Institutional diversity and local forest governance

NOTE: This is a pre-edited version of a paper that is now published as:

Andersson, K. Benavides, J.P. León, R. 2014. Institutional Diversity and Local Forest

Governance. Environmental Science and Policy 36: 61-72

Institutional Diversity and Local Forest Governance

Krister Andersson (corresponding author)

Associate Professor

Department of Political Science

Institute of Behavioral Science

University of Colorado at Boulder, Boulder CO 80309-0333, USA

Universidad de los Lagos, Chile

[email protected]

Phone: +1 303 735-2317, Fax: +1 303 492-0978

Jean Paul Benavides

Centro de Estudio de la Realidad Económica y Social (CERES), Bolivia

Rosario León

CERES, Bolivia

2

Acknowledgements

The paper benefitted from constructive comments from Arun Agrawal, Eric Coleman, Diego

Pacheco and participants at the North American Meetings of the International Association of the

Study of the Commons in Tempe, AZ on October 1, 2010. Financial support from the National

Science Foundation (BCS-0527165 and SES-0648447) as well as the USAID’s SANREM-CRSP

program is gratefully acknowledged.

3

Abstract

Scholarship on common-pool resource governance suggests that collective outcomes vary with

the strength of the local arrangements for compliance monitoring. Following Elinor Ostrom’s

approach to question panaceas, we explore the possibility that there are multiple institutional

designs can help sustain forests. We test these arguments with data from a sample of 200 forest

user groups in Bolivia and find broad empirical support for our propositions. Local monitoring

can be an important predictor of forest governance performance, but focusing on monitoring alone

can be misleading. Sometimes other aspects of the local governance system, such as self-

organized rule making and sanctioning, are more important in explaining why some groups

govern their forests more effectively than others. We also find that the more governance

functions that communities decide to organize themselves; the more likely it is that local forests

are sustained.

Keywords: Bolivia, forestry, rural development, governance, self-organization, institutions

4

1. Introduction

Several scholarly papers and reports have found that local users can be good stewards of

common-pool resources (i.e. Ostrom 1990, Arnold 1990, Ascher 1995, Clugston and Rogers

1995). Much of the empirical literature on natural resource management argues that local users

hold important time- and place-specific knowledge that is necessary for the creation of successful

governance regimes (Gibson and Koontz 1998; Berkes 1989; Bromley 1992; McCay and

Acheson, 1987; McKean 1992; Ostrom, 1999; Hayek 1948). Recent empirical studies suggest that

local groups are often at least as effective forest managers as national governments (Hayes, 2008;

Somanthan et al, 2009; Nelson and Chomitz. 2009; Porter Bolland et al 2012).

These findings notwithstanding, research on community-based forest management

activities suggest that these have had rather mixed outcomes: some communities govern their

resources successfully while others do not do as well (Bray et al 2003; Nagendra et al, 2005).

What factors help explain such divergent outcomes? Here, we address this question by focusing

on the role played by a variety of local institutional arrangements that local forest user groups

employ to control access to and regulate use of forest resources.

Several previous studies have found that one particular type of local institutional

arrangements—community-organized monitoring activities—is a key factor in explaining local

variation in forest conditions (i.e. Ostrom and Nagendra 2006; Agrawal and Chhatre, 2006;

Chhatre and Agrawal, 2008; Coleman, 2009; Gibson et al 2005). A recurring finding in all of

these studies is that local forest communities that carry out rule-compliance monitoring are more

likely to have forests in good conditions.

5

Here, we argue that there may be other aspects of self-governance apart from monitoring

arrangements that are critical in understanding why some groups enjoy better forest conditions

than others. Our central idea is that it makes good theoretical sense to distinguish between at least

three aspects of self-organized governance activities: rule-making, monitoring, and sanctioning.

By incorporating these three complementary measures of local governance into our study of local

forest governance, we believe we are in a position to gain a more nuanced understanding of the

particular aspects of self-governance that are most important for resource outcomes. We argue

that groups that have organized themselves to carry out at least some of these governance

functions increase the odds of being able to sustain forests.

The argument that we develop has four parts. First, we discuss the important contributions

of local forest users to the governance of forest resources. We argue that local involvement in

forest governance is critical because local users not only possess unique knowledge and skills

about the social and natural systems in which they live, but they also have a direct stake in the

sustenance of the common-pool resource system. Second, we explore the possibility that

successful local governance of forests does not always require local user groups to self-organize

and carry out all three governance functions—the complete “package” of self-governance—but

there may be circumstances under which user groups can manage their forests successfully by

carrying out just one or two of these functions. The consequence of this possibility leads us to the

third part of our argument, which is that local forest governance is not just about monitoring but

there are a multiple institutional designs of local governance that user groups may employ.

Finally, the fourth part of our argument is that while there may be circumstances under which it is

sufficient for user groups to carry out only one or two governance functions in order to protect

their forest, we predict that, in general, the more involved a user group is in organizing multiple

6

forest governance activities, the more likely it is that their forest will thrive. To test these four

linked arguments, we use data from a nationally representative sample of 200 rural settlements in

Bolivia and employ statistical techniques to analyze the relationships between local institutional

arrangements and three different measures of local forest conditions.

The rest of the paper is structured as follows: In the next section, we review previous

empirical research in the area of local governance of CPRs. We then present two cases of local

forest governance in the Bolivian context and use these cases to illustrate our theoretical

arguments. Section 4 describes the context in which we test this argument empirically: the

Bolivian forestry sector. After providing a brief outline of the data collection methods, we use

statistical analysis to test the paper’s central argument. We end by interpreting the statistical

results and discussing the broader implications of our study.

2. Previous Research

Several scholars have identified a large number of micro-level or so-called ‘contextual’

variables that affect the effectiveness of local communities’ efforts to govern common pool

resources. Agrawal (2001) reviews three major studies on community governance of CPRs (Wade

1994, Baland & Platteau, 1996, Ostrom, 1990) and identify over 40 variables that are potentially

influential determinants of community self-governance of CPRs. The characteristics of local

institutional arrangements that appear to be the most important, judging from the cases examined

in these three major studies, include the rule-making process, the characteristics of the rules,

monitoring and enforcement of rule compliance, and the imposition of sanctions on rule violators.

Our review of the empirical literature on local governance of forests identified nine recent

studies that engage in systematic comparative analyses of the ways in which local institutional

arrangements affect forest governance performance (Coleman, 2009; Coleman and Steed, 2009;

7

Gibson, Williams and Ostrom, 2005; Ostrom and Nagendra, 2006; Agrawal and Chhatre 2006;

Chhatre and Agrawal, 2008; Ghate and Nagendra 2005; Banana and Gombya-Ssembajjwe 2000;

Webb and Shivakoti 2008). With some variation in magnitude of the effects, all of these studies

find that that the local arrangements for monitoring and enforcement have a positive impact on

local forest conditions. None of these studies, however, include in their empirical analyses other

functions of self-governance—such as self-organized rule-making or sanctioning systems. This

means that the analyses in these studies may have overestimated the effect of monitoring and

enforcement vis-a-vis other aspects of self-governance. This is a possibility that we set out to test

in this paper.

Gibson, Williams and Ostrom (2005) analyze empirical data from 178 user groups in eight

different countries. Controlling for the influence of social capital and resource dependency, they

find that local monitoring and enforcement activities are positively associated with the people’s

perceptions of forest conditions. Consistent with these results, Ostrom and Nagendra (2006) show

that local community monitoring and enforcement of rules for entry into and use of resource

systems is significantly associated with positive changes in both forest basal area and tree stem

densities (p. 19230).

Chhatre and Agrawal (2008) examine the relationship between local enforcement and the

conditions of forest commons. Analyzing a multi-country dataset with observations from 152

community user groups in nine different countries, their analysis finds that high levels of

community enforcement has a strong positive effect on the probability that a given community’s

forest condition is improving. They also find that enforcement plays a moderating role in that it

tempers the negative effect of some of the known drivers of forest degradation. In Agrawal and

Chhatre (2006), the same authors analyze a set of communities in the Indian Himalayas and find a

8

positive effect of monitoring and enforcement arrangements (as well as using cash fines) on local

perceptions of forest conditions. Finally, Coleman (2009) uses longitudinal data on rural

communities’ use of 46 forests and finds that groups that carry out monitoring and enforcement

activities are more likely to have forests in stable or improving conditions.

Ostrom and Nagendra (2006) acknowledge that community monitoring and enforcement

activities are no panacea for governing forest commons, and in the words of the authors:

"focusing on monitoring alone is not sufficient" (page: 19230). Nevertheless, the empirical

analyses in the reviewed studies do not control for the possibly independent effects of other self-

governance variables on forest outcomes because none of these studies include independent

variables for local rule making or sanctioning activities in the analysis, only monitoring. Here,

we argue that there are both empirical and theoretical reasons for why future empirical studies

would benefit from expanding the current focus on monitoring and consider other self-governance

activities that are conceivably of equal theoretical importance.

Our distinction between the three different functions of local governance is not entirely

new—previous theoretical work on common pool resource (CPR) governance (i.e. Ostrom, 1990,

Bromley et al, 1992) has pointed out that users who seek to self-organize CPR governance face

three sequential collective-action dilemmas: first users need to agree on a set of rules to regulate

use and control access to the CPR; second they need to organize the monitoring of rule

compliance, and third users need to apply sanctions on individuals who break the rules. But much

of the empirical work on the local governance of forests has been less sensitive to such theoretical

distinctions (but see Agrawal, 1994; van Laerhoven, 2010). More often than not, the existing

empirical studies in this area, lift out local monitoring and analyze its effect on outcomes without

considering how other functions of governance, such as rule making and sanctioning, might

9

contribute to the overall performance of the governance system. The risk of such a perspective is

that it can give the impression that local compliance monitoring is all that is needed for good

outcomes.

Considering only monitoring and its effects on forest conditions may also cause problems

for the empirical analysis because the observed effects of monitoring may be linked to variations

in other variables that are related to self-organization more broadly. If these links are not

accounted for in the explanatory models—that is, omitting these non-monitoring variables from

the model—the results may produce a biased coefficient that overestimates the effect of

monitoring vis-à-vis other governance variables. Moreover, focusing on a single governance

variable called “monitoring” may result in theoretical ambiguity regarding the influence of

different aspects of local governance on forest outcomes. In this paper, we seek to shed light on

the potential differential effects of a broader set of self-governance functions.

3. Theory and Hypotheses

In this section, we develop an empirically-grounded theory about how self-governance

affects forest resource outcomes. Based on qualitative evidence from our own fieldwork, we

propose that the variability in forest conditions is systematically linked to three aspects of self-

governance: rule making, monitoring and enforcement, and sanctioning. Because these three

actions may be governed by separate and independent institutional arrangements, we hypothesize

that each of the three functions has an independent effect on forest conditions.

3.1 Empirical grounding

We draw on our own observations of two forest-user communities in Bolivia to illustrate the

possibility that there are multiple ways in which self-governance affects forest conditions. These

two cases represent two extremes of local forest governance—one user group that has developed a

10

strong set of governance activities and another with much more limited collective action to

govern their forests.

3.1.1 The Case of Villa Linda, Beni

The community of Villa Linda1, located in Department of El Beni, represents a case of strong

self-governance. Here, the local user group has developed its own set of rules for forest use, and

have been managed to set up an informal taxation system that generates income from local

logging activities that take place in forests beyond the village boundaries. The community

controls about 600 hectares of forest under a community-property title and has developed its own

rules to control access to and regulate use of these forest resources. These rules do not always

coincide with the national legislation. For example, the community authorizes each family one

annual cut from the communal forest but first each family must justify the harvesting in terms of

the family’s needs. Once authorized to harvest and sell forest products, each family is obliged to

pay a 10-percent tax on the income to the community authorities. It is extremely rare that these

rules are broken by community members, so the users have not implemented any extensive

monitoring of internal rule compliance. They don’t actively monitor outsiders’ use of the forest

either, because they consider it would be too costly to do so. Instead, they have invented a clever

way of making sure that outsiders do not steal resources from their forest without paying proper

compensation to the local community. So rather than patrolling the boundaries of the forest, they

have set up a toll both at a strategic location on the main road, where any truck carrying logs must

pass. They charge a fee to all vehicles carrying timber products, regardless of their origin.

Although this scheme is not actively supported by governmental authorities, it has been tolerated

1 We use a fictitious name to protect the identity of community members.

11

as a local, method of restraining timber harvesting in the area. With time, timber harvesters have

recognized the legitimacy of these rules.

The community is also implementing a comprehensive system of sanctions, which is not

only applied for infractions to the rules about timber harvesting but to all rules regarding land and

forest use. These sanctions are gradual in the sense that when a minor rule is broken for the first

time, the lightest degree of punishment is imposed and more severe types of sanctions are

reserved for repeat offenders of major rules.

3.1.2 The Case of Santa Barbara, El Chapare, Cochabamba

In the indigenous community of Santa Barbara2, located along the Chapare River in the

Department of Cochabamba, all of the community’s vast forest land is owned communally. The

community has a long history of interacting with and depending on its forest. Traditionally, the

community’s use of the forest was regulated by an elaborate, home-made rule system that defined

who could do what, when, how, and in which part of the forest. With time, however, these

institutions have lost significance and today they are no longer effective in constraining behavior.

Many of the community members, especially the elders are frustrated over this. One of them told

us that “the younger folks are not interested in living off the forest, they are more interested in

making money”. While the community has its own sanctions, these are rarely applied unless the

rule violations are related to individuals extracting high-value timber. Monitoring activities are

undertaken in a mostly hap-hazard and sporadic fashion: monitoring activities are largely reduced

to observations that community members happen to make while spending time in the forest.

The fading interest in the governance of the forest is starting to take its toll on the

community’s forest. While parts of the extensive forestland remain in stable condition, there are

2 This is not the real name of the community as we promised the local members that we would not reveal their

identities in our publications.

12

signs of increased degradation and reduction in parts of the forest. According to community

members we talked to, this degradation is due to unregulated timber harvesting and land clearing

by community members as well as outsiders—outsiders who have se settled at the perimeters of

the community’s territory. The community currently receives no support from external

organizations, but community members told us that they used to get support from two external

projects—one focusing on the production of cacao and the other on ornamental flowers—but

these did not have the desired results and were both discontinued in 2009.

This case represents “weak self-governance” as the local user group is losing its traditions

of self-organized forest governance, and although community members know about the

community’s traditional rule system, it is losing its relevance as community members struggle to

make all community members respect the rules, and have been unable to mobilize the users to

invest in effective monitoring and sanctioning activities.

3.2 Theoretical Argument

From these qualitative observations, we draw four lessons about local forest governance that are

central to our theoretical argument: (1) Local involvement in forest governance is critical; (2)

Local forest governance is not just about monitoring; (3) Multiple institutional designs can help

sustain forests, and (4) The more self-governance activities the better. Next, we develop the

theoretical rationale behind each of these propositions.

(1) Local involvement in forest governance is critical: we propose that local forest user

groups have a pivotal role to play in forest governance because they not only possess unique

knowledge and skills about the social and natural systems in which they live, but they often also

13

have a direct stake in the sustenance of the common-pool resource system. Without their active

involvement in governance activities, the odds for preserving forests are greatly diminished.

While we recognize that local forest users represent only one of the many actors who may be

involved in forest governance, and that sometimes other actors’ performance is very important,

we contend that local forests are extremely difficult to protect if the local user groups are not

committed to such protection, as we saw in the case of Santa Barbara. .

The case of Santa Barbara shows how outside pressures on the resource are more difficult

to withstand when the local governance system is broken, and in this case it was not just the

monitoring arrangements that were broken but the rules themselves that had been progressively

weakened.

(2) Local forest governance is not just about monitoring: Villa Linda’s experience

shows that despite the lack of systematic monitoring activities, harvesting activities are kept in

check by their home-grown rules system, graduated sanctions, and their innovative taxation

system. This suggests that it may not always be the community’s monitoring activities that make

the biggest difference for the forest. Depending on the circumstances, sometimes it could be the

community’s rule-making, sometimes the sanctioning, and, yes, sometimes it could be the

community monitoring that has the biggest influence on outcomes. By recognizing these distinct

possibilities, we believe scholars are in a better position to gain a deeper understanding of how

local institutional arrangements influence governance outcomes.

(3) Multiple institutional designs can help sustain forests: We propose that better conditions of

local forests are associated with a diversity of possible combinations of three distinct community

governance functions. The exact configuration of these self-governed institutions, or what we call

local institutional design, will vary from one community to another. In some cases, the

14

community faces circumstances that make the development of its own governance arrangements

very difficult—such as the size and heterogeneity of the group, the community members’

economic interests, and the level of trust, among other factors (Ostrom, 1990; Baland and

Platteau,1992; Agrawal, 2001). Given the specific circumstances, communities may not succeed

in organizing all three governance activities, but this, we argue, does not necessarily mean that

their forest is in danger: Even if they are able to organize just one or two governance functions

the odds of sustaining their forest will improve.

(4) The more self-governance activities the better: Rather than considering local

monitoring as the one and only function that matters for local forest conditions, the cases suggest

that achieving effective local forest protection is not so much about monitoring per se—or any

other particular governance function for that matter—but rather about the degree to which local

users are committed to and involved in forest governance activities. While local forest monitoring

can serve as a reasonable proxy indicator of forest governance performance, we propose that a

better indicator is the degree of the user group’s active engagement in the entire governance

process related to forests. We argue that having more governance functions organized by the

community, the better it is for conservation outcomes. In other words, having just one function

can sometimes make a difference, but having more functions organized by the community will, in

general, increase the likelihood that forest conditions are in good shape. In the next section, we

introduce the empirical context in which we seek to test this proposition: a large sample of forest-

dependent, rural villages in Bolivia.

15

4. Community Forestry in Bolivia

Rural communities in Bolivia represent an excellent context in which to examine the effects of

self-organized institutional arrangements on forest outcomes. First of all, the degree of scarcity

and salience of forest resources varies a great deal from one community to another (Andersson

and Pacheco, 2004). Places in which resource users perceive the resources to be salient and

increasingly scarce are more likely to have organized and implemented local institutional

arrangements for governing the resources (Gibson and Becker, 2000), which is the group of

independent variables that we are most interested in for this study.

Moreover, the forestry sector constitutes an important contribution in the current's

government's efforts to reduce rural poverty in Bolivia and national legislation has evolved to

become increasingly supportive of self-governance of forests (Pacheco et al, 2008; Larson et al.

2008). Since the mid-1990s, Bolivia has made efforts to create an enabling policy environment

for community forestry. The 1996 forestry and agrarian reform laws recognized smallholders and

indigenous groups as legitimate forest users, and decentralized many of the previously centralized

governance responsibilities to regional and municipal governments. United Nations agencies,

among others, have praised Bolivia for their political will to decentralize and modernize the

governance structure of the country’s forestry sector (FAO, 1999; UNDP, 1998).

The forestry and agrarian laws allow local communities and individuals to acquire formal

rights to manage forests, either as individual or as common property. The implementation of

these new sets of property rights, however, has been spotty and slow. Many rural communities are

still unclear about their formal rights to both land and resources. Community titling is far from

complete. As Larson et al note: "To date, state agencies, with their small regional staffs and

budgets, have manifested little capacity to defend the communally titled lands" (Larson et al,

16

2008: 112). Under these circumstances, the locally developed institutional arrangements are

likely to be particularly important for forest governance outcomes. Little systematic evidence

exists, however, on the environmental effects of local governance efforts.

5. Data and Methods

To test our hypothesis about the positive effects of self-governance on forest conditions, we

collected and analyzed data on self-organized use and management of forest resources in Bolivia.

In total, we visited 200 rural communities in the country. To reduce costs and logistical

complexity of data collection, we employed a clustered, two-stage sampling design. In the first

stage, we randomly selected 100 municipal territories. For each municipal territory, we

documented the distance between each rural village and all major roads. To ensure that sampled

communities would be accessible by field teams in a reasonable amount of travel time, we

systematically eliminated all villages from our sample list whose distance to a major road

exceeded 50 km. From the list of remaining villages, we randomly selected two villages in each

of the 100 municipalities.

In each community, our field teams collected data on forest characteristics, demographics,

decision making and governance activities related to forest use. The field teams, which consisted

of research assistants associated with the Centro de Estudios de la Realidad Económica y Social

(CERES) in Cochabamba, conducted a one-day workshop in which a series of questions were

discussed in a plenary focus group. The number of community members who attended these

workshops ranged from four to 65 individuals, and most workshops lasted for 1.5 to 2.5 hours.

During the workshops, we asked attendees a total of 15 questions. Their responses to these

questions allowed us to code 45 variables, many of which we use in the analysis for this paper. In

addition to the focus-group discussion data, we use the national census data for some of the

17

demographic and infrastructure control variables in the statistical tests. All of the data used in the

empirical tests are described in some detail in the next section. To assess the extent to which our

three measures of community self-organization has any systematic influence on forest conditions,

we employ Ordinary Least Square and Logistic regression techniques.

5.1 Dependent Variables

As outcomes, we employ one continuous and two binary variables that are proxy measures for

local forest conditions:

Forest Cover: This continuous variable indicates local group members' assessment of the

proportion of the community's land area that is covered by forest. The higher this number is the

higher the forest cover.

Forest Biodiversity: This is a binary variable that measures whether community members

perceive that the forest that they use is more biologically diverse relative to surrounding forests. A

value of 1 represents higher biodiversity, and a value of 0 means lower diversity than nearby

forests.

Forest Vegetation Density: This is a binary variable that indicates whether community members

perceive that the forest resources that they use has a higher density of trees compared to

surrounding forests in the area. A value of 1 represents higher density, and a value of 0 means

lower density than nearby forests.

The variation for all three dependent variables is presented in Figure 1, showing that local

residents in about 20 percent of all rural communities in Bolivia perceive that they have forests in

relatively good condition. We now turn to a description of the factors—and in particular the

multiple aspects of self-governance—that can help explain this variation.

18

[Figure 1 here]

5.2 Independent Variables

Given our theoretical focus and the findings from previous research, we specify an econometric

model that includes eleven independent variables, of which four are related to self-governance

activities and are therefore of particular theoretical interest here. The other seven variables are

included as control variables as they represent plausible determinants of local forest condition

variability. What follows is a brief description of these variables, including our theoretical

predictions with regards to their effects on local forest conditions in Bolivia.

Self-organized rule systems: This variable measures the existence of self-made rules concerning

(i) the type of forest use that is permitted; (ii) the quantity any local resident may harvest, (iii)

where harvesting may take place, and (iv) who may or may not extract certain products. When we

asked questions about rules, we defined rules as “any agreement between forest users that

affect what a person may, must, or must not do in accessing and using forest resources”. Our

measure of the existence of self-made rules gives us an additive index that ranges from 0 to 4.

Based on previous research, we expect that the more types of self-made rules that exist in a given

community, the more likely it is to have a forest in a relatively good and stable condition

(Nagendra, 2002; Hayes, 2006).

Self-organized monitoring and enforcement activities: On an ordinal scale from 0 to 2, this

variable describes the extent to which a community organizes its own monitoring and

enforcement activities. A value of 0 means that no such activities were carried out by the

community, 1 means that the community carried out some monitoring activities, and 2 means that

members carried out a great deal of monitoring and enforcement activities. Supported by findings

from earlier empirical work we predict that community-organized monitoring activities have a

19

positive effect on forest conditions (Gibson et al., 2005, Ostrom and Nagendra, 2006; Agrawal

and Chhatre, 2006; Chhatre and Agrawal, 2008; Coleman, 2009).

Self-organized sanctioning activities: This ordinal variable (0-2) indicates the degree to which

the community implements its own system of sanctions for punishing those individuals who have

been caught violating established rules (regardless of origin). A value of 0 means that no such

activities were carried out by the community, 1 means that the community carried out some

sanctioning activities, and 2 means that they themselves sanctioned infractions a great deal. The

stronger such self-organized systems are the more likely that the community enjoys relative good

forest conditions (Coleman and Steed, 2009).

Combinations of self-governance functions: Local communities choose among no less than

eight different possible institutional designs—unique combinations of the three basic self-

governance functions. Figure 2 presents the distribution of groups and their organization of these

three functions across all 200 groups. More than 85 percent of the groups in our sample have

chosen to organize at least one governance activity related to forest resources, and more than 30

percent of all have organized a full set of all three functions. To test the independent effect of

each unique institutional design we created a dummy variable to represent the existence of each

unique combination of local governance functions. For example, if a community self-organized

their monitoring activities and sanctioned rule breakers, but did not create any of their own rules,

the variable called “Monitor and Sanctions” would be coded as 1, and all the other

“combinations” would be coded as zero for that particular case. A community that had self-

organized all three governance functions would have the variable called “All functions” coded as

1 and all other combinations would take on a value of zero for that case. To test our final claim

that “the more community governance, the better” we create an ordinal variable called “Self-

20

governance”, for which a value of 0 means that the community has not organized any of its own

governance activities; a 1 means that it has organized just one of three possible governance

activity; a 2 means that the community has organized two out of the three governance activities,

and a value of 3means that the community has self-organized all three governance functions.

[Figure 2 about here]

5.3 Control Variables

De Jure Property Rights: In Bolivia, all communities are going through a national program for

land titling. We asked community members at what stage in this process they were. If they had

concluded the initial stage (study stage) we assigned a value of 1, if they had already received

their title, we assigned a 2. All others were coded as zero. Having formal property rights to forests

may, as suggested by many previous studies, affect the long-term stakes and interest that local

community members have in the resource (Alston et al, 1996; Nepstad et al., 2006; Chomitz et al.,

2006)

Forest area: This variable is continuous and indicates the size of the forest (in hectares) that the

community controls and uses. Recent empirical studies have found that the size of forest exerts a

significance influence on community efforts to manage forests and we predict that the larger the

forest the better shape the forest will be in (Chhatre and Agrawal, 2009; Coleman and Steed,

2009)

Indigenous: This continuous variable indicates the proportion of the community that considers

itself part of a lowland indigenous ethnic group. Previous qualitative studies have claimed that

indigenous communities in the lowlands have traditions and cultures that are often more amenable

21

to the preservation of natural resources (Becker and Leon, 2000; Birk, 2000; Nelson and Chomitz,

2009).

Population pressure: We asked local community members about the current population count

as well as the area of the community’s forest. We divided the number of inhabitants with the

forest area. All other factors held constant, we predict that places with high population pressure

are less likely to enjoy good forest conditions (Kaimowitz and Angelsen, 1999; Harrison, 1991).

Literacy rate: We asked what proportion of adult community members know how to read and

write. The higher the literacy rate—a proxy measure of human development—the more likely it is

for the community members to have off-farm employment and thereby reducing the extractive

pressure on the forest, resulting in better forest conditions (Rudel, 1989; Andersen, 2002;

Jorgensen, 2003; Andersson, 2004).

Soil Degradation: In one of our final questions, we asked community members about what

proportion of members cultivate soils that are degraded—a measure of measure of socioeconomic

status. We predict that this variable has a negative effect on forest conditions as poor households

will have shorter time horizons when making forest use decisions and will face fewer options

when it comes to securing their basic needs (Southgate et al, 1991; Skole and Tucker, 1993;

Chomitz and Gray, 1996).

Distance to primary road: We asked respondents about the distance (in km) to the nearest

primary road. We predict that this variable is negatively correlated with forest conditions because

residents of villages located near primary roads will perceive stronger short-term market

incentives to produce agricultural commodities rather than conserving forests (Pfaff, 1997; Hecht

and Cockburn, 1990). This variable is also important because it is a proxy for how accessible the

22

community is for external governance actors, who may be interested in intervening in forest

governance activities. Descriptive statistics for all variables are presented in Table 1 below.

[Table 1 here]

6. Results

The results from the nine multiple regression models, as shown in Table 2, lend support to

the propositions that (1) local user groups play a critical role in forest protection; (2) three self-

governance functions have independent and positive effects on forest conditions; (3) self-

governance is about more than local monitoring and multiple institutional designs affect forest

conditions, and (4) the more governance functions a local user group decides to organize, the

more likely it is that the local forest prospers.

Since the independent variables of interest are all related to self-governance, one might be

worried about collinearity among these three. Our tests indicate that these variables of interest are

not highly correlated with each other (pearson r< 0.4 for all combinations) and we calculated the

Variance Inflation Factor for each variable which confirmed that there is little or no concern for

multicollinearity (avg VIF=1.1).3

[Table 2 here]

In model 1, which uses the locally perceived forest cover as the dependent variable,

community rule-making and monitoring activities have statistically significant, positive, and

independent (p<0.05) effects on the percent forest cover in rural communities. Controlling for

3 The results are also robust to a number of other post-estimation tests. The F-statistics for the OLS model is 15.26 (p

< 0.001). The residuals for the OLS model are normally distributed. To avoid heteroskedasticity problems, all models

are estimated with robust standard errors. For the OLS model, we also examined whether individual observations

have a strong influence by using the Cook’s Distance and Leverage statistics. None of these diagnostic tests indicated

a violation of the standard regression assumptions.

23

other potentially influential variables, our analysis finds that a one-level increase in the rule-

making ordinal variable corresponds to a 3.4 percent increase in community forest cover. A one-

level increase in community monitoring corresponds to a 7.4 percent increase in forest cover.

Community rule making has a positive effect on relative forest biodiversity (model 2): holding all

other variables constant at their mean values, the average effect of a one-unit increase in this

variable corresponds to an eight-percent increase in the likelihood of a community's forest being

more biologically diverse than neighboring forests.

In model 3, we examine the effect of the independent variables on the local group's

perceived vegetation density of their forests relative to neighboring communities' forests. The

only self-governance variable that has a statistically significant effect on forest density is

community sanctioning. As a community strengthens its own sanctioning system from non-

existence to the strongest possible level, the probability of observing relatively dense forest

vegetation increases by about 19 percent, holding all other variables constant at their means.

The statistical results suggest the governance activities that the user groups themselves

have created play a critical role in explaining variable forest protection. In other words, the

performance of community institutions helps explain why some user groups are more successful

than others in preserving their forests.

The results also suggest that monitoring activities are not the only activities related to

local governance that affect forest conditions. In fact, in two of the first three models—which are

the models in which the independent effect of the three governance functions are tested—the

effect of community monitoring is not statistically distinguishable from zero and in none of these

three models is the M&E variable the sole self-governance variable that is statistically significant

(p<0.05). These results would suggest that the governance arrangements for community rule-

24

making and sanctioning are at least as important as community monitoring when it comes to

understanding local variations in forest conditions.

To test this possibility that previous studies may have over-estimated the effect of

community monitoring on forest conditions, we compared our model specification with an

alternative reduced model that includes community monitoring as the only local governance

variable, omitting the community rule making and sanctioning variables. Models 4, 5, and 6

present the results of this test, which show that the omission of other self-governance variables

overestimates the effect of community monitoring on all three measures of forest conditions.

Comparing the regression coefficients for the community monitoring variable in models 4-6 with

the coefficient for the same variable in models 1-3 in Table 2, it appears that the size of the bias

introduced by considering local monitoring without the other two governance functions only is

about 20 percent in model 4, 18 percent in model 5, and 250 percent in model 6.

Next, we examine the possibility that a variety of combinations of self-governance

functions matter for local forest conditions. This is a test of our argument that multiple

institutional designs may affect the conservation of forests. Figure 2 shows that far from all

groups consider these decisions about self-governance as an "all or nothing" decision. In fact, all

eight possible combinations of self-organized governance functions are represented in the sample.

Perhaps most striking is the finding that less than 15 percent of all user groups report no self-

organized forest governance functions in place. The majority of user groups —57 percent—said

they had organized a combination of self-governance functions that consisted of one or two

functions.

To assess the degree to which these different combinations of self-governance functions,

representing the locality’s specific institutional design, and how these affect forest conditions, we

25

reran the regression models with the governance combinations variables in Figure 2 as

independent variables. The results of these regressions are presented in Models 7, 8, and 9. The

existence of the different combinations of self-governance functions were coded as dummy

variables and included as independent variables.4 The control variables remain unchanged from

Models 1-3.

The results, which are presented in Table 2, suggest that the combination of all three local

governance functions (the variable called “All functions”) has a consistently positive effect on

forest conditions (p<0.05). But there are two other, more reduced combinations of functions,

which appear to also have an impact on forest conditions: the combination variable called “Rules

and Monitoring” as well as “Monitoring and Sanctioning. In two out of the three models, these

variables have a positive and statistically significant effect (p<0.10) on forest conditions. These

results would suggest that there are multiple paths to successful community-based forest

governance as there is a diversity of different institutional designs appears to be effective.

Finally, we investigate the idea that one of the critical drivers of good forest outcomes

among community forests in Bolivia is the degree to which the user groups are actively engaged

in the governance process. We hypothesize that the more involvement there is and the more

governance functions the local users organize themselves the better the forest conditions. To test

this final hypothesis the forest condition variables are regressed on the ordinal variable called Self

Governance. We find that the more governance functions organized by the community, the better

it is for conservation outcomes. Figure 3 shows how the probability good forest conditions

4 These dummy variables were created by considering all possible combinations of governance functions and

assigning a value of 1 to a variable that represents each unique combination—the same seven combinations as

represented in Figure 2. We chose to analyze the effects of these combinations in this way rather than creating

conventional interaction terms because of the large number of possible combinations and the collinearity problems

this would cause in the estimation of parameters.

26

(diversity and density) increase as the community adds governance functions. The probability

increases as a community moves from having no self-governance, to having just one function, to

having two functions, and finally to having all three functions –rules, monitoring and sanctions—

organized by the community. The effect is similar on the forest cover variable: for each additional

function that is added to the self-governance package, regardless of order in which these are

added, forest cover increases by an average of 40 percent. We conclude that for all three forest

condition measures, the more governance activities that the users have organized, the better the

condition of the local forest.

All studies on human institutions are susceptible to problems of reverse causality and

endogeneity. In this case, it might seem plausible that groups are more interested in organizing

governance arrangements because the forests they use are in relatively good conditions. There are

two reasons for why this is not a likely scenario for our study. First, the primary theoretical

prediction from the literature on common-pool resources is that groups are likely to be more

motivated to self-organize their governance arrangements when they perceive the resource to be

in decline and threatened, not in excellent conditions (Gibson and Becker, 2000; Ostrom, 1999).

Second, our primary interest is to examine the possibility that there are a variety of governance

functions—beyond compliance monitoring—that correlate with forest conditions, not to test a

particular causal path or direction. That said, future studies would benefit from longitudinal data

so that causal direction may be analyzed more thoroughly.

Taken together, the results suggest that it is important to look beyond monitoring, and

also beyond an “all or nothing” measure of local forest governance. Local forest governance may

look very different from one place to another, and sometimes it is not feasible or possible for

local groups to organize all three governance functions themselves. What these results show is

27

that organizing a limited set of governance functions does not necessarily mean that the forest is

doomed. It appears that there are context-specific characteristics that sometimes make rule

making more important than monitoring, and sometimes monitoring more important than

sanctioning. It also suggests that there are situations in which organized monitoring of rule

compliance is not required for achieving stable forest outcomes. And it would suggest that there

may very well be situations in which having effective local arrangements for compliance

monitoring will not make a difference for saving a forest from degradation. The role of a specific

function of governance seems to depend on what the local exigencies are when it comes to

regulating use or controlling access to the resources.

The findings are consistent with the proposition by Ostrom (2009) that” it is not the

general type of forest governance that is crucial in explaining forest conditions; rather, it is how a

particular governance arrangement fits the local ecology, how specific rules are developed and

adapted over time, and whether users consider the system to be legitimate and equitable” (429).

7. Conclusion

The findings of the paper both support and challenge parts of the existing research on forest

commons governance. Consistent with much of the extant literature, our results show that the

self-governance efforts of forest user groups are key to understanding inter-community variation

in forest conditions. At the same time, our findings from the Bolivian context challenge some

aspects of previous research results from global comparative studies, in particular when it comes

to the role of community monitoring in explaining local forest conditions. We show that there is

more to successful local forest governance than merely community monitoring; rule-making and

sanctioning systems also represent important ingredients of effective forest governance

arrangements.

28

Our findings carry implications for future directions in both the realms of research and

policy. In research, there are several ways in which future studies can build on and further

strengthen this field of inquiry. By collecting over-time data on self-governance, the robustness of

the findings will increase. While this paper has sought to unpack and operationalize self-

governance as a multi-dimensional concept by measuring three distinct functions there are

arguably more dimensions to consider. For example, future studies might go beyond the issue of

whether local rules matter to consider their origins in the local customs and practices. It would

also be interesting to explore the possibility that there are certain types of rules that are more

important than others when it comes to affecting governance outcomes. Moreover, it is often

argued in the governance literature that it is not so much the rule in and of itself that matters but

how it “fits” the specific local biophysical, socioeconomic, and cultural circumstances (Young,

2012, Ostrom, 2005; Basurto and Coleman, 2010), yet empirical tests of this argument are rare.

Since there is only so much one can take on in one paper, we leave these opportunities and

challenges to future studies in this exciting area of research.

For actors involved in the Bolivian forest policy arena, our study raises several issues for

debate. For instance, our results carry implications for Bolivia's recent decision to launch an

alternative, non-market approach to forest conservation as a response to the more market-oriented

initiative called Reduced Emissions from Deforestation and Forest Degradation (D. Pacheco,

2012). Our study shows that about half of all rural communities in Bolivia have forest resources

that are in relatively good conditions. This is a remarkable feat in that the overall deforestation

rate for the country as a whole during this time was the highest ever recorded in Bolivia's recent

history (FAO, 2006). This would suggest that it would make sense to allow rural communities to

play a greater role in forest protection and forest carbon sequestration activities. Their specific

29

roles should be given serious consideration when designing public policies to enhance the

country's forest carbon stocks.

Finally, our study also indicate that there may be limitations to what the national

government can do to affect directly the forest carbon stocks in forests used by rural communities.

Given the consistently positive effects of three different local self-governance functions, one

approach the government might consider is to find ways to support forest-user communities in

their home-grown efforts to govern their forest commons. Exactly how to do so, however, is a

subject for much-needed future research.

30

8. References Cited

Agrawal, A. 1994. Rules, Rule Making, and Rule Breaking: Examining the Fit between Rule Systems and

Resource Use. In ELinor Ostrom, Roy Gardner, and James Walker (Eds) Rules, Games, and Common-

Pool Resources. Ann Arbor, MI: The University of Michigan Press, pp 267-282

Agrawal, A. 2000. Small is beautiful, but is larger better? Forest-management institutions in the Kumaon

Himalaya, India. In C. C. Gibson, M. A. McKean, & E. Ostrom (Eds.), People and Forests.

Cambridge, MA: MIT Press.

Agrawal, A. 2001. Common Property Institutions and Sustainable Governance of Resources. World

Development. 29(10):1649-72.

Agrawal, A. 2005: Environmentality: Technologies of government and the Making of Subjets. Duke University

Press, Durhnam, NC;

Agrawal, A., and Chhatre, A., 2006. Explaining success on the commons: community forest governance in the

Indian Himalaya. World Development 34: 149-166;

Alston, L.J., Libecap, G.D. and Schneider, R. 1996. The determinants and impact of property rights: Land titles

on the Brazilian frontier. J. Law Econom. Org. 12 (x) 25–61

Andersen L.E., Granger C. W.,Reis E. J.,Weinhold D., Wunder S.(2002) The Dynamics of Deforestation and

Economic Growth in the Brazilian Amazon. New York: Cambridge University Press

Andersson, K. 2004. Who Talks With Whom? The Role of Repeated Interactions in Decentralized Forest

Governance. World Development 32(2): 233-249.

Andersson, K. and Ostrom, E. 2008. Analyzing Decentralized Natural Resource Governance from a Polycentric

Perspective. Policy Sciences 41(1):1-23.

Andersson, K. and Pacheco, D. 2004. The Historical Origins of Forestry Policy in Bolivia: Lessons for

Community Forestry. Policy Matters 13: 40-49

Arizpe, Lourdes, M. Priscilla Stone and David Major (eds) (1994) Population and Environment: Rethinking the

Debate. Boulder, CO: Westview Press.

31

Arnold, J. E. M. (1990). Social Forestry and Communal Management in India. ODI: Rural Development

Forestry Network (RDFN).

Ascher, W.1995. Communities and Sustainable Forestry in Developing Countries. San Francisco: Institute for

Contemporary Studies Press.

Baland, J.-M., & Platteau, J. P. (1996). Halting degradation of natural resources: Is there a role for rural

communities?. Oxford, England: Clarendon Press.

Banana, Abwoli Y. and William Gombya-Ssembajjwe (2000), “Successful Forest Management: The

Importance of Security of Tenure and Rule Enforcement in Ugandan Forests,” in Clark C. Gibson,

Margaret A. McKean, and Elinor Ostrom (eds.), People and Forests: Communities, Institutions, and

Governance, Cambridge, MA: MIT Press, 87–98.

Basurto, X., & Coleman, E. 2010. Institutional and ecological interplay for successful self-governance of

community-based fisheries. Ecological economics, 69(5): 1094-1103.

Becker, D. and R. León. 2000. Indigenous Forest Management in the Bolivian Amazon: Lessons From the

Yuracaré People. In Gibson et al (Eds), 2000. People and Forests. Cambridge, MA: MIT Press.

Berkes, F. (Ed.) 1989. Common Property Resources, Ecology, and Community-based Sustainable Development.

London: Bel- haven.

Berkes, F. 2004. Rethinking community-based conservation. Conservation Biology 18:621–630.

Berkes, F. (2007). Community-based conservation in a globalized world. Proceedings of the National Academy

of Sciences, 104(39), 15188.

Birk, G. 2000. Dueños del Bosque: Manejo de los Recursos Naturales por Indígenas Chiquitanos de Bolivia.

Santa Cruz de la Sierra, Bolivia: Apoyo para el Indígena-Campesino del Oriente Boliviano–Central

Indígena de las Comunidades Originarias de Lomerío.

Bray, D. B., Merino-Perez, L., Negreros-Castillo, P., Segura-Warnholtz, G., Torres-Rojo, J. M., & Vester, H. F.

M. (2003). Mexico's community-managed forests as a global model for sustainable landscapes.

Conservation Biology, 17(3), 672-677.

Bromley, D. (Ed.). 1992. Making the commons work: Theory, practice and policy. San Francisco, CA: Institute

for Contemporary Studies Press.

32

Chhatre, A. and Agrawal, A. 2009. Trade-offs and synergies between carbon storage and livelihood benefits

from forest commons. PNAS 106:17667-70.

Chhatre, A. and Agrawal, A. 2008. Forest commons and local enforcement. PNAS 105(36): 13286-91

Chomitz, K.M., Buys, P. De Luca, G. Thomas, T.S. and Wertz-Kanounnikoff, S. 2006. At Loggerheads?

Agricultural Expansion, Poverty Reduction and Environment in the Tropical Forests. Washington, DC:

The World Bank

Chomitz, K.M. and Gray, D.A. 1996. Roads, land use and deforestation, World Bank Econom. Rev. 10 (x) 487–

512

Clugston, R. and T. Rogers. (1995). Sustainable livelihoods in North America.“ Development 3 (9): 603.

Coleman, E.A. (2009). Institutional factors affecting biological outcomes in forest management. Journal of

Policy Analysis and Management 28(1): 122-146.

Coleman, E.A. and Steed, B.C. 2009. Monitoring and sanctioning in the commons: An application to forestry.

Ecological Economics. 68(7): 2106-2113

Comfort, L., Wisner, B., Cutter, S., Pulwarty, R., Hewitt, K., Oliver-Smith, A., et al. (1999). Refraining disaster

policy: the global evolution of vulnerable communities. Environmental Hazards, 1(1), 39-44.

Dietz, T., Ostrom, E. and Stern, P. 2003. The Struggle to Govern the Commons. Science 302(5652):1907–12.

Food and Agriculture Organization of the United Nations (FAO). 1999. State of the World’s Forests 1999.

Rome, Italy: Food and Agriculture Organization of the United Nations.

Food and Agriculture Organization of the United Nations (FAO). 2006. Forest Resources Assessment 2005.

Rome, Italy: FAO

Ghate, Rucha and Harini Nagendra (2005), “Role of Monitoring in Institutional Performance: Forest

Management in Maharashtra, India,” Conservation and Society, 3(2), 509–532

Gibson CC, McKean MA, Ostrom E, eds (2000) People and Forests: Communities, Institutions, and

Governance (MIT Press, Cambridge, MA).

Gibson, C. C., Williams, J. T., & Ostrom, E. (2005). Local Enforcement and Better Forests. World

Development, 33(2), 273-284.

33

Gibson, Clark and C. Dustin Becker. 2000. "The Lack of Institutional Supply: Why a Strong Local Community

in Western Ecuador Fails to Protect its Forest." In Gibson, Clark, Margaret McKean, and Elinor

Ostrom, eds. 2000. People and Forests: Communities, Institutions, and the Governance of Forests:

135-162.

Gibson, Clark and Tom Koontz. 1998. “When Community is Not Enough: Communities and Forests in

Southern Indiana.” Human Ecology 26: 621-647.

Government of Bolivia, 2007. Plan de Desarrollo Sectorial – Revolución Agraria y Forestal. La Paz, Bolivia.

Hardin, G. 1968. The Tragedy of the Commons. Science 162:1243–1248

Harrison, S. 1991. Population growth, land use, and deforestation in Costa Rica, 1950–1984, Interciencia 16 (x)

83–93.

Hayek, F. 1948. Individualism and Economic Order. Chicago: University of Chicago Press.

Hayes, T. M. (2006). Parks, people, and forest protection: an institutional assessment of the effectiveness of protected

areas. World Development, 34(12), 2064-2075.

Hayes, Tanya, and Elinor Ostrom. 2005. “Conserving the World’s Forests: Are Protected Areas the Only Way?

Indiana Law Review 37(3): 595-617.

Hecht, S. and Cockburn, A. 1990. The Fate of the Forest: Developers, Destroyers and Defenders of the

Amazon, Harper Collins, New York

Heckathorn, D. D. (1989). Collective action and the second order free rider problem. Rationality and Society, 1,

79–100.

IPCC, 2007. Fourth Assessment report. Bonn, Germany: International Panel on Climate Change (IPCC).

Jorgensen, A.K. 2004. Uneven Processes and Environmental Degradation in the World-Economy. Human

Ecology Review 11(2):103-117

Larson, A. M. Cronkleton, P. Barry, D. Pacheco, P. 2008. Tenure Rights and Beyond: Community Access to

Forest Resources in Latin America . CIFOR Occasional Paper No. 50. Bogor, Indonesia, Center for

International Forestry Research (CIFOR).

McCay, Bonnie J., and James M. Acheson. 1987. The Question of the Commons: The Culture and Ecology of

Communal Resources. Tucson: University of Arizona Press.

34

McKean, M.A. 1992b. Success on the commons: a comparative examination of institutions for common

property resource management. J. Theor. Polit., 4(3): 247 - 282.

Nagendra, H., Karna, B., & Karmacharya, M. (2005). Examining institutional change: Social conflict in Nepal's

leasehold forestry programme. Conservation and society, 3(1), 72.

Nagendra, H. (2002). Tenure and forest conditions: community forestry in the Nepal Terai. Environmental

conservation, 29(4), 530-539.

Nagendra, Harini. 2007. Drivers of Reforestation in Human-Dominated Forests. PNAS 104(39):15218–23.

Nelson, A. Chomitz, K.M. 2009. Protected Area Effectiveness in Reducing Tropical Deforestation: A Global

Analysis of the Impact of Protection Status. Independent Evaluation Group Evaluation Brief 7.

Washington, D.C. : The World Bank.

Nepstad, D., Schwartzman, S., Bamberger, B., Santilli, M., Ray, D., Schlesinger, P., Lefebvre, P., Alencar, A.,

Prinz, E., Fiske, G., and Rolla, R. 2006. Inhibition of Amazon Deforestation and Fire by Parks and

Indigenous Lands. Conservation Biology 20 (1): 65-73.

Ostrom 2007 A diagnostic approach for going beyond panaceas. PNAS 104(39): 15181–15187.

Ostrom, E. & Nagendra, H. 2006. Insights on linking forests, trees, and people from the air, on the ground, and

in the laboratory. PNAS 103(51): 19224–19231.

Ostrom, E. 1990. Governing the Commons: The evolution of institutions for collective action. Cambridge, UK:

Cambridge University Press.

Ostrom, E. 2001. Reformulating the commons. In J.Burger, E. Ostrom, R. B. Norgaard, D. Policansky, & B. D.

Goldstein (Eds.), Protecting the commons: A framework for resource management in the Americas (pp.

17–41). Washington, DC: Island Press.

Ostrom, E. 1999. Self-governance and Forests. CIFOR Occasional Paper No. 20. Bogor, Indonesia: CIFOR.

Pacheco, D. 2012. Paper presented at the Moore Foundation Workshop “Governance and conservation:

Towards more robust indicators”. Palo Alto, CA. April 25, 2012

Pacheco, P. Barry, D. Cronkleton, P. Larson, A. M. 2008. The role of informal institutions in the use of forest

resources in Latin America . Forests and Governance Programme Series No. 15. Bogor, Indonesia,

Center for International Forestry Research (CIFOR).

35

Pagdee, A. Yeon-Su Kim, and P. J. Daugherty. 2006. What Makes Community Forest Management Successful:

A Meta-Study from Community Forests throughout the World. Society and Natural Resources 19(1):

33–52.

Pfaff, Alexander S. P. 1997. "What Drives Deforestation in the Brazilian Amazon? Evidence from Satellite and

Socioeconomic Data." Policy Research Working Paper 1772. World Bank, Policy Research

Department, Washington, D.C. Processed.

Porter-Bolland, L., E. A. Ellis, M. R. Guariguata, I. Ruiz-Mallén, S. Negrete-Yankelevich and V. Reyes-García

(2012). "Community managed forests and forest protected areas: An assessment of their conservation

effectiveness across the tropics." Forest Ecology and Management 268(1): 6-17.

Rudel, T.K. 1989. Population, development, and tropical deforestation: A cross-national study, Rural Sociology

54 (x): 327–338.

Skole, D. and Tucker, C. 1993. Tropical deforestation and habitat fragmentation in the Amazon: Satellite data

from 1978 to 1988. Science 260: 1905–1910.

Somanathan, E., Prabhakar, R., and Singh Mehta, B. 2009. Decentralization for cost-effective conservation.

PNAS 106 (11) 4143-4147

Southgate, D., Sierra, R. and L. Brown. 1991. The causes of tropical deforestation in Ecuador: A statistical

analysis, World Develop. 19 (9):1145–1151.

United Nations Development Program (UNDP). 1998. Capacity 21 Program 1998 Annual Report. New York:

United Nations Development Program.

Van Laerhoven, F. (2010). Governing community forests and the challenge of solving two-level collective

action dilemmas--A large-N perspective. Global Environmental Change, 20(3), 539-546.

Webb, Edward and Ganesh P. Shivakoti, eds. 2008, Decentralization, Forests and Rural Communities: Policy

Outcomes in South and Southeast Asia, New Delhi: Sage India.

Wunder, S. 2001. Poverty alleviation and tropical forests—what scope for synergies? World Development

29:1817–1833.

36

FIGURES AND TABLES

FIGURE 1: Proportion of Communities that Enjoy Relatively Good Forest Conditions

0%

10%

20%

30%

Forest Cover >50% Forest Diversity Forest Density

Forest Condition Indicators

37

Figure 2: The Distribution of the Unique Combinations of Self-Governance Functions

0%

5%

10%

15%

20%

25%

30%

35%

No functions OnlySanctioning

OnlyMonitoring

Only Rules Rules andSanctioning

Monitoring andSanctioning

Rules andMonitoring

All functions

Combinations of Governance Functions Chosen by Communities

38

Figure 3: The effect on forest conditions of increasing the number of functions that the user

groups organize themselves.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

No self-governance

One self-governance

function

Two self-governance

functions

All self-governance

functions

Pro

bab

ility

More Self-Governance is Better for Forest Conditions

Density

Diversity

39

Table 1: Descriptive statistics for all variables included in the analysis

Variable Obs Mean StDev Min Max

Dependent

Forest Cover (2007) (%) 200 29.13 26.8 0 99

Relative Forest Vegetation Density (binary) 200 0.22 0.41 0 1

Relative Forest Vegetation Diversity (binary) 200 0.18 0.39 0 1

Independent

Community Rules for forest use (ord) 200 1.01 1.17 0 4

Community Monitoring System (ord) 200 1.04 0.75 0 2

Community Sanctions (ord) 200 0.71 0.82 0 2

Self Governance (ord) 200 1.60 1.11 0 3

Combinations

Rules + Monitor (binary) 200 0.10 0.30 0 1

Rules + Sanctions (binary) 200 0.05 0.22 0 1

Monitor + Sanctions (binary) 200 0.10 0.30 0 1

All functions (binary) 200 0.32 0.47 0 1

Controls

De jure property rights (ord) 200 0.57 0.51 0 2

Forest Cover (1000 ha) 200 5.60 31.04 0.00 370

Indigenous percentage (%) 200 0.46 0.50 0 100

People per ha of forest (cont) 200 60.64 190.9 0.00 1800

Distance to nearest primary road (km) 200 9.70 16.25 0.00 90.0

40

Table 2: The Effects of Self-Governance Regimes on Forest Conditions in Bolivia

Independent

Variables

Model 1

Cover

Model 2

Diversity

Model 3

Density

Model 4

Cover

Model 5

Diversity

Model 6

Density

Model 7

Cover

Model 8

Diversity

Model 9

Density

Rules 3.41(1.48)** 0.58(0.20)** 0.34(0.19) --- --- --- --- --- ---

Monitoring 7.50(2.45)*** 0.60(0.34) 0.19(0.29) 9.05(2.31)*** 0.71(0.31)** 0.47(0.27) --- --- ---

Sanctions 2.86(2.26) 0.20(0.28) 0.61(0.25)** --- --- --- --- --- ---

Rules, Monit. --- --- --- --- --- --- 0.14(0.53) 1.97(0.75)** 1.31(0.70)*

Rules, Sancts --- --- --- --- --- --- 0.45(0.65) 1.53(0.94) 1.72(0.93)*

Monit, Sancts --- --- --- --- --- --- 0.51(0.34) 1.56(0.78)* 2.25(0.63)***

All funct's (bi) --- --- --- --- --- --- 0.81(0.26)** 1.89(0.61)** 1.88(0.50)***

Prop rights 12.44(2.98)*** 0.67(0.44) -0.28(0.40) 12.58(2.97)*** 0.74(0.44) -0.27(0.39) 0.69(0.23)** 1.01(0.45)* -0.03(0.43)

Forest (ha) 0.00(0.00)** 0.00(0.00) 0.00(0.00) 0.00(0.00)** -0.00(0.00) 0.00(0.00) 0.000.00* 0.00(0.00) 0.00(0.00)

Indigen. (%) 0.00 (0.06) 0.00(0.01) -0.03(0.02) 0.11(0.06) 0.01(0.01) -0.02(0.02) 0.00(0.00) 0.01 (0.01) -0.03(0.01)

Pop/ha forest -0.03(0.01)*** 0.00(0.00) -0.00(0.00) -0.04(0.01)*** -0.00(0.00) -0.00(0.00) -0.00(0.00)*** 0.00(0.00) 0.00(0.00)

Literacy (%) -0.06(0.07) 0.00(0.01) 0.02(0.01) -0.06(0.07) -0.00(0.01) 0.01(0.01) -0.01(0.00) 0.00(0.01) 0.02(0.01)

Degr.soils (%) -11.72(4.27)** 0.14(0.79) 0.43(0.47) -8.60(4.05)* 0.64(0.55) 0.77(0.47) -1.27(0.29)*** 0.46(0.69) 0.62(0.46)

Dist.road -0.00(0.14) 0.01(0.01) 0.00(0.01) -0.04(0.14) 0.00(0.01) -0.00(0.01) 0.01(0.01) 0.01(0.01) 0.00(0.01)

41

Constant 15.85(5.53)*** -3.48(0.76)*** -3.34(0.88)*** 18.99(5.68)*** -2.85(0.73)*** -2.55(0.78)*** 2.69(0.42)*** -3.83(0.94)*** -3.67(0.93)***

F 15.28 - - 13.12 - - 12.90 --- ---

Wald chi2 - 34.03 27.73 - 24.01 11.81 --- 31.61 28.37

Prob > F 0.00 - - 0.00 - - 0.00

Prob > chi2 - 0.00 0.00 - 0.00 0.00 --- 0.00 0.00

R2 0.35 - - 0.31 - - 0.27

Pseudo R2 - 0.17 0.15 - 0.10 0.08 --- 0.16 0.17

Obs 200 200 200 200 200 200 200 200 200

Note: Columns list coefficients with robust standard errors in parenthesis.

Two-tailed hypothesis tests: *=p<0.10; **= p<0.05; ***=p<0.01