Social and Environmental Correlates of Palm Oil ...
34
Social and Environmental Correlates of Palm Oil Certification in Indonesia Matthew G. Borden (Advisor: Subhrendu K. Pattanayak) May 2, 2016
Transcript of Social and Environmental Correlates of Palm Oil ...
Social and Environmental Correlates of Palm Oil Certification in Indonesia
Matthew G. Borden
(Advisor: Subhrendu K. Pattanayak)
May 2, 2016
ii
Social and Environmental Correlates of Palm Oil Certification in Indonesia
Matthew G. Borden May 2, 2016
Abstract
Indonesia is losing more of its forests and, in Kalimantan, palm oil is contributing to increasing rates of forest loss. The Roundtable on Sustainable Palm Oil (RSPO) is a conservation scheme that harnesses market-based principles to internalize negative social and environmental externalities associated with palm oil production, through forest certification. To date, few studies use quantitative data to examine these certification schemes. This paper makes use of concession data and village-level environmental, economic, legal, and social data to identify correlates of RSPO in Kalimantan. Significant variables include: forest cover, total village carbon, average slope, proximity to markets, village protected areas, population density, and malnutrition. I find that RSPO concessions are generally correlated with suitable oil palm cultivation characteristics. However, RSPO concessions are somewhat correlated with higher total village carbon and negatively correlated with malnutrition, suggesting reduced potential for social and environmental co-benefits. The analysis reported here can be viewed as an important first stage in more targeted evaluations of RSPO and other certification schemes.
iii
Table of Contents
Acknowledgements ........................................................................................................................ iv
1 Policy Question ....................................................................................................................... 5
2 Introduction ............................................................................................................................. 5
2.1 Background ..................................................................................................................... 5
2.2 Certification and Sustainable Forest Management ......................................................... 7
2.3 Study Setting ................................................................................................................. 10
2.4 Research Objectives ...................................................................................................... 10
3 Materials ............................................................................................................................... 11
3.1 Time Period and Region ............................................................................................... 11
3.2 Unit of Analysis and Treatment Definition .................................................................. 12
3.3 Covariates ..................................................................................................................... 14
3.4 Data Sources ................................................................................................................. 16
4 Methods................................................................................................................................. 17
5 Results ................................................................................................................................... 18
5.1 Environmental ............................................................................................................... 18
5.2 Economic ...................................................................................................................... 19
5.3 Legal and Social ............................................................................................................ 19
Conclusions ................................................................................................................................... 22
References ..................................................................................................................................... 23
Appendices .................................................................................................................................... 28
Appendix 1: Evidence of Deforestation in Kalimantan ............................................................ 28
Appendix 2: The Forest Stewardship Council Certification Program ...................................... 29
Appendix 3: Initial covariate list and definitions ...................................................................... 31
iv
Acknowledgements
I owe many thanks to those who have helped me reach this point. I am grateful for my advisor, Professor Subhrendu K. Pattanayak, for his guidance and support from the beginning. I give thanks also to Dr. Daniela Miteva for her generous feedback and assistance with the data. Next, I appreciate my colleagues, Rahman Adi Pradana, Alisha Pinto, and Ariadne Rivera Aguirre, for their equal parts of technical advice and good humor. “Terima kasih” to the Green Warriors in East Java for opening my eyes to environment. Finally, I thank my family and friends for all they have done to encourage me.
“We are able to create new conditions, a new reality. We are not fated to swim forever among the realities that are here now.”
Pramoedya Ananta Toer
5
1 Policy Question What are the different environmental, economic, legal, and social conditions correlated with the
presence of Roundtable on Sustainable Palm Oil (RSPO) certified oil palm concessions in
Kalimantan?
Without establishing an initial sense of these factors, we cannot conduct more careful and
systematic evaluation of the impacts of new conservation schemes such as RSPO.
2 Introduction
2.1 Background In Indonesia, tropical forests and peat lands face increasing pressure for development to
supply food, fuel, and other forest products (DeFries et al. 2010; Gibbs et al. 2010; Ferraro et al.
2013; Miteva et al. 2015). Two patterns in Indonesia’s deforestation problem set it apart. First,
absolute rates of deforestation in Indonesia have increased dramatically (Hansen et al. 2013;
Margono et al. 2013). Second, the palm oil industry is expanding, thus leading to deforestation at
higher rates in the region (Gaveau et al. 2013; Stibig et al. 2013; Busch et al. 2014; Abood et al.
2015). See Appendix 1 for more evidence of deforestation.
Indonesia’s deforestation problem is well documented and the literature analyzes diverse
policy responses though few studies focus on policies targeting palm oil as a driver of
deforestation. Some conventional place-based policies, such as moratoriums and concessions,
fail to adequately protect against deforestation due to conflicting interests and ineffective
government responses (Auld 2008, Vincent 2008, Busch et al. 2015). Recent studies show that
other place-based policies, such as protected areas (PAs) and payments for environmental
services (PES), can protect against forest loss while reducing poverty (Snider et al. 2003; Weber
et al. 2011; Miteva et al. 2012; Heino et al. 2015). Innovative price-based policies, such as forest
certifications, have proliferated in Indonesia and elsewhere (Forest Stewardship Council 2014),
yet there is still little empirical evidence on their impacts (Pattanayak 2009; Blackman et al.
2010; Romero et al. 2013, Rainforest Alliance 2014).
6
Although certification programs contribute a new level of sophistication by using market
principles to deliver environmental protection, few programs anticipate and target the growth in
deforestation associated with palm oil. Through its certification programs, the Roundtable on
Sustainable Palm Oil (RSPO) initiative aims to protect forests and contribute toward the
wellbeing of local communities and also promote sustainable palm oil production and trade
(RSPO 2013). To my knowledge, this is the first quantitative study of the different
environmental, economic, legal, and social correlates of RSPO certified oil palm concessions in
Kalimantan.
In general, deforestation and forest degradation result in more than the simple loss of
natural capital. Tropical forests contribute to global goals of biodiversity conservation and
climate change mitigation, globally, while impacting livelihoods and wellbeing locally (Sills et
al. 2003; Gibbs et al. 2007; Ferraro et al. 2011; Primm et al. 2014). Deforestation leads to the
loss of these environmental services through the channels of forest cover, habitat fragmentation,
forest fires, and pollution. Deforestation also bears negative effects on household welfare and
village development (Sills et al. 2003). Forests are public goods which supply environmental
services by way of externalities. Decision makers often fail to consider these externalities,
leading to the exploitation of forests by way of deforestation and forest degradation and, hence,
the undersupply of environmental services.
Place-based approaches for environmental protection, such as PAs, and moratoriums and
concessions, have mixed effects on deforestation (Miteva et al. 2015). These are mandatory,
government-administered instruments focused on regulating the behavior or performance of
individual firms but they do not address underlying incentives driving deforestation (Keohane
and Olmstead 2007). The resulting restrictions on the global supply of forest products increases
both prices and incentives for deforestation in tropical regions (Keohane and Olmstead 2007;
Burgess et al. 2012). Protected Areas are the most commonly used instrument for biodiversity
conservation in developing countries (Miteva et al. 2012). They work by placing legal
restrictions on human access and use within their boundaries and impose penalties on offenders.
On a global scale, PAs covered an estimated 14.4 percent of the land and 2.8 percent of the
oceans in 2014 (Pimm et al. 2014). Despite their coverage, there exists far less evidence of their
impacts (Ferraro et al. 2015). The available studies show that PAs have delivered mixed impacts
across political-economic and geographical landscapes (Joppa et al. 2010a, 2010b; Miteva et al.
7
2012; Pfaff et al. 2014; Heino et al. 2015). Additionally, archetypal forms of the policy fail to
compensate rural people living near PAs who traditionally rely on forests for subsistence
(McElwee 2008). However, when paired with PES, PAs can deliver both forest protection and
poverty reduction (Weber et al. 2011).
Regarding Indonesia’s national forest moratorium, the evidence is less than promising.
Concessions employ the Indonesian Selective Cutting and Planting System (TPTI), which sets
rotation time and the diameter of the harvested trees, necessitates replanting in the presence of
low natural restocking rates, bans logging in sensitive areas, but does not require reduced logging
techniques (Ruslandi et al. 2014). Forest management under TPTI has been flawed, which,
coupled with weak law enforcement, has contributed toward a dramatic decrease in traditional
logging concessions (Resosudamo 2005; Burgess et al. 2012; Ruslandi et al. 2014). Considering,
also, the conversion of logged forests into oil palm plantations, forest mismanagement caused a
decrease in logging concessions from 59 million ha in 1990, to 25 million ha in 2011 (Ruslandi
et al. 2014).
The RSPO certification program adds to an existing collection of “information-based”
approaches for environmental protection (Blackman 2010). In contrast to place-based
approaches, ecolabeling and certification programs are voluntary and decentralized approaches
which aim to influence the behavior of individual consumers and firms by changing the
incentives they face (Keohane and Olmstead 2007). Information-based approaches aim to
overcome market failures from asymmetric information. Certification programs create incentives
for sustainable forest management (SFM); first, by providing a price premium to producers
engaging in SFM; and, second, by facilitating access to markets that otherwise would be closed
to producers with unsustainable practices (Auld et al. 2008; Araujo et al. 2009). Sustainable
forest management is no monolith and certification programs’ success in delivering
environmental and social co-benefits, depends on the robustness of each individual program’s
criteria and their application in the field. For an example of an established certification program,
the Forest Stewardship Council (FSC), see Appendix 2.
2.2 Certification and Sustainable Forest Management The RSPO initiative is an example of an emerging “partnered governance” model
intended to promote sustainable development (Nikoloyuk et al. 2010). Established in 2004, the
8
initiative functions as a voluntary, non-state, information-based approach to environmental
protection and development. It incorporates oil palm plantation certification for SFM, PAs, as
well as other programs targeting social outcomes (RSPO 2013). Unlike third-party certification
programs, and as implied by its name, the RSPO was established by an association of
stakeholders including palm oil producing firms as well as non-governmental conservation
advocacy organizations. RSPO manages two certification programs to support its two goals; to
promote the production as well as the trade of certified sustainable palm oil (RSPO 2013). The
principles and criteria certification program (P&C) aims to ensure that palm oil is produced
sustainably at the plantation level and the supply chain certification program aims to ensure the
integrity of trade in certified sustainable palm oil (RSPO 2013). This paper is focused on the
P&C program
The P&C program is designed to incentivize participating firms to internalize the social
and environmental spillover effects of palm oil production (RSPO 2013). In addition, the
program emphasizes commitment to and compliance with transparency; local laws and
international agreements; long-term economic and financial viability; use of appropriate best
practices by growers and millers; environmental responsibility and conservation of natural
resources and biodiversity; responsible consideration of employees, and affected local
communities; responsible development of new plantings; and continuous improvement in key
areas of activity (RSPO 2013). The P&C promotes environmental protection by requiring its
plantations to comply with the following: (i) plans for reducing pollution and emissions; (ii)
improved fire management; (iii) riparian and other buffer zones; and (iv) protected areas that are
maintained by plantations but cannot be cleared (RSPO 2013). The P&C program also requires
that plantations and millers contribute to local sustainable development, at the village level,
where appropriate (RSPO 2013) (Fig 1).
Unlike other certification programs, RSPO’s P&C program promotes a second,
seemingly contradictory, goal; palm oil production. Even given palm oil’s growing annual
contributions to deforestation, carbon emissions, and other undesirable outcomes, palm oil
production need not exist in conflict with conservation goals (Corley 2008). Palm oil production
can generate local benefits if cultivation follows sustainable planning and management practices.
Potential benefits include increased incomes, profits, government revenues, reduced poverty, and
9
Fig 1. Summary of the intended RSPO impacts. The figure is based on RSPO 2013.
improved natural resource management (Gingold et al. 2012). The World Resources Institute
(WRI) developed standards and online tools to identify potentially suitable areas for oil palm
cultivation, designed in accordance with established standards including RSPO’s P&C, relevant
Indonesian laws and policies, and proposed national REDD+ strategies to support palm oil
production on low-carbon degraded land (Gingold et al. 2012). WRI’s oil palm cultivation
suitability standard is based on a set of environmental, economic, legal, and social indicators.
The RSPO P&C program’s effectiveness in internalizing the social and environmental
spillover effects of palm oil production hinges on the robustness of program criteria and their
application in the field. This paper focuses on the program’s potential to promote sustainable
palm oil production, using SFM, at the plantation level. To describe potential, this paper
examines the conditions correlated with RSPO concessions within villages in Kalimantan.
Environmental • Maintenance of
High Conservation Value habitats
• Plantation and mill management mitigates impacts
Economic • Management plan
aims to achieve long-term economic and financial viability
Legal • Compliance with all
applicable laws and regulations
• Use of land for oil palm does not diminish other users’ rights
Social • Participatory approach
to plantation and mill management
• Growers and millers contribute to local development
10
2.3 Study Setting Indonesia exemplifies the difficult tradeoffs that consumers, firms, and policymakers
must consider with regard to development and deforestation. The country has pledged to reduce
greenhouse (GHG) emissions, driven in large part by palm oil, by 26 percent unilaterally, and up
to 41 percent with international support, compared to a projected business-as-usual baseline, by
2020 (Government of Indonesia 2015). At the same time, Indonesia has pledged to grow its
agriculture sector, including doubling palm oil production by 2020 (Maulia 2010). Since 2000,
more than 70 percent of palm oil expansion has occurred at the expense of peat lands and
primary, secondary, and agro-forests, leading to significant contributions to Indonesia’s GHG
emissions (Carlson et al. 2013; Gunarso et al. 2013; Ramdami et al. 2013). The dual goals of
improved conservation and increased palm oil production seem in conflict. If RSPO’s P&C
program can deliver increased production and social and environmental co-benefits, it could
serve as one example of a successful policy response in the face of Indonesia’s development and
deforestation tradeoffs.
2.4 Research Objectives Overall, this paper aims to examine which different environmental, economic, legal, and
social conditions are correlated with the presence of RSPO certified oil palm concessions in
Kalimantan. This is an important first step in evaluating RSPO effectiveness and impacts
because it sets up the basis for establishing the counterfactual (Pattanayak 2009). For example, if
RSPO concessions are typically in poor rural communities, then any evaluation of its impacts
must compare RSPO communities to other poor rural communities.
In order to shortlist potential correlates of RSPO, I rely on the logic presented in Lin et al.
(2012) which uses a simple cost benefit model to explain REDD+ proponents’ project site
selection in Brazil and Indonesia, consistent with the focus on additionality. At the subnational
level, proponents are likely to consider costs associated with reducing emissions, yet many of the
factors encouraging deforestation are also likely to increase the difficulty of project
implementation (e.g. high opportunity costs, high population density, unclear tenure and poor
governance) (Lin et al. 2012). The authors found that projects are more likely to be placed in
inaccessible areas and in municipalities with higher population density but lower poverty,
yielding mixed evidence on the role of expected poverty alleviation co-benefits in site selection
11
(Lin et al. 2012). They also found strong evidence of the potential for biodiversity co-benefits,
indicated by positively and strongly significant coefficients on percent of land in PAs (Lin et al.
2012). I expect RSPO concessions to be associated with signs on these same correlates that
suggest potential for social and environmental co-benefits, see section 3.3 “Covariates” for more.
To identify criteria specific to oil palm concession site selection, I refer to WRI’s oil
palm cultivation suitability standard. High performance is indicated by the presence of
concessions in villages associated with high potential suitability for oil palm cultivation (Gingold
et al. 2012). Some site selection considerations include: (i) preference for low-carbon degraded
land, to conserve carbon and biodiversity outside of plantations; (ii) preference for suitable
climatic and topographical conditions and access to markets, to promote productivity and
financial viability; (iii) preference for areas with the appropriate legal zoning classification and
mechanisms for addressing land use claims, to pre-empt illegal land use change; and (iv)
preference for areas where local communities’ land use and interests are recognized and voiced,
to avoid negative social impacts (Gingold et al. 2012). In other words, I expect that RSPO
concessions are associated with sustainable oil palm cultivation characteristics across
environmental, economic, legal, and social considerations.
3 Materials
3.1 Time Period and Region I focus on the presence of oil palm concessions with RSPO certificates issued between
2010 and 2013 in Kalimantan. The spatial and temporal frames were determined by data
availability, as described by Dr. Daniela Miteva. General oil palm concession data are publicly
available for download at WRI’s Open Data Portal. RSPO certified concession data are publicly
available at the same site, but only for reference. Although RSPO has issued certificates since
2013, data associated with those more recent concessions are not publicly available. Considering
that some RSPO certified concessions analyzed for this paper were certified as early as 2010, I
treat the year 2008 as my baseline for covariate data.
12
3.2 Unit of Analysis and Treatment Definition This paper uses villages as the unit of analysis for three reasons, following Miteva et al.
(2015). First, as there is no unclaimed land; village areas sit entirely within concession areas (Fig
2). Second, because RSPO impacts forests through targeting Low Conservation Value areas and
creating and supporting PAs, ecological indicators (e.g. forest cover, peat area) capture aggregate
impacts of any and all RSPO activities at the village level. Third, village-level characteristics
(e.g. malnutrition, population density) are likely to affect the presence of an RSPO certified
concession. The villages of interest are comprised of all villages in Kalimantan that partially or
fully intersect with RSPO concessions established between 2010 and 2013 (Fig 2).
Fig 2. Distribution of oil palm concessions in Kalimantan. RSPO concessions established between 2010 and 2013 appear in blue. The green polygons are all traditional oil palm concessions established after 2010.
13
In total, 113 villages in Kalimantan intersect with RSPO concessions (Table 1). Village area for
villages intersected by RSPO concessions comprise 2.69 million ha. Concession area for RSPO
concessions comprise .42 million ha. For a more robust assessment (Model 3), I consider those
villages comprised of at least 5 percent oil palm concessions (i.e., RSPO > 5 percent of village
area), which reduces the number of treatment villages from 113 to 83. Village area reduces from
2.69 million ha to 1.94 million ha and RSPO concession area reduces from .42 million ha to .41
million ha. Total village area reduces from 30.47 million ha to 25.33 million ha and RSPO
concession area reduces from 10.15 million ha to 9.38 million ha. Among RSPO-intersected
villages, 27.86 percent of total village area and 1.98 percent of concession area is associated with
villages comprised of less than 5 percent RSPO concession area.
The focus of this paper is the placement of RSPO concessions. I identify the
environmental, economic, legal, and social conditions correlated with RSPO certification. To
examine correlation, I use RSPO presence as a binary variable as well as fraction of a village
comprised of RSPO concessions as a continuous variable.
Table 1. Descriptive statistics for village groups. Models 1 and 2 (includes all villages)
Number of villages Village area (ha) Concession area (ha)
Percent of village as concession
Villages intersected with RSPO concessions 113 2,693,909 416,451 21%
Villages intersected with non-RSPO concessions 2,570 27,777,482 10,148,140 43%
All villages intersected 2,683 30,471,390 10,564,591 42%
Models 3 and 4 (drops villages: concession area < 5% of village area)
Number of villages Village area (ha) Concession area (ha)
Percent of village as concession
Villages intersected with RSPO concessions 83 1,943,394 408,186 29%
Villages intersected with non-RSPO concessions 2,266 23,383,887 9,379,721 48%
All villages intersected 2,349 25,327,281 9,787,907 48%
14
3.3 Covariates I focus on village characteristics that are likely to influence both the placement of RSPO
concessions and the outcomes. I selected covariates based on the logic summarized in Lin et al.
(2012) as well as indicators from WRI’s model for oil palm cultivation suitability. Although this
paper’s covariates do not match either model exactly, they still address environmental,
economic, legal, and social considerations. Selected covariates include baseline forest cover, rice
agriculture village area, total village carbon, fraction of village under peat, average village
elevation, average village slope, proximity to markets, village with some area falling under PAs,
village with some area adjacent to PAs, type of property rights within a village, presence of
malnutrition, and presence of street lights in a village. Proximity to markets is proxied by
distance to major cities, distance to provincial capitals, distance to the nearest market with
permanent structures, and the distance to the nearest port interacted with the sea depth at the port,
in order to distinguish between ports based on their commercial importance (larger vessels
necessitate deeper ports). All the covariates refer to the year 2008 and prior – the period
preceding certification.
I expect RSPO concessions to be associated with signs on correlates that suggest
potential for social and environmental co-benefits. For example, some correlate signs include:
total carbon in village (-); percent of land in PAs (+); population density (-); malnutrition (+).
Descriptive statistics for the covariates are reported in Table 2. The table also shows the
predicted effects the covariates will have on the presence of RSPO concessions within a village,
based on Lin et al. (2012) and WRI’s oil palm cultivation suitability standard (Gingold et al.
2012). A complete list of initial covariates and covariate definitions are reported in Appendix 3.
15
Table 2. Summary of the covariates used in the analysis.
Variable
Outcome Obs. Mean Std. Dev. Min. Max. Expected
sign Village area 2,683 11,357 30,163 94 460,737 N/A Concession area non-RSPO 2,683 3,782 5,560 0 76,407 N/A Concession area RSPO 2,683 155 1,251 0 27,023 N/A Fraction village non-RSPO 2,683 43% 31% 0% 100% N/A Fraction village RSPO 2,683 21% 19% 0% 100% N/A RSPO presence, 1 if 2,683 4% 20% 0 1 N/A
Environmental Forest cover 2,185 47 15 6 78 - Area rice agriculture 1,782 551 2,266 0 71,672 + Area non-rice agriculture 1,782 8,466 17,236 0 296,548 - Carbon in village 2,683 5 46 0 1,549 - Fraction village under peat 2,185 12% 24% 0% 100% -
Economic Elevation 2,185 66 79 1 938 - Slope 2,185 3 3 0 20 - Distance to major city 2,185 41 31 0 175 - Distance to capital 2,185 170 101 4 544 - Distance to port*depth of port 2,177 0.28 0.26 0.00 1.78 - Distance to permanent markets 2,149 31 32 0 99 -
Legal Fraction village under PA 2,185 3% 15% 0% 2% - Village PA protection, 1 if 2,683 16% 53% 0 1 + Village PA adjacent, 1 if 2,683 13% 33% 0 1 + Fraction village shared owner 1,797 0% 3% 0% 92% - Fraction village private owner 1,797 30% 35% 0% 100% +
Social Malnutrition 2,185 26% 37% 0 1 + Government electricity 1,892 307.7 647.28 0 10,130 - Population density 2,185 .6637 2.986 0 100 - Religious diversity 1,333 49.9% 37.34% 0 1 -
16
3.4 Data Sources The analysis makes use of two forms of data: RSPO concession data in the form of an
ArcGIS shapefile and dependent variable data. The RSPO concession shapefile is needed to
identify and map certified oil palm concessions and to identify those villages that intersect with
concession areas. The shapefile includes relevant metadata for each concession including, but not
limited to, the following: plantation name, date of certification and expiration, activity status, and
geographic coordinates.
RSPO’s most recent concession shapefiles are not public documents and obtaining these
data would require gaining access from the RSPO Secretariat. My petition to the Secretariat was
rejected due to legal considerations. Instead, I created my own shapefile using data publicly
available on WRI’s Global Forest Watch Commodities site. First, I downloaded a 2013 shapefile
from WRI’s Open Data Portal, which includes all of Indonesia’s oil palm concessions. Next, I
referenced a layer for RSPO concessions available on WRI’s Global Forest Watch Commodities
site. Unfortunately, the same layer (shapefile) was not available for download. Using ArcGIS, I
selected those concessions on the general shapefile that visually matched with those on the
reference site. I then intersected these concessions with a village shapefile to generate a list of
villages that intersect. I also calculated the percent of each intersected village that is comprised
of RSPO concessions.
Dependent variable data refers to the list of environmental, economic, legal, and social
indicators that this paper uses to determine correlation with RSPO presence. For a complete list
of initial variables, see Appendix 3. For forest cover, I used data collected from the MODIS
Vegetation Continuous Fields datasets. I used socio-economic data collected from PODES
datasets, which are constructed using village-level censuses. Some PODES indicators include:
malnutrition, availability of street lights in a village, population density, and religious diversity.
The data for the first set of variables are publicly available. The authors of Miteva et al. (2015)
made all these data available for this paper.
17
4 Methods I estimate various models to test for correlations between environmental, economic, legal,
and social conditions and the presence of RSPO concessions (y variable) in Kalimantan’s
villages. The various models are as follows:
• Model 1 uses a logistic regression with a parsimonious specification, including, forest cover,
distance to city, presence of PAs in village, and population density, for the full set of 113
villages.
• Model 2 also uses a logistic regression but a fuller specification including more covariates
for the full set of 113 villages.
• Models 3 and 4 serve as robustness checks. Model 3 uses ordinary least squares (OLS)
regression to examine if the same variables influence the amount of certified area within a
village. It is more than just a binary indicator.
• Model 4 also uses OLS regression but drops those villages comprised of less than 5 percent
concession area, leaving 83 villages from the original list of 113.
The outcome variable in Models 1 and 2, yj, is a binary variable signifying the presence of RSPO
certified concession area within village j; βk (k = 1, …, n) are the parameters that will be
estimated as described in Table 2; and εj~N(0,σ) are the normally distributed error terms.
(1 & 2) Ln $%&'$%
= )* + )&,&- + ⋯+ )/,/- +0-
The outcome variable in Models 3 and 4, Yij, is a measure of the percent of village area
comprised of RSPO concessions within village j; βki (k = 1, …, n) are the parameters that will be
estimated as described in Table 2; and εj~N(0,σ) are the normally distributed error terms.
(3 & 4) 12 = )*2 + )&2,&- + ⋯+ )/2,/- +02- (i = RSPO certification)
First, I checked pair-wise correlation between all the variables of each category of
variables (environmental, economic, legal, social). I noted significant collinearity amongst some
variables; among these, I selected the variables which best fit the theory and used them in my
18
models. Those collinear variables that I excluded are the following: fraction of village under PA
(collinear with village PA protection), fraction of village under private ownership (collinear with
fraction of village under shared ownership), and village reliance on private electricity (collinear
with village reliance on government electricity). Additionally, I excluded the following: non-rice
agricultural village area, distance to capital cities, fraction of village under shared ownership,
village reliance on government electricity, and presence of street lights in village because they
have no effect on RSPO presence. I include both elevation and slope although all observations
meet the suitability criteria: for elevation, < 1,000 m; and, for slope, < 30°).
5 Results The final logistic and OLS regression results for the four models are reported in Table 3.
In summary, Model 1 returned strongly significant results; 3 of the 4 covariates have expected
signs. Though distance to a major city has an unexpected sign, Model 2’s fuller specification
corrects this. In Model 2, I find that 4 of the 9 covariates have expected signs, but of the 5
covariates that returned significant and contrary results, I am concerned primary with total
village carbon; the results show that greater measures of carbon in villages are correlated with a
higher probability of RSPO certification. Of the remaining significant covariates that resulted in
signs contrary to my predictions, the negative sign on the malnutrition variable suggests that
RSPO does not target poorer or less developed areas, ostensibly reducing potential for social co-
benefits. The market proximity variables also returned surprising results though, in combination,
the results appear both logical and favorable because they suggest that proximity is correlated
with a higher probability of certification. Overall, the results show that RSPO concessions are
correlated with suitable oil palm cultivation characteristics. The remainder of this section
discusses significant outcomes for variables by categories. The results from Models 3 and 4 do
not, however, offer any support for the theory that covariates predict the amount of RSPO
concession area in villages.
5.1 Environmental Forest cover, under Model 2, is negatively correlated with the presence of RSPO
concessions, suggesting that RSPO concessions are found in villages with degraded land which
is more suitable to sustainable oil palm cultivation. Under Models 3 and 4, forest cover has no
19
effects. Total villages, under Model 2, is positively correlated with RSPO presence at the 1
percent level. Although its effect size is relatively small, in conjunction with a negative sign on
forest cover, the positive sign for total village carbon may represent belowground soil carbon, or
peat. This is potentially concerning because one of RSPO’s goals is to expand monoculture oil
palm cultivation. However, this may not be concerning if RSPO is conserving the carbon-rich
portions of its concessions in PAs as required under the P&C program. Contrasting with
environmental features of FSC villages from Miteva et al. (2015) – the closest to this paper – I
found that villages with greater total carbon are correlated with a higher probability of RSPO
certification.
5.2 Economic Market access is proxied by distances to (i) a major city, (ii) near market with permanent
features, and (iii) to a near port interacted with sea depth at port. Under Model 2, all three
covariates are significant. The first two covariates return relatively small yet positive effect sizes,
contrary to my prediction. However, the larger negative effect size of the third variable
dominates the smaller effects of first two. Under Models 3 and 4, all covariates retain some level
of significance and only near port interacted with sea depth at port retains any effect size. These
results suggest that RSPO is associated with areas further from cities and markets. Such rural
areas are potentially more profitable if gains from production on that land dominate costs of
transportation to markets. Alternatively, RSPO could be constrained in some way that pushes
concessions away from markets. The strong effect of the interaction variable is supported by
DeFries et al. (2010) which found agricultural exports to be a leading driver of deforestation.
Slope returned a relatively large negative effect size, as expected. Contrasting with the economic
features of FSC villages, from Miteva et al. 2015, I found a larger negative coefficient for the
interaction variable; suggesting that palm oil concessions are much closer to deeper ports than
are logging concessions. Distance to a major city and distance to permanent markets returned
similar results to those from Miteva et al. (2015).
5.3 Legal and Social Higher presence of PAs is negatively correlated with RSPO presence. This relationship is
significant at the 5 percent level under Models 2 through 4. This suggests some substitution
20
because RSPO’s Principles and Criteria 5.2 requires the maintenance and/or enhancement of
rare, threatened, or endangered species and other High Conservation Value habitats, if any, that
exist within plantations (RSPO 2013). I also considered other socio-economic covariates to
explore what type of village-level conditions correlate with RSPO presence. Malnutrition is
meant to proxy for village development. Based on Lin et al. (2012), I expected RSPO presence to
correlate with less developed villages. Contrary to my expectation, malnutrition is negatively
associated with RSPO presence and the relationship is significant at the 5 percent level. This
result contrasts with positive poverty coefficients for REDD+ villages from Lin et al. (2012) and
FSC villages from Miteva et al. (2015). As expected, population density is negatively correlated
with RSPO presence and at the 1 percent level, similar to negative coefficients for REDD+
villages from Lin et al. (2012) and FSC villages from Miteva et al. (2015). Religious diversity
was meant to proxy for the degree of homogeneity in a village, which is linked to a community’s
ability to exercise collective action.
21
Table 3. Presence of RSPO concessions: Logistic and OLS results Variable Model 1 Model 2 Model 3 Model 4
Environmental Forest cover -0.054***
-0.041***
-0.000**
-0.000**
(0.007)
(0.011)
(0.000)
(0.000)
Area rice agriculture
0.000
0.000
0.000
(0.000)
(0.000)
(0.000)
Village carbon
0.025***
0.000
0.000
(0.006)
(0.000)
(0.000)
Fraction village under peat
-0.427
-0.007
-0.007
(0.619)
(0.007)
(0.007)
Economic Average elevation
-0.002
0.000
0.000
(0.008)
(0.000)
(0.000)
Average slope
-0.398**
-0.002*
-0.002*
(0.184)
(0.001)
(0.001)
Distance to major city 0.014***
0.014***
0.000**
0.000**
(0.003)
(0.005)
(0.000)
(0.000)
Distance to permanent markets
0.014***
0.000***
0.000***
(0.004)
(0.000)
(0.000)
Distance to port * depth of port
-1.961***
-0.009*
-0.009*
(0.732)
(0.005)
(0.005)
Legal Village PA protection -0.960***
-2.621**
-0.006**
-0.006**
(0.362)
(1.317)
(0.003)
(0.003)
Village PA adjacent
0.483
0.006
0.006
(0.316)
(0.004)
(0.004)
Social Population density -0.687***
-0.668***
-0.000
-0.000
(0.222)
(0.226)
(0.000)
(0.000)
Malnutrition
-1.059**
-0.005
-0.005
(0.465)
(0.003)
(0.003)
Religious diversity
0.177
0.002
0.002 (0.484) (0.005) (0.005) Constant -0.969***
-0.521
0.020***
0.020***
(0.347)
(0.574)
(0.006)
(0.006)
Observations 2,185
1,507
1,507
1,507 R-squared 0.098 0.2535 0.034 0.034 Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
22
Conclusions This paper responds to recent reviews and calls for the rigorous study of certification
programs for sustainable forest management (Romero et al. 2003; Blackman et al. 2010; Miteva
et al. 2014, 2015). I focus on the first stage of any systematic evaluation of these new
conservation ideas: identifying correlates of RSPO certification in Kalimantan. To my
knowledge, this is one of the first few studies filling an important knowledge gap regarding the
conditions that are correlated with certification. My results support findings from previous
literature that the socio-economic and institutional context within a country could be of particular
importance (Van Kooten et al. 2005; Lin et al. 2012; Sargent 2014). I add to this literature by
showing that some environmental, economic, legal, and social factors are significantly correlated
with the presence RSPO. Broadly, these findings support my hypothesis that RSPO concessions
are generally associated with suitable characteristics based on WRI’s oil palm cultivation
suitability standard. Clearly, these findings do not suggest causality.
Palm oil production is arguably the largest threat to Kalimantan’s forests, given its
outsize contribution to forest loss. If past trends continue, one result of palm oil expansion will
be to jeopardize Indonesia’s GHG emissions commitment (Austin et al. 2015). Given palm oil’s
growing contribution to deforestation and emissions, as well as Indonesia’s plans for expanded
palm oil production, consumers, firms, and policymakers must face the costs and consider the
tradeoffs. Innovative price-based policies such as this one are designed to internalize the social
and environmental spillover effects associated with palm oil production. However, we cannot
make progress on evaluating such incentives unless we can determine where such interventions
are being located (Pattanayak 2009). This would be a first step towards rigorous evaluation of
RSPO’s impact, using concepts, methods, and data such as those presented in Miteva et al.
(2015). Unfortunately, RSPO does not make its current concession shapefiles publicly available.
Further, the socio-economic indicators of RSPO presence are not readily available and rely on
ground surveys by independent third parties. For these reasons, I call for greater transparency on
part of RSPO and for broader changes that make data collection and distribution a priority, in
line with RSPO’s commitment to transparency and in the interest of improving conservation
outcomes.
23
References Abood, S.A., Lee, J.S.H., Burivalova, Z., Garcia-Ulloa, J., Koh, L.P. (2015) Relative
contributions of logging, fiber, oil palm, and mining industries to forest loss in Indonesia. Conservation Letters. 8(1):58-67.
Araujo, M., Kant, S., Couto, L. Why Brazilian companies are certifying their forests? Policy Economics. 2009; 11: 579–585. Auld, G., Gulbrandsen, L.H., McDermott, C.L. Certification schemes and the impacts on forests
and forestry. Annual Review of Environmental Resources. 2008; 33: 187–211. Austin, K.G., Kasibhatla, P.S., Urban, D.L., Stolle, F., Vincent, J. (2015) Reconciling oil palm
expansion and climate change mitigation in Kalimantan, Indonesia. PLoS ONE 10(5) Baccini, A., Goetz, S. J., Walker, W. S., Laporte, N. T., Sun, M., Sulla-Menashe, D., Hackler J.,
Beck, P.S.A., Dubayah, R., Friedl, M.A., Samanta, S., Houghton, R. A. (2012). Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nature Climate Change, 2(3), 182-185.
Blackman A, Rivera J. The evidence base for environmental and socio-economic impacts of
“sustainable certification”. 2010. Brandt JS, Nolte C, Steinberg J, Agrawal A. Foreign capital, forest change and regulatory
compliance in Congo Basin forests. Environ Res Lett. 2014; 9:44007. Burgess R, et al. (2012) The political economy of deforestation in the tropics. Q J Econ
127(4):1707–1754. Busch, Jonah, et al. "Reductions in Emissions from Deforestation from Indonesia’s Moratorium
on New Oil Palm, Timber, and Logging Concessions." Proceedings of the National Academy of Sciences 112.5 (2015): 1328-33.
Carlson KM, Curran LM, Asner GP, Pittman AM, Trigg SN, Adeney JM (2013) Carbon
emissions from forest conversion by Kalimantan oil palm plantations. Nature Climate Change 3: 283–287.
Cashore, B., Auld, G., Newsom, D., 2004. Governing Through Markets. Forest Certification and
the Emergence of Non-state Authority. Yale University Press, New Haven and London. Cerutti PO, Tacconi L, Nasi R, Lescuyer G. Legal vs. certified timber: preliminary impacts of
forest certification in Cameroon. For Policy Econ. 2011; 13: 184–190. Corley, R.H.V. "How Much Palm Oil Do We Need?" Environmental Science & Policy 12:134
39. Print.
24
Cuypers, Dieter, Theo Geerken, Leen Gorissen, Arnoud Lust, Glen Peters, Jonas Karstensen,
Sylvia Prieler, Günther Fisher, Eva Hizsnyik, and Harrij Van Velthuizen. 2013. “The Impact of EU Consumption on Deforestation: Comprehensive Analysis of the Impact of EU Consumption on Deforestation.” Technical Report 2013-063 (Final Report), European Commission, Brussels, Belgium. doi: 10.2779/822269.
DeFries, R.S., Rudel, T., Uriarte, M. & Hansen, M. (2010). Deforestation driven by urban
population growth and agricultural trade in the twenty-first century. Nature Geoscience, 3, 178-181.
de Lima ACB, Luiz A, Keppe N, Alves MC. Impact of FSC forest certification on
agroextractive communities of the state of Acre, Brazil. 2008. Ferraro, Paul J., et al. (2015). "Estimating the impacts of conservation on ecosystem services and
poverty by integrating modeling and evaluation." Proceedings of the National Academy of Sciences 112.24: 7420-7425.
Ferraro, PJ, MM Hanauer, DA Miteva, GJ Canavire-Bacarreza, SK Pattanayak, KRE Sims. 2013. More strictly protected areas are not necessarily more protective: evidence from Bolivia, Costa Rica, Indonesia, and Thailand. Environmental Research Letters 8(2): 1-7.
Ferraro, P.J., and S.K. Pattanayak. (2006). Money for nothing? A call for empirical evaluation of
biodiversity conservation investments. PLOS Biology 4(4) Ferraro PJ, Lawlor K, Mullan KL, Pattanayak SK. (2011). Forest Figures: Ecosystem Services
Valuation And Policy Evaluation in Developing Countries. Rev Environ Econ Policy; 6: 20–44.
Forest Stewardship Council. FSC Facts & Figures. September 3, 2015. Forest Council Stewardship. Global FSC certificates: type and distribution. 2014. Forest Stewardship Council. FSC principles and criteria. 2012. Gaveau DLA, Kshatriya M, Sheil D, Sloan S, Molidena E, et al. (2013) Reconciling Forest
Conservation and Logging in Indonesian Borneo. PLoS ONE 8(8) Gibbs, H.K., Brown S., Niles J.O., Foley J. Monitoring and estimating tropical forest carbon
stocks: making REDD a reality. Environ Res Lett. 2007; 2: 45023. Gibbs, H.K., Ruesch, A.S., Achard, F., et al. (2010). Tropical forests were the primary sources of
new agricultural land in the 1980s and 1990s. Proc. Natl. Acad. Sci. U.S.A., 107, 16732-16737.
Godfray HCJ, Beddington JR, Crute IR, Haddad L, Lawrence D, Muir JF, et al. (2010) Food
25
Security: The Challenge of Feeding 9 Billion People. Science 327: 812–818. Gunarso P, Hartoyo ME, Agus F (2013) Oil Palm and Land Use Change in Indonesia, Malaysia,
and Papua New Guinea. In: Killeen T, Goon J, editors. Reports from the Science Panel of the Second RSPO GHG Working Group. Kuala Lumpur, Malaysia: Roundtable for Sustainable Palm Oil.
Hansen MC, et al. (2013) High-resolution global maps of 21st-century forest cover change.
Science 342(6160):850–853. Heino M, Kummu M, Makkonen M, Mulligan M, Verburg PH, Jalava M, et al. (2015) Forest
Loss in Protected Areas and Intact Forest Landscapes: A Global Analysis. PLoS ONE 10(10)
Indonesia, the government of. Intended Nationally Determined Contribution. August 30, 2015. Joppa, L. & Pfaff, A. 2010a Re-assessing the forest impact of protection: the challenge of non
random location and corrective methods. Annu. Rev. Ecol. Econ. 1185, 135–149. Joppa, L. & Pfaff, A. 2010b. Global protected area impacts. Proceedings. Biological
Sciences/The Royal Society, 278(1712), 1633–8. Keohane, N.O., Olmstead, S.M. Markets and the Environment. Washington, DC: Island, 2007. Lin, L, SK Pattanayak, EO Sills and W Sunderlin. 2012. Site selection for forest carbon projects.
Chapter 12 in A. Angelsen, M. Brockhaus, W. Sunderlin and L. Verchot (ed.) Analyzing REDD. Center for International Forestry.
Margono B, Potapov PV, Turubanova S, Stolle F, Hansen M (2014) Primary forest cover loss in
Indonesia over 2000 – 2012. Nature Climate Change 4: 730–735. McElwee, P.D. 2008. Resource use among rural agricultural households near protected areas in
Vietnam: the social costs of conservation and implications for enforcement. Environmental Management 45:113-131.
Maulia, E. 2010. Indonesia pledges to “feed the world.” Jakarta Post. Medjibe VP, Putz FE, Romero C. Certified and uncertified logging concessions compared in
Gabon: changes in stand structure, tree species, and biomass. Environ Manage. 2013; 51: 524–540.
Miteva, D.A., Pattanayak, S.K., Ferraro, P.J. (2012) Evaluation of biodiversity policy
instruments: what works and what doesn’t? Oxford Review of Economic Policy 28(1): 69-92.
Miteva DA, Pattanayak SK, Ferraro PJ. (2014). Do biodiversity policies work? The case for
26
Conservation Evaluation 2.0. In: Helm D, Hepburn C, editors. Nature in the Balance. Oxford: Oxford University Press. pp. 251–284.
Miteva, D.A., Loucks, C.J., Pattanayak, S.K. (2015) Social and environmental impacts of forest
management certification in Indonesia. PLoS ONE 10(7) Miteva, D.A., Murray, B.C. and Pattanayak, S.K., 2015. Do protected areas reduce blue carbon
emissions? A quasi-experimental evaluation of mangroves in Indonesia. Ecological Economics, 119, pp.127-1
Muller, J., and H.J. Albers. 2004. Enforcement, payments, and development projects near
protected areas: How the market setting determines what works where. Resource and Energy Economics 26:185–204.
Nikoloyuk, J., Burns, T.R., de Man, R. (2010) The promise and limitations of partnered
governance: the case of sustainable palm oil. Corporate Governance 10(1):59-72. Oliver JGJ, Janssens-Maenhout G, Peters JAHW (2012) Trends in global CO2 emissions; 2012
Report. PBL Netherlands Environmental Assessment Agency, Institute for Environment and Sustainability of the European Commission’s Joint Research Centre.
Pattanayak, S.K., 2009. Rough guide to impact evaluation of environmental and development programs. SANDEE Working Paper 40.
Pattanayak, S.K.; S. Wunder; Fereraro, P.J. (2010). Show me the money: do payments supply
environmental services in developing countries? Review of Environmental Economics and Policy, volume 4, issue 2, pp. 254-274.
Pfaff, A., Rica, C., Rica, C., & Herrera, L. D. (2014). Governance, Location and Avoided
Deforestation from Protected Areas : Greater Restrictions Can Have Lower Impact , Due to Differences in Location. World Development, 55, 7–20.
Primm SL, Jenkins CN, Abell R, Brooks TM, Gittleman JL, Joppa LN, et al. The biodiversity of
species and their rates of extinction, distribution, and protection. Science. 2014; 344: 1246752
Putz FE, Zuidema P, Synnott T, Peña-Claros M, Pinard MA, Sheil D, et al. Sustaining
conservation values in selectively logged tropical forests: the attained and the attainable. Conserv Lett. 2012; 5: 296–303.
Rainforest Alliance. Certified forestry operation summaries for Indonesia. 2014. Available: Ramdani F., Hino M. (2013) Land Use changes and GHG Emissions from Tropical Forest
Conversion by Oil Palm Plantations in Riau Province, Indonesia. Plos ONE 8. RSPO. (2013) Principles and criteria for the production of sustainable palm oil.
27
Ruslandi, Klassen A., C. Romero, and F. E. Putz. "Forest Stewardship Council certification of
natural forest management in Indonesia: Required improvements, costs, incentives, and barriers.” Bogor, Indonesia: CIFOR. 2014.
Romero, C., Putz, F.E., Guariguata, M.R., Sills, E.O., Cerutti, P.O. and Lescuyer, G. 2013. An
overview of current knowledge about the impacts of forest management certification: A proposed framework for its evaluation. Occasional Paper 91. CIFOR, Bogor, Indonesia.
Sargent, M. (2014). “Global drivers of forest certification.” M.EM. Thesis, Duke University. Sills, E., S. Lele, T. Holmes, and S. K. Pattanayak. 2003. “Non-timber Forest Products in the
Rural Household Economy.” In Forests in a Market Economy, E. Sills and K. Abt (eds.), p. 259-282, Forestry Sciences Series, Volume 72, Dordrecht: Kluwer Academic Publishers.
Snider, Anthony G., et al. "Policy innovations for private forest management and conservation in
Costa Rica." Journal of Forestry 101.5 (2003): 18-23. Stibig, H.J., et al. "Change in tropical forest cover of Southeast Asia from 1990 to
2010." Biogeosciences 11.2 (2014): 247. Tilman D, Balzer C, Hill J, Befort BL (2011) Global food demand and the sustainable
intensification of agriculture. Proceedings of the National Academy of Sciences 108: 20260–20264.
Van Kooten, G.C., Nelson, H.W., Vertinsky, I., 2005. Certification of sustainable forest
management practices: a global perspective on why countries certify. Forest Policy and Economics 7, 857–867.
Vincent JR. The tropical timber trade and sustainable development. In: Brown K, Pearce D, editors. The causes of tropica deforestation: The economic and statistical analysis of factors giving rise to the loss of tropical forests. Vancouver: UBC Press; 1994. pp. 298–308.
United Nations. UN Commodities Trade Database. 2014 Weber J. G., Sills E. O., Bauch S., Pattanayak S. K. (2011) ‘Do ICDPs Work? An Empirical
Evaluation of Forest-based Microenterprises in the Brazilian Amazon’ Land Economics 87(4)661–81
World Resources Institute. Climate Analysis Indicators Tool (CAIT). Accessed September 21,
2015. Zagt RJ, Sheil D, Putz FE. Biodiversity conservation in certified forests: an overview.
Biodiversity conservation in certified forests. 2010.
28
Appendices
Appendix 1: Evidence of Deforestation in Kalimantan
Table 5. Evidence of Deforestation Rate of deforestation (general)
Rate of deforestation (palm oil)
Contributing sectors
Study Data Sensitive* Years Palm oil Pulp/Paper Logging
Burgess et al. (2012) MODIS (Hansen) ~ 2001 - 2008 ~ ~ ~ ~ ~
Cuypers et al. (2013) FAO
~ 1990 - 2008 1.41 Mha/yr 186,900 ha/yr 24% ~ 8%
(Indonesia) (Indonesia)
Gaveau et al. (2013)
Ministry of Forestry Yes 2000 - 2010 142,120 ha/yr 289,313 ha/yr 14% ~ 1%
(Kalimantan) (Kalimantan)
Stibig et al. (2013) FAO
~ 2000 - 2010 .82 Mha/yr ~ ~ ~ ~
(Indonesia)
Busch et al. (2014)
Ministry of Forestry No 2000 - 2010 1.14 Mha/yr 226,860 ha/yr 19.90% 12.60% 5.20%
(Indonesia) (Indonesia)
Abood et al. (2015) Greenpeace
~ 2000 - 2010 1.47 Mha/yr 161,700 ha/yr 11% 12.80% 12%
(Indonesia) (Indonesia)
*Sensitive refers to whether or not data distinguishes between natural forests and tree plantations, or includes regrowth.
29
Appendix 2: The Forest Stewardship Council Certification Program
From the late 1980s, policymakers have been motivated by the failure of previous efforts
to both curb deforestation and improve forest management. Since then, third-party actors have
introduced various certification programs with unique sets of objectives, principles, and criteria.
Initiated by non-governmental conservation organizations, the Forest Stewardship Council (FSC)
certification program is one of the largest. The program’s general goals are to promote
“environmentally appropriate,” “socially beneficial,” and “economically viable” timber
harvesting in logging concessions (FSC 2012). By the end of September 2015, a total of 1,358
certificates were active in 80 countries around the globe, spanning over 183 million ha (FSC
2015). Given the extensive reach of just one certification program, the broader intervention holds
great potential for delivering positive outcomes at scale.
Forest certification, serving as a voluntary, non-state, market-driven governance system
(Cashore et al. 2004), promises to address environmental and social spillover effects due to weak
or poorly enforced forest legal frameworks that allow the unsustainable, even if legal, use of
forests (Van Kooten et al. 2005, Auld et al. 2008). As practiced by FSC, certifications work by
creating incentives for participating timber producers to use sustainable forest management
(SFM), thereby delivering improved social and environmental outcomes. On the demand side,
the incentives include a price premium or reputational benefits for “green” forest products and
access to markets closed to producers with unsustainable practices (Auld et al. 2008, Araujo et
al. 2009). On the supply side, third-party auditors ensure producers’ compliance with law and
international agreements, tenure security and conflict resolution, recognition of indigenous
people’s property rights, community relations and workers’ rights, investments to maintain
biodiversity, the ecological productivity of the area, minimalized waste and damage to other
resources such as soil and water, enhanced forest regeneration, monitoring and assessments of
activity impacts, and maintenance of high conservation value forests (FSC 2012). Empirical
studies on the effects of certifications reveal no impact as well as modest to moderate impact.
See Table 4 for a summary on FSC programs and their impacts.
30
Table 4. Evidence of FSC impacts on deforestation, environment and other Outcomes & Impacts
Study Design Location Unit Land-use Change Environment & Other
de Blas et al. (2008) Observational Congo basin Logging concession ~ Improved RIL* practices
de Lima et al. (2008) Observational Brazil CFM** No effect on deforestation ~
Medjibe et al. (2008) Observational Gabon Samples (-) 6.3% loss of AGB*** (-) 11.8% tree damage
(-) 2.9% skid trail area
(-) 4.7% surface disturbance
Cerutti et al. (2011) Observational Gabon Logging concession (-) 18 to 34% Deforestation ~
Griscom et al. (2013) Quasi-experimental Indonesia Samples ~ (-) 50% skidding emissions
Blackman et al. (2015) Quasi-experimental Mexico Ejido**** No effect on deforestation ~
Miteva et al. (2015) Quasi-experimental Indonesia Village (-) 5% Deforestation (-) 31% air pollution
(-) 33% firewood dependence
Panlasigui et al. (2015) Observational Peru Logging concession (-) .07% Deforestation ~
Cameroon (-) .02% Deforestation ~
* RIL refers to reduced-impact logging.
** CFM refers to community forest management. The unit of analysis is a community managed forest.
*** AGB: Above-ground biomass serves as a proxy for deforestation.
**** Ejido: In Mexico, an area of communal land used for agriculture, on which community members possess individual parcels.
31
Appendix 3: Initial covariate list and definitions
Table 6. Initial covariate list and definitions.
Variable Definitions Code
Outcome Village area Area of villages, in hectares area_vil_ha Concession area non-RSPO Area of non-RSPO concessions, in hectares area_conc_nonRSPO_ha Concession area RSPO Area of RSPO concessions, in hectares area_conc_RSPO_ha Fraction village non-RSPO Fraction of village comprised of non-RSPO concessions fr_vil_conc_nonRSPO fraction village RSPO Fraction of village comprised of RSPO concessions fr_vil_conc_RSPO RSPO presence 1 if RSPO concessions are present in village RSPOpresent
Environmental Forest cover Average forest cover in village in 2008 avgforest08 Area rice field Rice area in 2008 tot_rice_area08 Area not rice field Non-rice area in 2008 non_rice_ag08 Carbon in village Total carbon per village, in metric tons tot_vil_carbon Fraction village under peat Fraction village area under peat, in 2004 fr_peat
Economic Elevation Average elevation, in meters aveelev_m Slope Average slope, in degrees aveslope_d Distance to major city Euclidean distance from the village boundary to the nearest dist2city_km
district capital or major trading centers, in kilometers
Distance to capital Euclidean distance from the village boundary to the nearest distocap_km
province capital, in kilometers
Distance to port*depth of port Euclidean distance from the village boundary to nearest dis2ports_depth
port (in km)*sea depth at the port (in km).
Distance to permanent markets Proximity to nearest permanent market in 2000, in kilometers dist2permmkt00, km
Legal Fraction village under PA Fraction of village under protected areas, using 2008 PA layer fr_pa PA protection 1 if protected by PA established after 2008 pa_post2008 PA adjacent Indicates whether a village is adjacent to a PA adj_nopa Fraction village shared owner Fraction of village under shared ownership in 2000 fr_adat00 Fraction village private owner Fraction of village under private ownership in 2000 fr_privateland
Social Government electricity 1 if village reliant on government electricity in 2008 govt_electricity_hhs08 Private electricity 1 if village reliant on private electricity in 2008 private_electricity_hhs08 Malnutrition 1 if malnutrition present in 2008 malnutrition_present08 Main street lit 1 if main street lit in 2008 main_str_lit08 Population Population in 2008 population_08 No. of Catholic churches Number of Catholic churches in a village num_catholic_church08 No. of Protestant churches Number of Protestant churches in a village num_protest_church08 No. of Mosques Number of mosques in a village num_mosques
