Intergovernmental Panel on Climate Change, Geneva, Switzerland.

60

Hufty, M., and A. Haakenstad. 2011. Reduced Emissions for Deforestation and Degradation: A
Critical Review. Consilience: The Journal of Sustainable Development, Vol. 5, Iss 1, pp
1-24.
Kaimowitz, D. and J.C. Ribot. 2002. Services and Infrastructure versus Natural Resource
Management: Building a base for Democratic Decentralization, International Conference
on Decentralization and the Environment, 18-22 February, Bellagio.
Lund, J.F. and T. Treue. 2008. Are we getting there? Evidence of decentralized forest
management from the Tanzanian miombo woodlands. World Development 36: 27802800.
Manor, J. 1999. The Political economy of Democratic Decentralisation, Washington, DC: The
World Bank.
Miles, L., and V. Kapos. 2008. Reducing Greenhouse Gas Emissons from Deforestation and
Forest Degradation: Global Land-Use Implications. Science 320 (5882), Pp. 1454-1455.
Morgan, D.L. 1988. Focus Groups as Qualitative Research. Thousand Oaks: Sage
Publications.
Oyono, P.R. 2004. One step forward, two steps back? Paradoxes oof natural resources
management decentralization in Cameroon. Journal of Modern African Studies, Vol. 42,
No.1, Pp. 91-111.
Persha, L. and T. Blomley. 2009. Management Decentralisation and Montane Forest Conditions
in Tanzania, Conservation Biology, Volume 23, No. 6, Pp. 1485-1496.
Reinhraz, S. 1992. Feminist Methods in Social Research. Oxford University Press, NewYork.
Ribot, J.C. and A. Larson. 2005. Democrtaic Decentralisation through a Natural Resource Lens,
eds J.C. Ribit and A.M. Larson, Routledge, Oxon.
Ribot, J.C. 2004. Waiting for Democracy: The Politics of Choice in natural Resource
Decentralization. Washington, DC, USA, World Resources Institute.
Ribot, J.C. 2003. Democratic Decentralization of Natural Resources: Institutional Choice and
Discretionary Power Transfers in Sub-Saharan Africa, Public Administration and
development, Vol. 23, No. 1, Pp. 53-65.
Ribot, J.C. 2002. Democratic Decentralization of Natural Resources: Institutionalizing Popular
participation, Washington, DC: World Resources Institute.
Ribot, J.C., A. Chahtre, and T. Lankina. 2008. Introduction: Institutional Choice and Recognition
in the Formation and Consolidation of Local Democracy. Conservation and Society, 6(1),
Pp. 1-11.

61

Ribot , J.C., A. Agrawal and A.M. Larson. 2006. Recentralising While Decentralizing: How
National Governments Reappropraite Forest Resources, World Development, Vol. 34.
No. 11, Pp. 1864-1886.
Smoke, P. 2000. Fiscal Decentralisation in Developing Countries: A Review of Current
Concepts and Practice in Y. Bangura (ed.) Public Sector Reform, Governance and
IInstitutional Change, Geneva: UNRISD.
Sunderlin, W., J. Hatcher and M. Liddle. 2008. From exclusion to ownership? Challenges and
opportunities in advancing forest tenure reform. Washington, DC, USA: Rights and
resources Initiative.
Wily, L.A. and P.A. Dewees. Undated. From Users to Custodinas: Changing Relations Beteween
People and The State in Forest Management in Tanzania.
Wily, L.A. 2000. Forest law in eastern and Southern Africa. Moving towards a community based
forest future? Unasylva, 203, Pp. 29-62.
UN-REDD. 2011. http://www.un-redd.org/AboutREDD/tabid/582/Default.aspx. Accessed April
19, 2011.
URT (United Republic of Tanzania). 2001. Project Implementation Mannua: Annex 2.
Improving Service Delivery for participatory Forest Management. Unpublished Report.
Ministry of Natural Resources and Tiurism, Forestry and Beekeeping Division, Dar es
Salaam.
URT (United Republic of Tanzania). 2003. Framework for participatory Forest Management.
Ministry of Natural resources and Tourism, Dar es Salaam, Tanzania.
White, A. and A. Martin. 2003. Who owns the World’s Forests? Forest Tenure and Public
Forests in Transition. Washington, DC, USA: Forest Trends.

62

CHAPTER 3: PAYMENTS FOR ENVIRONMENTAL SERVICES AND WOMEN’S
LAND RIGHTS IN THE ULUGURU MOUNTAINS, TANZANIA

3.1. Introduction
Payment for environmental services emerged in the 1990s as a way to help overcome the
problem that often conservation is not undertaken because the market does not value it, and as a
response to earlier approaches based on regulations and indirect incentives that had yielded
disappointing results (Ferraro and Kiss, 2002; Pagiola and Platais, 2002). It is based on a
recognition of the very real opportunity costs that land mangers often face in adopting
conservation practices whose benefits may accrue largely to others (e.g. Pagiola and Platais,
2002; Wunder, 2005). PES thus aims to help align the private interests of land stewards with the
value that wider society places on the services that nature provides (Engel et al., 2008). In that
sense, PES aims to mimic a market transaction where the price determines how much of the
good is produced.

The rise of PES brought hope that it could serve as a win-win approach for improving
environmental management while also offering a source of income to poor people (e.g.
Fearnside, 1997). At first glance the logic appears straightforward: important environmental
services such as biodiversity, carbon sequestration and watershed protection are produced by
landscapes widely populated by poor people, so paying those people to protect that landscape
should result in benefits to both the natural environment and the local people. NGOs and
development activists quickly began to consider ways to put PES to work as a tool for poverty

63

alleviation, while scholarly literature pointed out the (sometimes restrictive) conditions under
which in fact PES could help the poor (e.g. Landell-Mills and Porras, 2002; Pagiola et al., 2005).

On the other hand, Pagiola et al. (2005) and Wunder (2005) caution that PES is intended as a
way to promote environmental conservation and that it is unrealistic to expect a single policy
tool to achieve conservation and poverty alleviation objectives simultaneously. Tradeoffs
between the two objectives may arise in which case multiple policy tools may be required.
Landell-Mills and Porras (2002) pointed out conditions under which PES could actually harm the
poor. Kosoy and Corbera, (2010) argue that PES simplifies complex socio-ecological systems
and masks the social relationships embedded in the processes of producing and selling ecosystem
services. They emphasize the need to understand the power asymmetries and social relationships
embedded in the process of ecosystem service production. Sullivan (2009) stresses that as
monetary values for ecosystem services arise, other values (held by other people) may be left
behind, resulting in enclosures of newly valuable land and dispossession of the poor and
powerless. Swallow and Meinzen-Dick, (2009), Corbera et al., (2007) and Kerr et al. (2006) all
comment on the potential of PES projects to alter property rights relations, making certain lands
and land-uses “off-limits” for the poor.

The need for a sharper understanding of the problematic complexities of the relationship between
incentive-based conservation instruments and local social processes is important as there is much
discussion but little evidence of negative social impacts of PES. Moreover, the literature gives
little if any attention to potential gendered and intra-household outcomes of PES.

64

While actual empirical studies on the gendered outcomes of PES remain limited, there are
lessons to be gleaned from earlier efforts at environmental conservation that made land more
valuable. Scholars across the social sciences have long noted that commercialization and natural
resource improvement initiatives in developing countries risk transforming social relationships
around production in ways that harm the interests of less powerful segments of society (e.g.
Carney, 1993, Schroeder, 1997; and others discussed below). These studies show that when
market forces or a development project intervention raises the value of land or another natural
resource, people with power develop an interest in controlling that resource, often at the expense
of others who previously used it. It is logical to anticipate that PES interventions could have
similar impacts in situations characterized by weak institutions.

This paper illustrates the potential of PES to create contestations around gendered command over
resources. It presents a case study from the Uluguru Mountains in Tanzania in which monetary
incentives for tree planting have created tension around access to and control over land
resources. The case also illustrates the concern expressed by Kosoy and Corbera, (2010) that
valuation of ecosystem services tends to mask the local social relationships embedded in their
production and ignore resulting contestations and exclusions.

These ideas of exclusion of local people and invisibility of social relationships are explored in
the context of PES-based tree-planting projects that are being developed in the study area. The
Waluguru people of Eastern Tanzania are a matrilineal society, where traditionally women have
rights to own land. Local socio-cultural changes are leading to a patrilineal and patrilocal social
order, with women’s rights to own land shifting in favor of men. These social processes of

65

change are unfolding alongside increasing population pressure, land scarcity, and environmental
degradation. It is in this state of social and environmental flux that PES-based tree-planting
projects are being introduced. Local social norms link tree planting to proprietary rights to land,
and men in the area have manipulated these norms to their advantage. While these processes are
to a large degree endogenous, the proliferation of payment-based tree-planting projects has the
potential to accelerate it by providing men the motivation to plant trees on and thus claim rights
to lands belonging to their married women kin who live in other villages. This paper places these
ongoing processes of change as a backdrop to an ongoing PES project that aimed to promote tree
planting, and cautions that these dispossessions might worsen if the complex social context of
the area remains invisible within the PES discourse. It demonstrates the need for project
designers to understand and account for complex relationships around land rights if PES is to
help rather than harm the poor.

The paper is organized as follows. Following this introduction, the second section reviews three
interrelated strands of literature: first, the evidence of differential gendered and intra-household
effects from previous natural resource improvement projects; second, the cultural norms linking
tree planting and land ownership in much of Africa and the implications for women, and third,
the changing cultural norms among matrilineal communities in Africa and their implications for
women’s land rights. The third and the fourth sections introduce the methods and the study area
respectively. The fifth section presents the findings. The paper concludes with some
implications for future PES projects.

66

3.2. Gender and intra-household effects of environmental stabilization projects
A vast array of studies examining the intra-household effects of agricultural development and
environmental conservation projects in developing countries demonstrate that women’s interests
often are harmed in cases where monetization approaches to natural resource management
converge with customary tenure regimes undergoing transitions. Gray and Kevane, (1999)
provide evidence from across Africa about the local tensions that emerge when customary tenure
regimes that honor the claims of women undergo changes through processes of titling or
commercialization that render land valuable. Men have used their position of dominance to take
advantage of these processes to challenge women’s usufruct rights to such lands.

For instance, Carney (1993) in her work with Mandinka women in the Gambia documents new
claims over communal land tenure systems when introduction of irrigated paddy raised the
profitability of land use. Serious intra-household conflicts over land and labor resources resulted
wherein men as household heads claimed control over lands that were traditionally reserved for
women’s use. In another example from the Gambia, Schroeder (1997) illustrates that the
introduction of commercial fruit orchards by international development agencies led to the
imposition of new forms of property claims that worked against women. Such problems exist
beyond Africa; Rochleau and Ross, (1995) provide evidence from the Dominican Republic,
where the introduction of a so-called miracle tree species (Acacia mangium) for ecological
restoration generated unpredictable outcomes that undermined rural women’s land use strategies.

These exclusionary processes have occurred even in societies where matrilineal kinship systems
have prevailed. Gray and Kevane (1999) cite Tambiah (1989) as reporting that introduction of

67

perennial cash crops like tea or cocoa transformed matrimonial land into patrimonial properties
capable of being owned and transferred, with men exercising rights of ownership and excluding
women.

3.2.1 Tree planting and land tenure
Land tenure and tree-planting are interrelated in much of Africa, Latin America and Asia
(Fortmann, 1987). In principle, rights to trees do not extend to rights to land on which they are
planted, but in practice planting trees on a particular piece of land has come to signify rights to
that land under customary land tenure systems (Berry, 1988; Fortmann, 1985). Fortmann (1987)
discusses examples from Africa, where farmers enthusiastically adopted tree planting to mark
their proprietary claims to land. Evidence from West Africa also indicates that relatively strong
individual rights are granted to those who plant trees (Otsuka et al., 1997).

For decades, researchers have argued that promotion of tree planting projects has triggered the
privatization of rights to land. When trees were planted for commercial reasons, increased
planting led to a growth of a market in trees. This resulting commercialization of trees, together
with individual ownership of trees have been taken to mean that the spread of tree crop
cultivation has acted as an impetus towards privatization of rural land rights (Kobben, 1963 cited
in Berry, 1988). Some others point out that because of this relationship between tree-planting and
land ownership, an individual community member who acquires land through inheritance or
allocation may have a strong incentive to plant trees in order to consolidate their individual land
rights more permanently (Quisumbing et. al., 1999).

68

Tree planting projects tend to have gender and equity effects similar to those that have been
described for other agricultural development and environmental projects. This happens for two
reasons: first, trees can make land valuable thereby creating avenues for its appropriation by the
powerful. The case of horticulture development in the Gambia illustrates this point wherein fruit
crops added value to women’s garden lands resulting in men taking control (Schroeder, 1997).
Fortmann (1987) explains that the introduction of commercial tree planting in parts of Africa
resulted in the privatization of communally owned land, which in turn resulted in women losing
their usufruct rights to land. Secondly, more men avail of tree planting and acquire new rights to
land because in some societies there are explicit restrictions on women planting permanent crops.
This has resulted in tree planting being used as a means to get or maintain de-facto ownership
rights to land with prohibitions against women from planting trees lest they become owners
(Fortmann, 1985).

While both formal and customary restrictions may discourage planting of trees on communal
lands, land scarcity induced by population pressures and the erosion of customary authority have
led to poor enforcement of these rules (Fortmann, 1985). For instance, trees were planted on
lineage lands in the Uluguru mountains despite rules forbidding it (Maack, 1992).

If an agroforestry or land use project does not recognize the influence of tree planting on land
tenure, project-driven tree planting opportunities in a community may enable individuals to
establish de-facto ownership of land. Since PES projects provide cash incentives for planting
trees, similar gendered and intra-household contestations over claims to land can be expected. In
the context of prevailing customary norms that link tree-planting and land tenure it becomes

69

essential to examine who is planting the trees and where, and to be forewarned about the
unintended gendered outcomes of such projects.

3.2.2. Transition in matriliny and women’s land rights
Ethnic groups in Malawi and southeastern Tanzania such as the Chewa, Yao, Makonde, Makua,
and Mwera share many cultural traits and form part of a wider belt of matrilineal peoples
extending from the coast of East Africa to Central Africa (Dondeyne et al., 2003). The concept
of matriliny as a closed and complete system of social organization has been the subject of
debate. Among social theorists there is now an emerging agreement to consider matriliny in
terms of a cluster of characteristics (Peters, 1997). Among matrilineal peoples of East and
Central Africa these include a greater degree of autonomy for women in local politics and ritual,
control over land and income, uxorilocal marriage for men (explained below), tracing descent
from the mother, rights of the maternal uncle (instead of father) over his sister’s children, and
inheritance to men’s sisters’ children rather than to their own sons (Phiri, 1983; Richards, 1934
(cited in Peters, 1997); Holy, 1986).

Uxorilocal residence upon marriage means that upon marriage the husband comes to stay in his
wife’s village and provides labor on farms owned by his wife and her family. Thus as against a
patriarchal group, matrilineal land inheritance and uxorilocal residence together allow women to
maintain economic rights including independent control over land. Lovett (1997) illustrates the
significance of these two characteristics of matrilineal social organization for gender relations.
Among the lakeside Tonga people in Malawi, matriliny implied that as members of their
matrilineage women could claim certain economic rights, the most important of which was

70

independent control over land (Lovett, 1997). Uxorilocal residence upon marriage further
provided these women with a chance to retain control over their land and labor. This is in
contrast to the position of women in patrilineal groups where virilocality is practiced, meaning
that upon marriage a woman goes to live in her husband’s village. In this case patriarchal norms
guide a married woman’s access to land and it is mediated through her husband. Much research
on matrilineal organization has pointed out that the greater autonomy of women in these societies
is due to their control over land (Peters, 1997; Dondeyne et al., 2003).

Numerous studies point to gradual but major changes in matrilineal social organization such that
in some cases that virilocality has replaced uxorilocality (e.g. Phiri, 1983; Lovett,1997). These
processes of change have been going on over more than half a century and have proceeded
unevenly, such that certain features of matriliny have endured and others have declined.
However, there appears to be a general consensus among these research findings that a change in
residence upon marriage has largely undermined women’s authority, control of their labor and
3

rights to land in affinal villages (see Peters, 1997 for a review). Lancaster (1981), cited in Peters
(1997), in his study of the matrilineal Goba people in the Zambezi Valley, describes how
commercialization of agriculture and male migration led to a reduction in uxorilocal marriage,
resulting in an improvement in men’s independence as compared to women. He feared that the
intensification of these trends would lead to women becoming helpful wives of market-oriented
peasants rather than the more independent women of traditional villages.

3

Affinal means related to by marriage ties.
71

While these studies give an account of progressive loss of women’s authority and control over
labor as virilocal wives, they do not shed light on the status of these women as owners of land in
their natal villages. As members of a matrilineage these women have rights to land in their natal
villages as sisters and daughters. It is important to understand how the land rights of such
daughters and sisters have undergone changes. Lovett (1997) describes that traditionally,
compared to virilocally married women, Tonga women as sisters held superior authority in their
matrilocal villages as “owners of villages’. However, by the1950s there were moves away from
matrilineal practices, and uxorilocal residence was replaced by virilocal residence. The practical
consequence for women of adoption of these aspects of patriliny was the diminution of sociopolitical authority and economic security as land owners. Stripped of her identity as a sister who
exercised authority and owned land, a woman who resided as a virilocal wife became just a wife
and a daughter-in-law.

These social and cultural changes in matrilineal organization and the rise of patriarchal
organization enhanced men’s position as fathers and husbands. At the same time, women lost the
status and authority associated with their role as sisters, and they were repositioned to the
subordinate status of wives who had little control over land and their labor.

3.3. Methods
The field research for this paper was conducted in the Uluguru Mountains in Tanzania. The
Uluguru Mountains are an important center of floral and faunal diversity (Polhill, 1968) and
among Africa’s ten most important conservation sites (Burgess et al., 2002). The Ulugurus are
inhabited by the Waluguru, a matrilineal community.

72

The research was conducted during 2008-2009 in five villages in the eastern part of the
Ulugurus. Another round of follow-up field work to triangulate and validate findings was
conducted in late 2010 and early 2011. The five villages in the study area are located in the
wards (administrative sub-unit of the district) Kinole, Lundi, Matombo. During discussions with
government officials in Morogoro town, the administrative headquarters for the district, it was
found that villages in wards Kinole and Matombo are considered as the hub of Uluguru culture
and society. Thus, these wards were selected purposively to understand gender roles in a
matrilineal community. Ward Lundi was selected through a snowball approach after discussions
with village elders in Matombo and Kinole.

Qualitative methods, including interview and focus groups were used to collect data for this
research. Qualitative data are non-numeric, textual and visual data derived from interviews,
observations, documents or records, gathered from a small number of informants (Chung, 2000).
Qualitative methods were chosen because the research topic entailed understanding the meanings
and nuances of gendered dynamics around land. Qualitative methods are useful for exploring
processes and meanings of phenomena under study, and accessing participants’ ideas, thoughts
and memories in their own words rather than the words of the researcher (Reinharz, 1992).

Interviews: Qualitative interviews are conversations in which the researcher guides the
informants to elicit details and depth about the research topic (Rubin & Rubin, 2005). Rubin &
Rubin (2005) justify the use of open-ended interviewing methods in research topics that require
capturing the details and nuances of the social phenomenon under study. Interviews yield direct

73

quotations from people about their experiences, opinions and knowledge (Patton, 2001). These
interviews involve the use of an open-ended interview guide that outlines a set of issues to be
explored with the informant before the start of the interview (Patton, 2001). Topical interviews
are used to study research problems that are highly visible at the beginning of the study, and the
researcher interviews several informants to gather facts about events and piece together a
coherent explanation. Accordingly informants with a variety of perspectives are chosen who
have experience and knowledge about the problem (Rubin & Rubin, 2005).

Semi-structured, topical interviews were conducted with key informants such as village elders,
village government representatives, and individual women and men at various phases of the
research. The informants were selected purposively to include narratives of people who have
knowledge about gendered access to land in a matrilineal community and the changes that have
come about. A total of eleven open-ended interviews were conducted. All interviews were
conducted in Swahili. While a few of these interviews were tape-recorded the others involved
note taking.

Focus group discussions: A focus group consists of a small group of individuals from a welldefined target population in order to generate discussion focused on pre-selected topics of
interest to the researcher (Knodel, et. al., 1990). The main feature of the focus groups is the
explicit use of the group interaction to produce data and insights that would be less accessible
without the interaction found in the group (Morgan, 1988).

74

Ten focus groups were conducted with groups of men and women in the study villages. Nine
focus groups were with either men or women and one was with both women and men together.
The participants of the interviews and focus groups were purposively recruited after conducting
key informant interviews with village elders across the selected villages to include informants
who have knowledge about matrilineal social and cultural organization and gendered land rights.

Each focus group discussion had six members. Each focus group discussion followed a similar
question guide which consisted of a number of open-ended questions around themes related to
traditional social and cultural organization, traditional and current gender disaggregated land
inheritance patterns, relationships between trees and land rights and household division of labor.
All focus group discussions were conducted in the local language Kiswahili, audio recorded and
transcribed verbatim.

Phases of research: The study was conducted in two phases. The objective of the first phase
was to gain an in-depth understanding of the traditional and current gendered patterns of resource
management and control. This was achieved through focus groups and interviews with key
informants (such as school teachers, ward executives) and village elders (including men and
women) in Kinole and Matombo. This information was consolidated in the form of field memos
and was used to design questions for the second phase of the study. The objective of the second
phase was to get an in-depth understanding of the changes in gendered patterns of land
management through in-depth individual cases. Thus, in-depth interviews were conducted with
individual men and women to get details of land management and any contentions that occur
over women’s land. During the triangulation phase in 2010-2011, focus group discussions were

75

conducted where findings from the first two phases was shared with the respondents to get their
feedback on the validity of narratives collected.

Data analysis: The original data set from focus groups and interviews discussions consisted of
audio-recorded tapes. In case of interviews that were not recorded, interview notes were
reconstructed. All audio-data were transcribed in Kiswahili, and translated in English to produce
the textual data for the purpose of analysis. This resulted in a rich set of data that were analyzed
using Carney’s ladder of analytical abstraction for qualitative data. The ladder has three steps: (i)
summarizing and packing the data, (ii) repackaging and aggregating the data and (iii) developing
an exploratory framework to address the research questions (Miles and Huberman, 1994). This
ladder follows three levels:-

Summarizing and packing the data: The aim of this step is to create a text to work upon.
Transcription of recorded data from discussions and reconstruction of written notes was done at
this stage, using a word processor. At this stage, the transcripts of the initial data set were read
with an eye to look for codes. Codes are tags or labels used for assigning units of meaning to the
descriptive information gathered during a research (Miles & Huberman, 1994). A provisional list
of codes thus emerged from this early coding. These first rounds of codes were descriptive in
nature (Miles & Huberman, 1994) and entailed little interpretation, and related directly to the
topics covered in the group discussions. These descriptive codes were later applied to the data to
look for relationships among codes and to arrive at a revised and enlarged code list. These
second order codes were recorded in a memo.

76

Repackaging and aggregating the data: The aim of this step along the analytical ladder is to
search for relationships in the data and find out the areas of emphases and gaps. At this stage, the
transcriptions from all sets of data were searched for relationships and patterns among codes.
Definitions and operational rules for application of codes were developed for each of these
codes, to help in the process of standardization and cross-comparison of the data (refer to table
3.1 for code book). Upon completion of coding a final family of inferential codes emerged,
termed as a structure. The final structure thus contained “larger” (more conceptually inclusive)
and “smaller” (more differentiated instances) codes (Miles & Huberman, 1994). This final
structure thus tied related codes or themes into a pattern or a meta-code. These meta-codes
elaborate upon explanations for the research problem (Miles & Huberman, 1994).

After the entire data had been coded, case summaries were prepared for each theme or concept
that emerged from the list of codes. For instance, case summaries were prepared for the themes
traditional land management, women’s land rights as daughters, wives, widows etc. These case
summaries were then combined in order to have an overall summary statement for each major
theme or concept. This step resulted in reduction of the entire transcribed data, into three-four
line summaries around major themes and concepts in the data.

Developing and testing hypotheses to construct an exploratory framework: The aim of this
level of analysis is to cross-check findings through matrix-analysis of major themes in data and
finally integration of data into an exploratory framework. At this stage, the summary statements
developed in the previous step are sorted and organized as per important concepts that emerge
from the data. These summary statements are organized in form of matrices or grids, and can

77

take many forms, depending upon the nature of relationships and patterns among themes. For
instance, matrix diagram was prepared to compare and contrast various examples in data that
illustrate how women have lost their land rights (refer table 3.2 for matrix on women’s land
rights) .

3.4. Study Area
The Uluguru Mountains are part of the Eastern Arc Mountain range, an important center of floral
and faunal diversity and endemism (Polhill, 1968). The Ulugurus are home to at least 16
endemic vertebrate and 135 endemic plant taxa, a degree of endemism exceptional in tropical
Africa that puts these mountains among the continent’s ten most important conservation sites
(Burgess et al., 2002). The Ulugurus also form the most important water catchment in Tanzania
as the source of the Ruvu River, which supplies water to Dar-es-Salaam, the largest city. The
forest reserves in the Uluguru mountains have been ranked at the highest level of importance in
terms of their water catchment and biological value. However, recent surveys across the range
report a decline in forest area and potential disappearance of a few endemic faunal species. They
point to conversion of forest to agriculture as the major reason for forest loss (Burgess et al,
2002).

3.4.1. PES and smallholder carbon and tree planting projects in the Ulugurus
PES is being explored as an instrument to promote landscape conservation in these mountains.
Several pilot PES projects are being implemented, including projects that pay people to conserve
existing forests and plant trees on their agricultural land to provide carbon sequestration and
hydrologic services (Katoomba Group, 2007). Valuing the Arc is a Cambridge University project

78

that aims to create an inventory of important environmental services that the mountains provide
and use this information for project design and implementation (www.valuingthearc.org). A pilot
project by Sokoine University of Agriculture pays communities to protect existing forests against
deforestation and associated carbon emissions (Zahabu, 2008 ). CARE is implementing the
Equitable Payments for Watershed Services (EPWS) project, which aims to compensate farmers
for investing in new land use practices such as tree-planting to reduce the flow of silt into the
Ruvu River. Another active PES project in the region is the Pro-poor Rewards for Environmental
Services in Africa (PRESA) project, initiated by the World Agroforestry Centre (ICRAF).
EPWS and PRESA promote agroforestry and farm forestry on farmlands to provide both national
environmental benefits in the form of reduced silt loads in the Ruvu River and carbon
sequestration as a global environmental service. The CARE and ICRAF projects are already
underway, with tree planting having been introduced.

3.4.2. Social organization among the Waluguru
Traditionally the Waluguru were a matrilineal society, differentiated by clan and lineage (Young
and Fosbrook,1960). Among the matrilineal Waluguru, marriages were uxorilocal, i.e. in the
locus of the wife. Traditionally under matriliny, all land belonged to the matrilineage and was
referred to as lukolo. The maternal uncle, the mjomba, was the custodian of this land.
Membership in a lineage ensured rights to lukolo lands. There was no concept of individual
ownership for lukolo land; only usufruct rights were granted to this land and the land could revert
back to the lineage if the occupants did not use it.

79

Both male and female members of the lineage were eligible for allocation of lukolo land in the
first generation. However, in the second generation the land could be inherited by heirs on the
female side only (as they remained full clan members) as opposed to the heirs on the male side
(since they belonged to the matriclan of their mother). This meant that while a male member in
the first generation may be entitled to lukolo land, his share of this land would be inherited by his
nieces and nephews (on his sister’s side) and not by his own children. In this way, the mjomba as
custodian of the interest of the lineage land and as a facilitator of its inheritance to his nephews
and nieces exerted a greater authority over his nieces and nephews as compared to his own
children (Young and Fosbrook, 1960).

Traditionally it was not permitted to plant trees on lineage land. This is because when trees are
planted and they mature they become the property of the planter. Thus, even while the land
belonged to the lineage the trees would belong to the planter. During the 1950s with the
introduction of plantation crops such as coffee, and tree-planting to produce timber or control
soil erosion, this distinction between ownership of land and ownership of trees was lost. Even
though technically the land belonged to the lineage, since it was covered with trees that the
planter controlled and the lineage group was not in a position to buy back the trees, the rights
over land were de facto associated with the planter (Young and Fosbrook, 1960).

As with other matrilineal systems, the Waluguru society has undergone a social transition, in part
due to inroads from colonialism, Islam and Christianity that brought pressure against matriliny
(Maack, 1992). Over the years, lukolo land has been privatized. According to historical accounts
of the area, the privatization of lukolo land occurred under a complex set of social and

80

environmental factors. Land scarcity emerged with population growth and became pronounced
during the 1960s and has continued since. Around the same time, colonial policies curtailed the
authority of the institution of mjomba (Young and Fosbrook, 1960, Maack, 1992).

Lukolo lands once allocated were not returned to the lineage pool under the reversionary right.
Simultaneously, there was a rise in awareness among Waluguru men about fathers’ rights.
Fathers used tree planting on lukolo lands to express their investment in land and eventually to
bequeath this land to their children. Thus, more and more men started bequeathing their share of
lukolo land to their own children rather than to their nieces and nephews. The mjomba, having
lost his power and operating in a context of land scarcity, could do little to safeguard the
common pool of lineage lands (Maack, 1992).

This movement of fathers’ rights to children and subsequent substitution of individual rights of
ownership for clan ownership were more pronounced in the eastern part of the mountains where
the villages in this study are located (Maack, 1992). These transitions have hybridized the
Waluguru social structure with elements of both matriliny and patriliny. As a result, there a rise
in patriliny along with a redefinition of rules of inheritance, locus of marriage and access and
control over land, with significant implications for women’s social status and control over land.

3.5. Findings
3.5.1 Matriliny, Patriliny and women’s land rights
Currently, the Waluguru social organization can be described as being on the crossroads of
matriliny and patriliny. While elements of patriliny such as virilocal marriages (in the man’s

81

village) and land inheritance from the father as opposed to the mother are dominant, at the same
time matrilineal tendencies such as children inheriting the clan identity from their mother are
also practiced. The decline of matrilineal traditions in the realm of land inheritance has had an
effect on women’s rights to access resources as daughters in their natal and as wives in affinal
villages.

As discussed in the previous section, women as daughters in their natal villages traditionally
inherited clan lands or lukolo. The presence of lukolo land was a kind of security for a woman, as
her rights to this land were not subject to any challenge. She could maintain her rights to this
land and they would be upheld by the matrilineal custom. However, over time the lukolo has
been privatized. As a result, lukolo land as a category is non-existent as corroborated by a
majority of respondents in the study villages. Of all the respondents in the present study, only a
small percentage report to have inherited lukolo land from their mother. The loss of lukolo has
meant that women’s land rights as daughters and full members of the matriclan are diluted.

A woman from Tambuu explains what the loss of lukolo lands signifies for women:

“On lukolo, it is from your mother’s clan and so mjomba had decision [about its
allocation]. But now mjomba is not powerful and this land is not there. So you cannot get
this land.”
The privatization of lukolo has meant that most land is now inherited from the father. Both male
and female children are entitled to this and. The inheritance usually follows Islamic law,
according to which daughters receive half the share of a son. Increasing land scarcity and Islamic

82

inheritance rules have meant that women’s shares of these family lands are smaller as compared
to their male kin.

In principal, the rights of women to lukolo land are inalienable. So if a woman is married within
her natal village, she maintains her rights to such land by cultivating it. And if a woman is
married virilocally and stays in another village or town far from her natal village, her share of
lukolo land is held in the custody of the male members of her family. As per the focus group
discussions, the rights of virilocally married women to this land must be upheld by their male
relatives. Should a woman get divorced and return to her natal village, she can claim this land as
the only land over which she has exclusive rights to. An elder man from of Ngongolo explains
how inheritance of lands from the father works:

“Yes, the children will inherit the land [from father]. They also inherit some land from
her mother’s side. But mostly it is from her father where they inherit land. The son would
look after land that was for her sister.”
With a change in the locus of marriage from uxorilocal to virilocal, there is another category of
land over which women have rights. Referred to as lima, this is the land on which a woman has
rights as a wife in her husband’s village. However, as opposed to lukolo or the land inherited
from the father where women have exclusive rights of ownership, rights over lima do not
translate into ownership. Women as virilocal wives have usufruct rights to lima. As wives they
invest labor to produce subsistence food crops, cash crops, and permanent tree crops.

83

The rights of women to lima are tenuous and they do not own it upon widowhood or upon
divorce. In the event of widowhood, the woman holds the land in custody for her children. In no
case does the widow own this land. Should a man die without leaving hiers, conflicts among his
kin and his widow over the custody of land are common with the widowed woman often forced
out of her husband’s land. To avoid such conflicts and secure the claims of his wife a number of
respondents in the study villages mentioned that a husband may buy a piece of land for his wife
so that her rights are not extinguished in the event of widowhood. However, such cases where
men buy land for their wives are still quite uncommon.

While widowed women hold use rights to lands as their children’s custodians, the land rights of
divorced women are non-existent. In the event of a divorce, a woman is not entitled to any
parcels of land from the husband. In principal, divorced women are entitled to a share of land or
cash value of permanent tree crops that she helped plant on the lima. However, in practice it is
difficult for women to enforce their claims.
A divorced woman in Tandai explains this situation:

“I am tired and I have not gotten anything so far. Its been so long, I was

supposed to

get a share in two shambas, I have left trying now”
A change in the availability of clan lands along with a shift in the locus of marriage among the
matrilineal Waluguru has resulted in a diminished status for women in terms of their authority,
control of land in natal villages and rights to their labor in affinal villages. Since lukolo as a land
category is non-existent, their rights as daughters in a matrilineal system are largely extinct. As
daughters, their rights to land in natal villages extend to land that they inherit from their father.

84

However, their rightful share of these lands is diminishing under the influence of Islamic
inheritance rules. Finally, as virilocal wives, not only do they have no rights to own land, but in
the event of divorce the labor they invest in planting and maintaining permanent crops and trees
on their husband’s farm remains inadequately compensated. Thus, changes in socio-cultural
organization among the Waluguru have resulted in women losing control over resources in both
natal and affinal villages.

3.5.2 Contestations on women’s lands
Contestations around land remain common in the study villages. Often they have to do with the
land of absent owners. It is in this context of absentee ownership that contestations around
women’s land are happening. Study respondents agree that more and more marriages in the study
villages are virilocal, wherein women move away from their natal village to their husband’s
village. These respondents also mention that the rights of virilocally married women to the lands
inherited from their father in their natal villages are customarily protected by their male relatives.
Interviews with older men suggest that women’s rights to these lands are secure; however, in
reality there are conflicts around this issue.

Discussions with virilocally married women reveal that these women are aware of their rights to
lands that they inherit from their father and expect that their rights will be upheld. However, their
responses also suggest a sense of despair in that they know that it would be difficult for them to
claim their rights from their male relatives. A woman in village Tambuu discusses the divide
between what women think are their rights, their expectation about the fulfillment of these rights
and the current status.

85

“I have inherited land from my father. After I married I moved to my husband’s place and
the land that I inherited from my father is currently being used by my brother. [The rule is]
The lands belonging to daughters (if used by sons) are to be handed to daughters. I think that
I can claim it [my land] but I think that it will be difficult.”
The lands that virilocally married women inherit from their father in their natal villages are the
locus of conflict. These conflicts revolve around the fact that the male relatives who are
supposed to be custodians of women’s land are unable to honor women’s rights. Interviews and
discussions with men on the issue of rights of virilocally married daughters revealed a general
sentiment that women have rights to land from their father and that these should be upheld. But
at the same time, men also expressed a sentiment that virilocally married daughters should farm
with their husbands and thus their hold over rights to land in their natal villages should be
allowed to lapse. While women expect that they might be able to claim their lands, men express
their inability to honor women’s claims. Male respondents express these mixed feelings and
place them in the context of land scarcity and the fact that land needs to be bequeathed to
children as they do not inherit anything from their mother’s side. A quote by a man from Tandai
village resonates these sentiments:

“The land from father is divided among my brothers and sisters. The sisters are entitled
to rightful claim on this [paternal land]. However, I might not be able to honor their claims as I
have children and she [sister] is farming in her husband’s village just like my wife… My children
are entitled to land from my wife’s clan but they cannot get it as it is finished.”

86

3.5.3. Tree planting and dispossessions
As evident from historical accounts of the area there have been ongoing struggles among the
Walgurus to consolidate claims to land. These struggles have been driven by both material
exigencies to own land in face of acute land scarcity as well as for symbolic reasons as land
ownership confers a sense of belonging. As discussed in earlier sections, tree planting in the area
is a step towards ownership of the land where trees are planted. Maack (1992) describes that men
as fathers planted trees on clan lands to secure these material and symbolic aspects.

Historically, tree planting on lukolo lands has been used to privatize lineage lands and prevent
their reversion to the lineage. Currently, men have used tree planting as a means to claim lands
that their virilocally married sisters have inherited from their father.

While it is clear to men that tree planting on these lands confers ownership and may dispossess
their sisters of rights to land, men’s narratives do not reflect this. They explain that they have
planted trees on behalf of their sister and leaving this land unclaimed and thus unplanted may
invite contestations from strangers. Another explanation that men provide is that they are
planting trees on the land of their sister so that she may not be able to bequeath this land to her
children. This is because as per matrilineal inheritance, children are entitled to land that their
mother owns, and thus men as brothers perceive that their nephews and nieces may make claims
to their mother’s land. These men as fathers are also aware that their children may not get any
land from their maternal clan, so they need to secure land to bequeath to their children.

87

Interviews with men reveal this concern for their children when they say that they have used tree
planting on their sisters’ land not as a way to dispossess them, but for the benefit of their
children. A man from Tandai explains that he has planted trees on his sister’s land so that in
future his children may have land.

“My sisters they have a share in father’s land. But I might not be able to honor their
claims as I have planted trees on this land. My children need land. There is no clan land;
brothers take over the sister’s land.”
Women’s narratives on the other hand allude to a sense of dispossession. As daughters in natal
villages, their rights to lands that they inherit from their father are coming under threat as male
relatives fail to uphold their rights and in many cases attempt to control their land through
planting of trees. A woman from Tandai explains,

“And now if she [a woman] has to leave on marriage, this land that she can get from her
father, she does not get it. Brothers take care of sister’s land, use it but do not return [it].
Now there are conflicts, when brothers are taking care of sisters’ land and they [sisters]
stay far but cannot get their land back. Yes and I know of women where brothers have
conflicts on land of their sisters”
There are numerous such examples in the study villages where men have used tree planting to
appropriate women’s land. Many of these cases pre-date the PES based tree planting projects
that have been introduced in the area. In order to ascertain the incidence of such cases among the
local farmers who were allocated PES contracts under the PRESA project, we planned to launch
a survey of these farmers. However, during the focus group discussions organized before

88

designing the survey, two local key informants who had helped organize the distribution of treeplanting contracts objected to the survey as being too sensitive and obstructive. They suggested
that asking about land rights would make it difficult to get any answers from the farmers about
other aspects of the PES project. At the same time, one of these same key informants admitted
having planted trees on the lands of their own sisters, albeit not under the PES project (in which
they were not awarded contracts). When asked about the potential of the project in accelerating
these dispossessions they were dismissive of such a possibility, explaining that they had ensured
that only those farmers with clear land tenure be supplied trees. In fact, given the method of
contract allocation, it would have been impossible for them to impose such restrictions. To date,
under the PRESA project there is no procedure to ensure that the project awards contracts only to
farmers with clear, uncontested land tenure, so it is likely to unintentionally lead to more
contestations.

3.6. Implications for PES
While tree planting and the resultant dispossession of women’s lands are ongoing, there is reason
to expect that the tree-planting PES projects may accelerate the situation. This might happen as
the incentives offered under PES projects create new opportunities for men to adopt tree planting
and in the process make claims to lands of their female relatives. This scenario of dispossession
of women from their lands is plausible because such dispossession is already happening, and
because people in the study villages are attracted by the idea of receiving money to plant and
maintain trees. The ongoing PES projects are specifically designed to be pro-poor and have
invested effort in understanding the conditions under which poor people can participate in them,
but they operate on the incorrect assumption that land rights are stable and well defined. At the

89

same time, they have not focused attention on the issue of gendered land rights and gendered
division of labor, the underlying link between tree planting and land rights, and the changing
social and cultural context. While some of the PES literature has suggested the need to account
for multiple, competing claims to land, in practice such contextual factors tend not to be
accounted for. The analysis presented here has addressed this critical gap in the literature and
provided field evidence of what can be expected if PES projects continue to neglect social
dynamics surrounding landscape conservation and production of environmental services.

New PES projects need to better account for the social context of the places in which they aim to
work, including power asymmetries and the politics around natural resources that emerge when
monetary incentives for conservation are introduced into complex and changing social and
environmental contexts. The social conditions found in the Ulugurus related to the transition
from matriliny to patriliny and the interlinkage between land and tree rights also exist in many
other parts of sub-Saharan Africa, so the findings of this paper are directly applicable to such
places. More generally for PES programs that specifically aim to help the poor, it is important
that they go beyond simply targeting poor households and try to understand and address the
subtle intra-household relationships that further influence the welfare impacts of a project
intervention. It is not too late for conservation projects operating in the area to take these issues
into consideration, and it is important that new projects elsewhere work harder to gain an
understanding of the social context that could help anticipate problems that may arise when
insecure land rights, power asymmetries and monetary incentives cross paths.

90

Appendices

91

Table 3.1: Code Book: Matriliny, Tree tenure, and Women’s land rights
Theme
Traditional
matriliny
practices
Changes in
matriliny
practices
Women’s
land rights

Code
Traditional cultural
and social practices of
matrilineal
organization
Changes in social and
cultural practices of
matrilineal
organization
Traditional land rights
under matriliny

Code name
TRDMAT

Current land rights

LRTS CURENT

CHNMAT

LRTS MATRI

Land rights of
daughters

Current rights as wife

LRTS WIFE

Current rights as
widow

LRTS WIDW

Definition
Description of any social or
cultural practices associated
with matrilineal social
organization
Description of changes in
traditional matrilineal
organization

Rule
Direct reference to specific cultural
and social practices

Description of ways in
which women could own/use
land in matrilineal social
order
Description of women’s
rights to own/use land as
they stand now.

Direct mention of specific rights to
specific lands (for instance, clan
land) that women had under
matriliny
Direct reference to the rights of
access/control that women have over
specific lands (such as clan,
inherited from father, mother or
husband’s land )

Description of how brothers
are responsible to look after
the share of sisters land in
father’s estate
Description of rights to
own/use land as a wife as of
today

Description of rights to
own/use land as a widow as
of today

92

Direct reference to changes that
have occurred in cultural and social
practices

Direct reference to the kinds of
rights women have to own/use
specific lands (such as clan,
inherited from father, husband’s
land) marriage
Direct reference to the kinds rights
women have to own/use specific
lands (such as clan, inherited from
father, husband’s land) as a widow.

Table 3.1 (cont’d)
Current rights as
divorcee

LRTS DIVORC

Description of rights to
own/use land as a divorcee
as of today

Description of various
categories of land that
women could hold rights to
in natal and patrilocal
villages.
Description of changes
(whether weakening or
strengthening) in women’s
rights to land as owners or
under conflicts around these.
Descriptions of patrilocality
of marriage as a reason of
change in women’s land
rights
Description of tree planting
as a reason through which
women lose control of land

Categories of
Land over
which
women had
rights
Change in
women’s
land rights

Traditional categories

LNDINH
CATEG

Examples of changes

CHNG EXMPL

Reasons for
changes in
women’s
land rights

Patrilocality of
marriage

RSN
PATRIMARG

Tree planting

RSN
TREEPLTG

Scarcity of lands

RSN
LNDSCRCE

Religion

RSN RLGION

Description of land scarcity
as a reason for loss or
change in women’s land
rights
Description of religion as a
reason for loss or change in
women’s land rights

93

Direct reference to the kinds rights
women have to own/use specific
lands (such as clan, inherited from
father, husband’s land) as a
divorcee.
Direct reference to various
categories of lands that women
could have rights over- clan lands,
lands from father, lands from
husbands
A record of specific cases or
conflicts where women’s rights to
land as owners have changed are
under contestation
Direct reference to patrilocality of
marriage as being a reason for a
change/contestation in women’s
land rights.
Direct reference to tree planting as
being a reason for a
change/contestation in women’s
land rights.
Direct reference to decline in clan
lands or land in general as being a
reason for a change/contestation in
women’s land rights.
Direct reference to religious
influence as being a reason for a
change/contestation in women’s
land rights.

Table 3.1 (cont’d)
Tree tenureLand tenure
links
Women’s
strategies to
cope with
dilution of
land rights

Trees and land rights
relations
Strategies to address
loss of land rights or
changes in land rights

TREELANDRTS Description of land rights
and their relationship with
tree planting
STRCOPWOM
Description of how women
cope with loss or change in
land rights

94

Direct reference to tree planting
influencing/altering land rights as
being a
Direct reference to strategies or
mechanisms that women adopt to
deal with loss in land rights

Table 3.2: Matrix: Women’s land rights
Traditional rights
under matriliny
Under a matrilineal
system of social
organization all land
belonged to the clan (s.
Ukoo). The clan
elders/maternal uncle
was responsible for
allocation of these
lands. This land is
called the lukolo. The
women and their
children were entitled
to a share in this lukolo
land. The maternal
uncle facilitated access
of rightful heirs to
these lands. Permanent
crops could not be
planted on this land and
this land reverted back
to the clan in case it
was not being used by
the owners. Another
category of land is the
lima which women
farm at their husband’s
house. Women may (if
jointly acquired) or

Rights as
wives

Rights as
daughters

Rights as
widows
The rights of
widowed women
The women get Nowadays, women are mediated
through their
as a rule get land
land rights as
from their fathers. children. They
daughters and
do not own any
This is the gani
also as wives.
land of the
land. They may
As daughters
deceased
also inherit from
they inherit
husband but are
land from their their mother if
considered
land is available.
father (or
custodian of
However, as pr
mother) and
these lands for
Islamic law the
thus have
their children. A
women’s share is
ownership
half of what a man husband can buy
rights on it.
a piece of land
receives. So in
Upon
for his wife as it
marriage, they many cases, it is
is possible that
farm the lands the brothers who
upon his death
that are owned inherit more than
sisters. In case the the husband’s
by their
kin may take
husbands. The women are
back the land
married locally,
women have
that belonged to
they maintain
usfructuary
him and the wife
rights over this
rights over
land. It is this land may remain
farms and
landless,
that they claim
homestead
especially if she
when they get a
gardens at her
does not have
divorce.
husband’s
children. If she
place. In some

95

Rights upon
divorce
Upon divorce, a
woman has to
leave her
husband’s house.
She is not entitled
to any land from
her husband except
in the case if it was
acquired jointly. If
any permanent
crops were planted
she may or may
not be entitled to
them or collect
fruits. The trees
belong to the
children. She goes
to her father’s
house with her
young children and
starts a life with
land from her
father.

Grand Summary
Lukolo lands were
clan lands which
were allocated to men
and women.
However, this land
could not be owned.
Women as virilocal
wives also gained
rights to lima, land
from their husbands.
Whereas widows
hold rights to lima,
divorced wives’
rights to lima do not
exist. With changes
in land ownership,
lukolo was privatized
and lands started to
be inherited from
father. Sons and
daughters were both
given land.
Daughters’ land is
called the gani. With
virilocal marriages,
gani lands are
contested.

Table 3.2 (cont’d)
may not (if it is
inherited by the man
from his family) have
ownership to it. If
divorced, women
cannot claim this
category of land unless
it was acquired jointly.
A third category of
land is the gani, which
it seems as a category
has come up with the
individualization of
land and inheritance to
ones own children
rather than to one’s
nieces or nephews.
This gani is the
category of land that a
women inherits from
her parents (mostly
father) and is her
exclusive property.
Upon divorce, this is
the land that she can
claim (neither lukolo
nor lima).
Since most of lukolo
land has become
individualized and
there is hardly any land

cases, the
husband can
buy a separate
piece of land
for the wife so
that she owns
it and in the
event of
widowhood
that land is not
claimed by the
husband’s kin.
In case a
daughter is
married
locally, she has
the land from
her father and
her husband’s
lands in the
same village. If
a new piece of
land is to be
acquired, it is
the man who
approaches the
village
government.
This new land
belongs to the
husband and he

marries again,
then she has to
leave everything
for her children.
If there are no
children from the
marriage, the
land is divided
between the
widow and the
deceased
husband’s kin.

96

Table 3.2 (cont’d)
in this category left for
allocation to women.
At the same time, with
the scarcity of this land,
the social institution of
the maternal uncle or
mjomba has also
become weak so that
lukolo land inheritance
to women has declined.
Thus, land inheritance
of lukolo lands on
which women had
rightful claims has
come to a decline
because of social
change and material
scarcity of land.
Since marriages are
virilocal distant, it is
the gani land which has
come under
contestation.

may decide to
apportion a
part of it to his
wife. Thus,
while in the
past a wife
could own land
independently
(as clan lands)
now her land
rights are
mediated
through a man
as a husband or
father.

97

References

98

REFERENCES

Berry, S. 1988. Property rights and rural resource management: the case of tree crops in West
Africa, Cah. Sci. Hum. Volume 24, No. 1, pp 3-16.
Burgess, N., Doggart., N., Lovett, J.C. 2002. The Uluguru Mountains of eastern Tanzania: the
effect of forest loss on biodiversity. Oryx Vol 36, No. 2. pp. 140-152.
Carney, J. 1993. Converting the Wetlands, Engendering the Environment: The Intersection of
Gender with Agrarian Change in the Gambia, Economic Geography. Vol. 69, No. 4, part
2, pp 329-348.
Corbera, E., Kosoy, N., Martinez-Tuna, M. 2007. The equity implications of marketing
ecosystem services in protected areas and rural communities: case studies from MesoAmerica, Global Environmental Change. 17, pp 365-380.
Chung, K. 2000. Using qualitative Methods to Improve Data Collection and Analysis of the
Living Standards Measurement Surveys, In: Designing and Implementing the Living
Standards Measurement Surveys: Lessons from Ten Years Of Experience, M. Grosh and
P. Glewwe, eds. Washington DC: The World Bank.
Dondeyne, S., Vanthournout, E., Wembah-Rashid, J. A.R., Deckers, J.A. 2003. Changing Land
tenure regimes in matrilineal village of South Eastern Tanzania, Journal of Social
Development in Africa. Vol. 18, No. 1, pp 7-32.
Engel, S., Pagiola, S., Wunder, S. 2008. Designing Payments for Environmental Services in
theory and practice: An overview of the issues, Ecological Economics. Volume 63, No. 4,
pp 663-674.
Fearnside, P. M. 1997. Environmental services as a strategy for sustainable development in rural
Amazonia, Ecological Economics. 20, No. 4, pp 53-70.
Ferraro, P., Kiss, A. 2002. Direct Payments for Biodiversity Conservation, Science. 298: pp
1718-1719.
Fortmann, L. 1985. The Tree Tenure Factor in Agroforestry With Particular Reference to Africa,
Agroforestry Systems. 2, pp 229-251.
Fortmann, L. 1987. The Importance of Land and Tree Tenure In Agroforestry. Perspectives in
Agroforestry, Washington State University, Pullman, WA,8pp.
Gray L., Kevane, M. 1999. Diminished Access, Diverted Exclusion: Women and Land Tenure in
Sub-Saharan Africa, African Studies Review. Vol. 42, No. 2, pp 15-39.

99

Holy, L. 1986. Strategies and Norms in a Changing Matrilieal Society: Descent , succession and
inheritance among the Toka of Zambia, Cambridge University Press.
Katoomba Group. 2007. East and Southern Africa Katoomba Group Newsletter,
http://www.katoombagroup.org/~katoomba/documents/newsletters/Vol1.No3.html#anch
or2, accessed on 27 August 2010.
Kerr, John, Foley, C., Chung, K., Jindal, R. 2006. Sustainable Development in the Clean
Development Mechanism: Constraints and Opportunities. Journal of Sustainable
Forestry. 23(1): 1-18.
Kobben, A.K.F. 1963. Land as an object of gain in a non-literate society, In: Biebuyck, (Ed.).
Kosoy, N., Corbera, E. 2010. Payments for ecosystem services as commodity fetishism,
Ecological Economics. Vol. 69, pp 1228-1236.
Lancaster, C.S. 1981. The Goba of Zambezi: Sex Roles, Economics and Change. Norman:
University of Oklahoma Press.
Landell-Mills, L., I. Porras. (2002).Silver Bullet or Fool’s Gold: a global review of markets for
forest environmental services and their impact on poor. Instruments for sustainable
private sector forestry series, IIED, London.
Lovett, M.L. 1997. From Sisters to Wives and ‘Slaves’ : Redefining Matriliny and the Lives of
Lakeside Tonga Women, 1885-1955. Critique of Anthropology. Volume 17, No. 2, pp
171-187.
Maack, P. 1992. The Waluguru are not Sleeping: Poverty, Culture, And Social Differentiation in
Morogoro Tanzania. Unpublished PhD dissertation, Northwestern University, Evanston.
Miles, M., Huberman, A.M. 1994. Handbook of Qualitative Research. Thousand Oaks, Sage
Publications.
Morgan, D.L. 1988. Focus Groups as Qualitative Research. Thousand Oaks: Sage
Publications.
Otsuka, K., Suyanto, S., and Tomich, T.T. 1997. Does Land Tenure Insecurity Discourage Tree
Planting? Evolution of Customary Land Tenure And Agroforestry management In
Sumatra. Environment and Production Division, International Food Policy Research
Institute, Washington, D.C.
Pagiola, S., Platais, G. 2002. Payment for Environmental Services, Environment Strategy notes,
No.3. World Bank, Washington, D.C.

100

Pagiola, S., Arcenas, A., Platais, G. 2005. Can Payments for Environmental Services Help
Reduce Poverty? An Exploration of the Issues to Date from Latin America, World
Development, Vol. 33, No. 2, pp 237-253.
Patton, Quinn, P. 2001. Qualitative Research & Evaluation Methods-3rd edition. Sage
Publications.
Peters, P. 1997. Revisiting the Puzzle of Matriliny in South-Central Africa, Critique of
Anthropology.Vol. 17 (2), pp 125-146.
Phiri, K.M. 1983. Some Changes in the Matrilineal System among the Chewa of Malawi since
Nineteenth Century, The Journal of African History. Vol. 24, No. 2. P 257-274
Polhill, R.M. 1968. Tanzania, Conservation of Vegetation in Africa South of Sahara. In:
Hedberg, I. and O. Hedberg, O. (Eds.) Acta Phytogeographica Suecica, 54. pp. 166-78.
Quisumbing, A., Payongayong, E., Aidoo, J.B., Otsuka, K. 1999. Women’s Land Rights In The
Transition To Individualized Ownership: Implications For The Management Of Tree
Resources In Western Ghana, FCND Discussion Paper No. 58 Food Consumption and
Nutrition Division, International Food Policy Research Institute, Washington D.C.
Reinhraz, S. 1992. Feminist Methods in Social Research. Oxford University Press, NewYork.
Richards, A.I. 1934. Mother-right among the Central Bantu, In: Evans-Pritchard, E.E., Firth, R.,
Malinowski, B., Schapera, I. (Eds.), Essays Presented to C.G. Seligman, Kegan-Paul,
London, pp 269-279.
Rocheleau, D., Ross, L.1995. Trees as Tools, Trees as text: Struggles over Resources in
Zambrana-Chaucey, Donminican Republic, Antipode. Volume 27, No. 4, pp 407-428.
Rubin, H.J., Rubin, I.S. 2005. Qualitative Interviewing: The Art of Hearing Data. (Second
edition). Sage Publications, UK.
Schroeder, R.A. 1997. “Re-claiming” Land in the Gambia: Gendered Property Rights and
Environmental Intervention, Annals and Association of American Geographers,. Volume
87, No. 3, pp 487-508.
Sullivan, S. 2009. Green Capitalism and the Cultural Poverty of Constructing Nature as Service
Provider, Radical Anthropology. Issue 3, pp 18-27.
Swallow, B., Meinzen-Dick, R. 2009. , Payment for Environmental Services: Interactions with
Property Rights and Collective Action. In: Beckmann, V., and M. Padmanabhan (Eds.),
Institutions and Sustainability, Springer, New York, pp 243-265
Tambiah, S. 1989. Bridewealth and Dowry Revisited. Current Anthropology, Vol. 30, No. 4, pp
413-27.
101

Valuing the arc, http://www.valuingthearc.org/about-us/index.html. accessed on 21 March 2011.
Wunder, S., 2005. Payments for Environmental Services: Some Nuts and Bolts. CIFOR
Occasional Paper No. 42. Center for International Forestry Research, Bogor.
Young, R. and Fosbrooke, H. 1960. Smoke in the Hills: Political Tension in the Morogoro
Distict of Tanganyika, Northwestern University Press, Evanston.
Zahabu, E. 2008. Sinks and Sources: A strategy to involve forest communities in Tanzania in
global climate policy. Doctoral Dissertation. University of Twente, The Netherlands.

102

CHAPTER 4: PRO-SOCIAL BEHAVIOR AND INCENTIVES: EXPERIMENTAL
EVIDENCE FROM THE ULUGURU MOUNTAINS, TANZANIA

4.1. Introduction
In the past decade many payment for environmental services (PES) projects have been
established in developing countries (Wunder et al. 2008, Swallow et al. 2009). In most of these
projects payments are made to individual landholders who make individual land-use decisions,
but increasingly they are being introduced for environmental services that are provided
collectively by a group of natural resource managers. A group of users may jointly manage a
common pool resource (CPR) or separately manage individual parcels that are bundled together
in a PES contract to minimize transaction costs (Kerr and Vardhan 2010, Swallow and MeinzenDick 2009). The prevalence of CPRs in developing countries (e.g. Baland and Platteau 1996,
Agrawal 2007) makes it likely that group PES contracts will continue to spread. For example,
PES is being explored as a way to compensate communities all across the tropics under the new
Reduced Emissions from Deforestation and Forest Degradation (REDD) regimes (Miles and
Kapos 2008).

The PES literature has given little attention to the question of how introduction of incentive
payments influences collective action to manage a common pool resource that provides an
environmental service. The explicit assumption underlying PES is that offering sufficient
payment will motivate users to adopt conservation practices (Ferraro and Kiss 2002; Wunder
2005); where contracts are collective there is an implied assumption that collective action will
take care of itself. However, a long line of literature on common property resource management

103

makes it clear that eliciting such collective action can be difficult (e.g. Baland and Platteau 1996,
Ostrom 2002). This literature also documents that rural areas of developing countries often have
longstanding traditions of norm-based collective action that govern various aspects of civic life
(Cleaver 2000, Jodha 1985). Meanwhile, research from social psychology and behavioral
economics shows that where collective action is driven by social, non-pecuniary norms,
introduction of monetary incentives can undermine these social norms and thus weaken instead
of strengthen collective action (e.g. Fehr and Falk 2002). To the extent that natural resource
protection requires collective action, such perverse outcomes of incentive payments could
undermine environmental service provision.

It becomes important to examine the incentive structures under which PES promotes rather than
undermines collective action to deliver environment conservation, and whether alternative,
noncash payment types could have better properties for collective action (Kerr and Vardhan
2010). Interaction between payments and social norms, and the incentive types that local actors
might prefer are critical areas for research with strong implications for PES and REDD. In this
paper we investigate the effect of different levels and types of payment or reward, whether
individual or group and cash or kind, on individual decisions to contribute to manage a common
pool resource. We also examine contextual factors that can affect these relationships and
preferences.

We undertake this work through survey and experimental methods carried out in six
communities in the Uluguru mountains in Tanzania. Our main findings are that a high percentage
of people contribute to communal tasks regardless of the compensation regime; high payments

104

can increase participation but low payments can reduce participation relative to when no
payment is made; payments can reduce the satisfaction people derive from participating; social
approval incentives play a strong role in motivating participation; and people express preference
for collective rewards in the form of infrastructure rather than cash payments made either to
individuals or groups.

4.2. Human motivation, incentives and collective action
Sources of motivation to cooperate
The literature describes two perspectives on the sources of human motivation to cooperate:
individual rational utility maximization and social utility maximization (Vatn 2008). Models of
individual rational utility maximization assume self-regarding individuals, who perform actions
based on some internal cost-benefit calculations (eg. von Neumann and Morganstern 1947).
When faced the decision of whether to contribute to a public good, this model assumes that the
best option for self-interested, rational individuals is either to free-ride or to contribute only
when they can obtain private gains from doing so or face sanctions if they do not. Thus, within
this viewpoint, human motivations are extrinsically determined through incentives, whether
rewards or sanctions. Based on these models, early theorists like Olson (1965) and Hardin (1968)
were sharply pessimistic regarding the prospects for collective action among independent actors.

Social rationality implies behaving in a way that supports what is best for a group that one is part
of, even if it involves foregoing some specific personal gains or opportunities (Vatn 2008).
Literature on management of common pool resources (CPRs) (Ostrom 2002) and experimental
economics (see Vatn, 2008 for a review) supports the social rationality viewpoint and shows that

105

individuals exhibit a much stronger propensity to cooperate for provision of a common good than
expected under the rational actor paradigm.

People might be intrinsically motivated and contribute effort without any expectation of a reward
(Deci 1971), or they might be acting out of “impure altruism,” deriving a “warm glow” and
improved self-image as a byproduct of their own contribution to a public good (Andreoni 1990).
Reciprocity has been explained as another source of motivation for cooperation (Fehr and
Gachter, 2000). Experimental evidence from behavioral economics shows that people are strong
reciprocators who tend to cooperate voluntarily, responding to cooperative behavior by
maintaining or increasing cooperation and are willing to punish those who do not cooperate,
even at some personal cost (Gintis 2000, Gachter and Fehr 1999).

Similarly, social norms are seen as strong motivators for cooperative behavior. Social norms are
informal rules of behavior and are enforced by social sanctions (Coleman 1987) that can take the
form of approval or disapproval from others, and related feelings of pride and shame (Lindbeck
1997). Gachter and Fehr (1999) provide experimental evidence that the desire for social approval
gives rise to a large and significant reduction in free-riding among groups that share social ties.

Effects of monetary incentives on individual effort and pro-social behavior
A positive response to a monetary incentive is fundamental to economic theory and easy to
envision in daily life, where people work in exchange for payment and search for bargains while
shopping. On the other hand, in situations where intrinsic motivation exists to supply a good or
service, research shows that raising monetary incentives may reduce rather than increase the

106

supply of the good or service, as extrinsic motivation displaces or “crowds out” intrinsic
motivation (Frey and Jegen 2001, Mellstrom and Johannesson 2008, Deci 1971). In a metaanalysis of 128 related studies, Deci et al. (1999) found broad support for the argument that
extrinsic incentives undermine intrinsic motivation and as a consequence of rewards, people take
less responsibility of motivating themselves. Gneezy and Rustichini (2000a) found in
experimental work that subjects who were paid performed better the more they were paid, but
that on the whole, those who were paid performed poorly compared to those who were not paid.

Solow (1971) and Arrow (1972) assumed that monetary incentives were complementary to
altruistic incentives, so that offering payment would expand behavior that was otherwise driven
by the desire to help others. On the other hand, a range of recent literature shows monetary and
altruistic motivations can compete with each other – that introducing payment can actually
reduce pro-social behavior (e.g. Frey and Oberholzer-Gee 1997, Gneezy & Rustichini 2000b,
Frey and Goette 1999). Bowles (1998) reviews literature suggesting that socially embedded
other-regarding norms related to performance of and contribution to collective tasks decline as a
consequence of integration of traditional societies with markets. Heyman and Ariely (2004)
found that people are sensitive to price when they operate in monetary market situations but not
when they are in social situations, where altruistic motives and social norms guide behavior.
They found that when monetary market signals are imposed on social situations, people begin to
behave in a way that more resembles monetary markets.

107

Gaps in the literature and contributions of this study
The above discussion shows conflicting paradigms regarding the likely effects of monetary
incentives on collective action to undertake conservation. Individual rationality and in turn PES
implies that external payments will have a positive effect on conservation. Conversely, studies of
pro-social behavior driven by social norms suggest that payments could crowd out intrinsic
motivation to contribute towards a public good, thus possibly hampering the performance of
PES. Resolving this dichotomy has strong theoretical and practical implications for designing
PES projects in settings where contracts are undertaken with groups.

Evidence of crowding-out of pro-social behavior emerges mainly from experimental work done
with student subjects in laboratory settings in western or developed country contexts that are far
removed from natural resource management. While these insights remain useful, much remains
unknown about their applicability to natural resource management settings in developing
countries. Further, in the latter domain, evidence for crowding out comes mostly about the
effects of sanctions or regulations (Cardenas et al. 2000, Vollan 2008) rather than on the effects
of rewards.

This research addresses these gaps and explores the relationship between economic incentives
and pro-social behavior in the context of collective resource management in rural Tanzania
through a choice experiment and a field experiment that examine people’s willingness to
participate in communal tasks for different types and levels of reward. Building on evidence
described above that people respond differently to “social market” and “money market”
situations (Heyman and Ariely 2004), we assess the extent to which payment affects people’s

108

contribution to collective tasks. Further, we assess people’s preferences for different types of
payment or reward, whether individual or group payment and whether cash or kind. Following
Gneezy and Rustichini (2000a), we measure the possible differences between the effects of small
payments from those of high payments for the same task. Finally, following Gachter and Fehr
(1999), we assess the strength of social approval incentives among community members with
close social ties and the extent to which they curb free riding.

4.3. Study context
The experimental work in this study was conducted in six villages in the Uluguru mountains in
Tanzania. The six villages were chosen because they participate in decentralized forest
management arrangements and they have a school. Three of the villages had been purposively
selected for qualitative research conducted several months prior to the research reported here; the
additional three were selected for this research because they met the selection criteria and were
located near the other three villages. The Uluguru mountains form a part of the Eastern Arc
Mountain range, an important center of floral and faunal diversity and endemism (Polhill 1968).
The Ulugurus are home to at least 16 endemic vertebrate and 135 endemic plant taxa. This
degree of endemism in the Ulugurus is exceptional in tropical Africa, putting these mountains
among the continent’s ten most important conservation sites (Burgess et al. 2002). The Ulugurus
also form the most important water catchment in Tanzania as the source of the river Ruvu, which
supplies water to Dar-es-Salaam, the country’s largest city. However, recent surveys report a
decline in forest area and potential disappearance of a few endemic faunal species, with
conversion of forest to agriculture as the major reason for forest loss (Burgess et al. 2002).

109

Some endogenous forest conservation efforts exist, for example in the form of voluntary forest
patrols. Voluntary behavior is also evident in some other tasks that community members
undertake together such as planting trees to mark forest boundaries or contributing labor for
construction and maintenance of village schools. In our field survey, 88 percent of respondents
had participated in some collective task within the last three years, indicating the presence of
strong social ties in the area. People mentioned having contributing labor, materials and/or cash
to build a village school, dispensary, and village government office.

There is great interest in exploring alternate conservation approaches to address forest
degradation in the Ulugurus. The Tanzanian government and NGOs have been exploring PES as
an instrument to promote landscape conservation in these mountains. Several pilot PES projects
are being implemented, including projects that pay people to conserve existing forests and plant
new forests to provide carbon sequestration and hydrologic services (Katoomba Group 2007).
Discussions are ongoing around the design of appropriate incentive structures under PES that are
applicable to collective action settings for forest management under REDD.

4.4. Methods
This study used a survey and experimental methods including a survey-based choice experiment
and a field experiment to elicit people’s preferences and test their responses to different reward
types and levels. The fieldwork began with focus groups exploring traditions of collective action
around natural resources and other issues with village elders, schoolteachers and officials. These
initial discussions laid the groundwork for a household survey that incorporated a choice

110

experiment about participation in communal tasks and guided the selection of the task in the field
experiment.

Field experiments are increasingly used to understand decision-making in real world settings.
Studies of populations of interest in field experiments potentially generate more relevant findings
than do studies of a convenience sample of students in laboratory experiments (Harrison and List
2004). Based on the typology of field experiments proposed by Harrison and List (2004), our
experiments correspond to the category of a framed field experiment using a nonstandard subject
pool and a field context. The task structure of the experiment was designed to embed subjects in
the experiment and it resembled closely the collective action situations they face in their daily
lives.

Household survey to elicit preferences for group versus individual rewards
In a questionnaire survey respondents were asked what type of reward – individual cash, group
cash, or group in-kind rewards – they would prefer in a hypothetical PES-type arrangement in
which they would patrol the village forest to protect against fire and illegal logging. This
question was asked of 194 respondents randomly selected from a village level listing of
households across each of the four villages. The following three hypothetical payment options
were offered to the survey respondents:
1. An individual payment of Tshs 2,000 (equivalent to wage for one day of casual work in the
village, about US $1.50) for participating for one full day in a forest patrol.
2. No individual payment, but the Tshs 2,000 would be given to the village forest committee to
be used for village forests on behalf of the respondent.

111

3. No individual payment, but investment in village infrastructure (roads, school) would be
made on behalf of each participant.

Stated choice experiment: stated pro-social behavior under various reward types and levels
The household survey was also used to launch a choice experiment asking respondents to state
their preference for different types and levels of cash and non-cash rewards. The choice
experiment asked each respondent whether or not people in their village would participate in a
collective task in return for a specific reward that was randomly assigned to the respondent from
a list of five different options.

The hypothetical questions in the survey-based choice experiment made use of an inferred
valuation method to elicit subjects’ preference. This aims to counter the effects of a social
desirability bias, which implies that in both real and hypothetical treatments to estimate people’s
values for public goods, respondents provide answers that they think will please the interviewer
or be consistent with societal norms (Fisher 1993, Lusk and Norwood 2009). In general, people
have the propensity to misrepresent their “true” preferences out of concern for how they are
viewed by others. Inferred valuation entails asking a person how much another person would be
willing to pay for a good. In our experimental work we asked whether the respondent’s neighbor
would be willing to engage in a pro-social behavior under various levels and types of rewards.

The hypothetical collective task involved a half day of work to slash grasses in the village school
yard under different levels of cash and non-cash rewards. We designed five treatments
representing different rewards:

112

1. Payment neither mentioned nor offered.
2. Mentioned that no payment would be made.
3. Low individual cash payment (Tshs 200).
4. High individual cash payment (Tshs 1000).
5. No individual payment, but group reward of Tshs 1000 per participant paid to the village
school.

Each of the 194 survey respondent was randomly assigned to one of the five treatments, with all
treatments conducted in each village. We found no differences in mean age, gender and number
of years in school of the subjects assigned to the five treatments.

Field experiment: revealed pro-social behavior under various reward types and levels
A field experiment elicited people’s actual decisions when invited to contribute to a communal
task under different incentives. The experiment involved asking people to participate in a treeplanting activity in the schoolyard in four villages. The four villages included two where the
survey and choice experiment were conducted, and two others. Individuals in each group were
each asked to plant at least 4-5 plants. This included carrying the plants from a central location to
the planting site, clearing the planting area of grasses and weeds, digging pits one cubic foot,
planting the trees and refilling the pits, fetching water from the nearest water source and watering
the plants. The activity would last about four hours, a half day of wage work in the area.

113

We chose tree planting as the experimental activity because it simulates a scenario that is real
enough for people to think about contributing to a communal task, but not something that people
normally do. So while this activity is good for testing the hypotheses, we were not concerned
that it would interfere with people’s motivation to participate in communal tasks that they
currently perform without pay.

Logistical challenges in Tanzania made it impossible to randomly select participants and we
stress that the interesting findings from the field experiments are not the participation rates under
each treatment but the variation in exit interview responses by treatment. In most of rural
Tanzania, people live in disbursed homesteads rather than clustered together in villages. For
many people, going to the village schoolyard to plant trees involves walking a long distance and
they would need to be notified earlier than just the previous evening. Notifying people earlier,
however, would increase the likelihood that participants are not randomly selected and only
those participants who are known to the village leaders who were used to recruit participants get
invited. With help from village officials, we recruited 124 people across the four villages to
participate in the activity. Each treatment was administered in each of the four villages, making a
total of 12 groups. Comparison of mean age, gender, and number of years of school of subjects
showed that there were no significant differences across treatment groups, nor were there any
strong differences between groups in the field experiment and the groups in the choice
experiment described above.

In the experiment we asked people to perform the activity under the following three randomly
assigned treatments:

114

1. Participants would perform the activity on a voluntary basis, with compensation neither
mentioned nor offered.
2. Participants would receive an individual cash payment of Tshs 1000 for successfully
completing the task.
3. For each participant a payment of Tshs 1000 would be donated to the village school.

Once the subjects arrived at the school, they were randomly allocated to the treatments. The
subjects in each treatment were seated in three different classrooms and could not interact with
subjects in the other groups. A monitor accompanied each group to the planting site. Three
separate experimental scripts were developed that detailed a common task but differed in the
nature of reward offered. The script was read to each group in Kiswahili, the local language, and
after a five-minute wait during which subjects could decide to participate or drop out, all
participants were escorted by the monitor to a planting area in the school yard, which was
sufficiently far from the planting areas assigned to the other two groups. The monitor stayed with
his assigned group at all times and made sure that there was no inter-group communication. After
the subjects declared that their task was over, the monitor escorted the group back to their
classroom for individual exit interviews. Each exit interview was conducted privately and
consisted of assessing the subject’s satisfaction or dissatisfaction with the task and the incentive
offered. After the interview, a show-up amount of Tshs 1000 was paid to each subject in every
group. The subjects did not know of the show-up payment in advance; it aimed to achieved
parity in the actual reward for subjects across all groups and thus avoid any conflicts.

115

4.5. Results and discussion
Preferences for group versus individual rewards
Responses to the survey question regarding preference for group versus individual rewards
reveal that across all villages, respondents have a high preference for an investment in social
infrastructure followed by individual cash payments and lastly payments to be shared among the
group (Table 4.1). This suggests a high willingness among respondents to contribute to activities
that result in collective rather than private benefits.

A high preference for an investment in social infrastructure as a reward for undertaking
conservation is not surprising because in rural Tanzania it is common to pool village resources to
help finance investments in village infrastructure. The particularly strong preference for public
infrastructure as a reward in Kibangile and Ngongolo may be because these two villages are
remotely located and lacking in infrastructure compared to the other two villages. The low
preference for group-cash rewards to village forest protection committees probably reflects the
low trust the people have in the abilities of these institutions. Focus group discussions in the
study villages reveal that there are issues of local corruption and misuse of monies in at least
three of the four study villages.

Stated choice experiment: stated pro-social behavior under various types and levels of
reward
Table 4.2 presents the results of the stated choice experiment. For all treatments the majority of
respondents indicated that their neighbor would participate in the communal work to maintain
the village schoolyard. The response was highest at 97% in treatment 4 (high individual cash

116

payment equivalent to the prorated daily wage), and lowest in treatment 3 (low individual cash
payment of about 20% of the prorated daily wage) at 64%. The positive response rate for
treatment 1, with payment neither mentioned nor offered, is 82%. Treatment 1 is akin to what
Heyman and Ariely (2004) call a social market situation and the subject’s statement that his
neighbor would be willing to participate can be interpreted as being driven by either altruistic
motives or norms toward participation in collective tasks of social importance.

We used the non-parametric Mann-Whitney paired comparison test (Ott and Longnecker 2001)
to compare the fraction of subjects across treatments agreeing to participate in the communal
task. Two findings stand out. One is that the positive response rate with a high individual
payment is statistically different from that in all other treatments, as shown in table 4.2. This
indicates that participants in this experiment respond to monetary incentives, consistent with the
rational utility maximization paradigm guiding PES. The other is that the positive response rate
for a low payment is statistically significantly lower than when payment is neither mentioned nor
offered. This finding is consistent with that of Gneezy and Rustichini (2000a) that low payments
reduce effort relative to no payment, and with that of Heyman and Ariely (2004) that the
introduction of payment changes people’s mindset from operating in a “social market” to a
“monetary market.”

In treatment 5, where a social cash payment was offered in the form of a payment to the school,
76% of the subjects said that their neighbor would be willing to participate in the collective task.
The response rates with the social reward are not statistically different from response rates in
treatments 1 (82%) and 2 (74%), in which no payment was offered. This indicates that the

117

charity option has neither the positive effect of the high individual payment nor the negative
effect of the low individual payment.

Field experiment: revealed pro-social behavior under various incentive types and levels
The results from the field experiment exhibit no variation in contribution levels across the three
treatments. All 124 subjects decided to participate in the activity regardless of the absence or
presence of an individual cash incentive or a social reward (Table 4.3). First, as mentioned
above, there is reason to suspect that village officials may have recruited people whom they
might expect to participate. Second, the long distance to school implies a huge sunk cost just to
show up, in which case a subject might as well participate in the activity regardless of the
incentive. Third, participation is consistent with the existence of a strong social norm for
participation in communal tasks. And fourth, 100% participation rates could result from peer
pressure or what Gachter and Fehr (1999) call an “expectation of social approval.” The field
experiment was set up to measure individuals’ revealed behavior in a public situation. Therefore,
not only did subjects know each other very well but it was possible for everyone to observe
everyone else’s contribution. This public setting likely triggered the subjects’ desire to gain
social approval and to be seen doing the right thing by people they know, and it might have
motivated them to participate in the activity irrespective of the level of extrinsic rewards offered.

As mentioned above, the interesting findings from the field experiment come from the exit
interviews, which revealed large differences in subjects’ satisfaction levels across the three
groups despite universally high participation in communal task. Table 4.4 shows that despite
universal participation, subjects had very different opinions about their experience across

118

treatments. Subjects in treatment 1 (no individual payment) and treatment 3 (no individual cash
payment but a donation to village school) had a high degree of satisfaction with the task
performed and the payment offered. 90% of the subjects in treatment 1 and 86% of the subjects
in treatment 3 stated that they were “happy” or “very happy” about the task. Many subjects in
both treatments 1 and 3 also remarked that given a chance, they would gladly do this kind of
activity again. However, only 43% of the subjects in treatment 2 (individual payment) stated that
they felt happy, with the remainder saying that they were not satisfied or unhappy with the
payment offered and grumbled that it was low for the amount of work they did (even though it
was commensurate with the local market wage for casual work). These findings support the
motivation crowding-out effect of monetary rewards as suggested by Deci (1971) and others.
This result is similar to that in the choice experiment where the low payment offer yielded a
significantly lower expression of willingness to participate than the no-payment offer.

4.6. Conclusion: implications for PES
The literature on PES makes a strong case for compensating farmers to conserve fragile
ecosystems where they bear opportunity costs from doing so (e.g. Pagiola and Platais 2002,
Wunder 2005). Where conservation comes at the expense of livelihood, it is understandable if
livelihood receives priority, and the logic of raising the value of conservation to be economically
competitive is sound. However, this paper has shown that the equation is more complicated due
to the potentially ambiguous effects of monetary incentives on the kind of pro-social, collective
behavior that often will be required in group-based PES. Further, it has shown that cash
payments may not be the preferred incentive for motivating conservation behavior. These
findings suggest the need for great care in designing incentives in PES.

119

As discussed above, despite environmental degradation in the study area, there is at least some
evidence of social norms for environmental protection, in particular the voluntary forest patrols
in which people participate. Thus in such a scenario, where people undertake collective action
for environmental protection, it is important not to design incentive-based conservation programs
that may undermine existing pro-conservation motivations people have for engaging in this
work. Moreover, when funding for PES projects runs out (as would be the case for many
publicly financed programs), there is a fear that if people become accustomed to payment they
might stop conservation efforts altogether or demand to be paid for activities that they already
undertake voluntarily. Conservation planners in Tanzania already express concern about this
issues (Pers. Communication, Director, Eastern Arc Mountain Conservation Endowment Fund).

Interestingly, however, our results also suggest that payments will not necessarily have adverse
impact on pro-social behavior in the presence of pre-existing norms favoring collective action
among group members. In the choice experiment, the positive response was highest when a high
payment was offered. In the field experiment, even though subjects grumbled about the payment
they had received for planting trees in the school yard, they still went ahead and contributed the
same level of effort as those who were not offered any payments. Going back to our original
quest, these results suggest that the two dominant paradigms of individual rationality and social
utility may both need some fine tuning to capture the complexity in human behavior. Payments
may undermine the satisfaction that individuals derive from pro-social behavior, but in the
presence of strong norms, they may just become embedded in the social approval incentives that

120

exist among group members. If as in this case there is a social norm in favor of participation in
communal tasks, the desire for social approval may tip the scale in that direction.

Another important result is that payments can have non-monotonic effects on cooperation
behavior. A trivial payment can significantly reduce willingness to cooperate, even below the
level when the payments are mentioned but not provided (table 4.2). Even though PES theory
stresses the need for payments to cover opportunity costs, in practice payments in some PES
schemes may be lower than opportunity costs. Kosoy et al. (2007) document such cases but
suggest that with the right design they can contribute to other counterbalancing measures, for
example designing improved institutional arrangements for cooperation. Our research supports
the importance of arrangements that strengthen social norms of cooperation, but suggests risks in
tying them to low monetary payments.

Finally, the survey yielded a strong preference for group-level, in-kind rewards for
environmental services but a very low preference for group-level cash payments, and the choice
experiment revealed a statistically significantly lower preference for a cash payment to the
school than for individual cash payments of the same value. The low preference for cash
payments made to local institutions such as forest protection committees or the school reflect
respondents’ lack of confidence in these institutions and concerns about misuse of funds. On the
other hand, there appears to be a stronger preference for group-based in-kind rewards. Heyman
and Ariely (2004) found that responses to in-kind payments were not susceptible to the same
potentially perverse outcomes as cash payments, whereby a low payment could yield a reduced
level of effort compared to no payment. Similarly, Mellstrom and Johannesson (2008) found

121

that alternative reward types could counteract the tendency of cash incentives to crowd out
intrinsic motivation. Kerr and Vardhan (2010) argue that in-kind payments can have favorable
properties for promoting collective action, and they discuss ways to also make them consistent
with conditionality. All of this suggests potential benefits in using in-kind rewards in groupbased PES, especially since individual cash payments are not likely to be feasible in group-based
PES due to high transaction costs.

The importance of social norms and social approval incentives in driving participation in
communal tasks suggests value in designing conservation programs in ways that strengthen
social norms and intrinsic motivations. This is not to suggest a return to the days of Integrated
Conservation and Development Projects (ICDPs), which according to Brandon (1999) tended to
rely on social norms and local institutions that in fact often did not exist, and relied on incentive
structures that were not clearly linked to conservation (Ferraro 2001). The failure of ICDPs led to
the emergence of PES. In its emphasis on conditional incentives and the importance of not
relying on unfavorable social norms and institutional arrangements, PES appears to have
neglected the potential benefits from developing social norms and intrinsic motivations that
could work in favor of conservation. A logical next step is to work toward approaches that
incorporate all of these potentially favorable components while avoiding possible perverse
responses to monetary incentives. Recent literature has begun to address these issues (Reeson,
2008; van Noordwijk et al., 2010) and it is likely to remain an area ripe for testing new ideas.

122

Appendices

123

Table 4.1: Village-wise preferences for group versus individual rewards for hypothetical
†
work to patrol village forests (percentage of respondents)
Village
Gwata
Kibangile
Muhungamkola
Ngongolo
Total (all villages)
†
‡

Individual
cash
32.5
8.5
37.8
24.6
23.6

Group cash
15
15.3
16.2
1.8
11.2

Social
infrastructure
52.5
76.3
43.2
68.4
62.8

Other
suggestions‡
2.5
0
2.7
5.2
2.5

Source: authors’ survey, 2009, n=194
Other suggestions was an open-ended response category.

Table 4.2: Results of stated choice experiment about participation in communal work to
†
maintain the village school yard under various types and levels of reward

1.
2.
3.
4.
5.

Treatments
Payment neither mentioned nor offered; n=39
Mention that no payment will be made; n=39
Individual payment of Tshs 200; n=39
Individual payment of Tshs 1000; n=39
No individual payment but Tshs 1000 for each
participating individual awarded to the village school;
n=38

Willing to participate‡
32 (82%)
29 (74%)
25 (64%)
38 (97%)
29 (76%)
§

Hypothesis tests (p-value of difference)
Treatments 1 and 3
Treatments 4 and 1
Treatments 4 and 2
Treatments 4 and 3
Treatments 4 and 5
†
‡

0.076
0.026
0.004
0.000
0.006

Source: authors’ survey, 2009. Total n=194
Chi-square = 13.96, 4 df, statistically significant at 1%.

§

Comparisons between categories use the Mann-Whitney test. Only cases statistically
significant at 10% probability or less are displayed

124

Table 4.3: Results of field experiment about participation in communal work to plant trees
in village school yard under various types and levels of reward
Treatments
2. Individual cash
payment 1000 Tshs

1. No individual
payment offered
Number of subjects in
the treatment
Number and % of
subjects completing the
task

40

42

3. Social
payment to
school
42

40 (100%)

42 (100%)

42 (100%)

N=124, Source: authors’ field experiments, 2009.

Table 4.4: Summary of exit interview responses under various treatments in the field
†
experiment

Treatment
1. No individual payment
2. Individual cash payment
3. Social payment to the school

Response to task under different payments
Happy or very
Unhappy
happy
36 (90%)
4 (10%)
18 (43%)
24 (57%)
36 (86%)
6 (14%)

‡

§

Hypothesis tests (p-value of difference)
Treatments 1 and 2
Treatments 1 and 3
†
‡

.000
.000

Source: authors’ exit interview with field experiment subjects. Total N = 124
Chi-square = 28.16, 2 df (statistically significant at 1%).

§

Treatments 1 and 3 are both statistically different from treatment 2 at 1% using the MannWhitney test.

125

Table 4.5: Record form for participants of field experiments

Part I
1. Group no_____________

2. Form number________________

3. Name____________________________________________________
4. Gender (M/F)_____________

5. Age___________________

6. Education status
(i). No schooling ______(ii) Primary School_______ (iii). Secondary
school____

Part II
1. How many trees did you plant? ____________
2. Compared to others in the group, how would you judge your
performance?
(i). I worked harder than others__________
(ii). I worked just as hard as others ______
(iii). I worked less hard than others ______
3. Were you satisfied with the payment offered in your group?
(i). Completely satisfied ______________
(ii). Just satisfied ____________________
(iii). Not satisfied ____________________
4. Why did you complete or not complete the activity?
____________________________________________________________

126

Table 4.6: Script for the experiment

Habari zenu!
Thank you all for coming today.
We have done a research in this village about environment protection. In this research, the
people of this village have shown interest to plant trees to improve the village
environment. People of this village have identified tree planting in the school-yard as an
important activity.
You have been invited to plant trees in the school-yard. The tree planting activity will last
for one hour only. We have three trees for each person. If you decide to plant trees, each
one of you, can plant these three trees in one hour. Your decision to plant trees is
completely voluntary. You can decide to leave the activity at any point in time and return
to your home.
If you decide to participate in the tree planting activity today,
1. None of you will receive any payment.
2. You will each receive TSHs 1000 for one hour of work.
3. There will be no individual payments for any of you. Instead, TSHs 1000 for

every person who participates in the activity will be contributed to the school for
maintaining the classrooms.

127

References

128

REFERENCES

Agrawal, A. 2007. Forests, Governance, and Sustainability: Common Property Theory and its
Contributions. International Journal of the Commons 1(1). pp. 111-136.
Andreoni, J. 1990. Impure Altruism and Donations to Public Goods: A theory of Warm-Glow of
Giving. The Economic Journal, Vol. 100. No. 401. pp. 464-477.
Arrow, K. 1972. ‘Gifts and exchanges.’ Philosophy and Public Affairs 1(2): 343-62.
Baland, J-M. and J-P. Platteau. 1996. Halting Degradation of Natural Resources: Is There a
Role for Rural Communities? Oxford: FAO and Clarendon Press.
Blomley, T. and H. Ramadhani. 2007. Participatory Forest Management in Tanzania: An
overview of status, progress and challenges ahead, The Arc Journal Tanzania Forest
Conservation Group. Issue 21, pp. 3-5.
Bowles, S. 1998. Endogenous Preferences: The Cultural consequences of Markets and other
Economic Institutions, Journal of Economic Literature, Vol. 36, No. 1. pp. 75-111.
Brandon, K. 1999. Moving beyond integrated conservation and development projects (ICDPs) to
achieve biodiversity conservation. In D.R. Lee & C.B. Barrett, eds. Tradeoffs or
synergies? Agricultural intensification, economic development and the environment.
Wallingford, UK & New York, CABI Publishing.
Burgess, N., N. Doggart., J.C. Lovett. 2002. The Uluguru Mountains of eastern Tanzania: the
effect of forest loss on biodiversity. Oryx Vol 36, No. 2. pp. 140-152.
Cardenas, J.C., J. Stranlund and C. Willis. 2000. Local Environmental Control and Institutional
Crowding Out. World Development. Vol. 28, No. 10. pp. 1719-1733.
Cleaver, F. 2000. Moral Ecological Rationality, Institutions and the Management of Common
Property Resources, Development and Change, Vol. 31, Issue 1, pp. 361-383.
Coleman, J.S. 1987. Norms as social capital. In. Radnitzky, G., Bernholz P. (eds) Economic
Imperialism: The Economic Approach Applied Outside the Field of Economics. Paragon,
NewYork, pp. 133-155.
Deci, E.L. 1971. Effects of Externally Mediated Rewards On Intrinsic Motivation, Journal of
Personality and Social Psychology, Vol. 18, No. 1, pp. 105-115.
Deci, E.L., R. Koestner, and R.M. Ryan. 1999. A meta-analytic review of experiments
examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin,
125. pp. 627-668.
129

Fehr, E. and A. Falk. 2002. Psychological Foundations of Incentives, European Economic
Review, No. 46. Issues 4-5, pp. 687-724.
Fehr, E. and S. Gachter. 2000. Fairness and Retaliation: The Economics of Reciprocity. Journal
of Economic Perspectives, No. 14, pp. 159-181.
Ferraro, P.J. 2001. Global Habitat Protection: limitations of development interventions and a role
for conservation performance payments. Conservation Biology 15(4): 1-12.
Ferraro, P., and A. Kiss. 2002. Direct Payments for Biodiversity Conservation. Science 298:
1718-1719.
Fisher, R.J. 1993. Social desirability Bias and The Validity of Indirect Questioning. Journal of
Consumer Research. Vol. 20. pp. 303-315.
Frey, B.S. and R. Jegen. 2001. Motivation Crowding Theory, Journal of Economic Surveys, Vol.
15, No. 5., pp. 589-611.
Frey, B.S. and L. Goette. 1999. Does Pay Motivate Volunteers? Institute for Empirical Research
in Economics. University of Zurich. Working paper Series, ISSN 1424-0459.
Frey, B.S. and F. Oberholzer-Gee. 1997. The Cost of Price Incentives: An Empirical Analysis of
Motivation Crowding-Out, The American Economic Review, 87, 4, pp. 746-755.
Gachter, S. and E. Fehr. 1999. Collective action as a social exchange. Journal of Economic
Behavior & Organisation. Vol. 39.pp. 341-369.
Gintis, H. 2000. Beyond homo economicus: evidence from experimental economics, Ecological
Economics, Volume 35, Issue 3, pp. 311-322
Gneezy, U. and A. Rustichini. 2000a. Pay Enough or Don’t Pay at All. The Quarterly Journal of
Economics, Vol. 115. No. 3.pp. 791-810.
Gneezy, U. and A. Rustichini. 2000b. A fine is a price. Journal of Legal Studies. Vol. 29. No. 1.
pp.1-17.
Hardin, G. 1968. The tragedy of the commons. Science, 162, 1243-1248.
Harrison, G.W. and J.A. List. 2004. Field Experiments. Journal of Economic Literature. Vol. 42.
No. 4. pp. 1009-1055.
Heyman, A, and D. Ariely. 2004. Effort for Payment: A tale of Two Markets. American
Psychological Society. Vol. 15. No. 11.pp. 787-793.
Jodha, N.S. 1985. Population Growth and the Decline of Common Property Resources in
Rajasthan, India. Population and Development Review, Vol. 11, No. 2. pp. 247-264.

130

Katoomba Group. 2007. East and Southern Africa Katoomba Group Newsletter,
http://www.katoombagroup.org/~katoomba/documents/newsletters/Vol1_No3.html#anch
or2, accessed on 27 August 2010.
Kerr, J., and M. Vardhan. 2010. Designing appropriate incentives for group-based payment for
environmental services. In review, World Development.
Kosoy, N., M. Martinez-Tuna, R. Muradian, and J. Martinez-Alier. 2007. Payments for
environmental services in watersheds: Insights from a comparative study of three cases in
Central America. Ecological Economics, 61. pp. 446-455.
Lindbeck, A. 1997. Incentives and social norms in household behavior. American Economic
Review, Vol. 87, No. 2. pp. 370-377.
Lusk, J.L. and F.B. Norwood. 2009. An inferred Valuation Method. Land Economics. 85 (3): pp.
500-514.
Mellstrom, C. and M. Johanneson. 2008. Crowding Out in Blood Donation: Was Titmuss Right?
Journal of the European Economic Association, 6 (4), pp. 845-863.
Miles, L. and V. Kapos. 2008. Reducing Greenhouse Gas Emissions from Deforestation and
Forest Degradation: Global Land-Use Implications. Science 320(5882): pp. 1454-1455.
Olson, M. 1965. The logic Of collective action. Harvard University Press, Cambridge.
Ostrom, E. 2002. The drama of the commons. Washington, DC: National Academy Press.
Pagiola, S., and G. Platais. 2002. Payment for Environmental Services, Environment Strategy
notes, No.3. World Bank, Washington, D.C.
Polhill, R.M. 1968. Tanzania. In Conservation of Vegetation in Africa South of Sahara. Eds. I.
Hedberg and O. Hedberg. Acta Phytogeographica Suecica, 54. pp. 166-78.
Reeson, A. 2008. Institutions, Motivations And Public Goods: Theory, Evidence And
Implications For Environmental Policy. CSIRO Working Paper Series, 2008-01,
Solow, R. 1971. Blood and thunder. Yale Law Journal. 80(2): 170-83.
Swallow, B., and R. Meinzen-Dick. 2009. , Payment for Environmental Services: Interactions
with Property Rights and Collective Action. Chpt 12, pp 243-265 in Beckmann, V., and
M. Padmanabhan (eds.), Institutions and Sustainability, New York: Springer.
Swallow, B., Kallesoe, M., Iftikhar, U., Van Noordwijk, M., Bracer, C., Scherr, S., Raju, K. V.,
Poats, S., Duraiappah, A., Ochieng, B., Mallee, H. and Rumley, R. (2009) Compensation
and Rewards for Environmental Services in the Developing World: Framing PanTropical Analysis and Comparison. Ecology and Society 14(2): 26.
131

Van Noordwijk, M., and B. Leimona. 2010. Principles for fairness and efficiency in payment for
environmental services in Asia: payments, compensation, or coinvestment? Ecology and
Society, in press.
Vatn, A., 2008. Institutions and Rationality. In Batie, S., and N. Mercuro (eds.): Assessing the
Evolution and Impact of Alternative Institutional Structures. London, Routledge Press,
pp. 113-139.
Vollan, B. 2008. Socio-ecological explanations for crowding-out effects from economic field
experiments in southern Africa. Ecological Economics 67(4):560-573.
von Neumann, J., and O. Morganstern 1947. Theory of games and economic behavior. Princeton
University Press, Princeton, NJ.
Wunder, S., 2005. Payments for Environmental Services: Some Nuts and Bolts. CIFOR
Occasional Paper No. 42. Center for International Forestry Research, Bogor.
Wunder, S., S. Engel, and S. Pagiola. 2008. Taking Stock: A comparative analysis of payments
for environmental services programs in developed and developing countries. Ecological
Economics. Vol. 65, 834-852.

132