Content uploaded by Sebastien Desbureaux
Author content
All content in this area was uploaded by Sebastien Desbureaux on May 22, 2018
Content may be subject to copyright.
International Forestry Review Vol.XX(X), 2016 1
The ‘virtual economy’ of REDD+ projects: does private
certification of REDD+ projects ensure their environmen-
tal integrity?1
C. SEYLLER1, S. DESBUREAUX2,3, S. ONGOLO2,4, A. KARSENTY2, G. SIMONET2,5, J. FAURE6 and L. BRIMONT7
1Consultant, Rome, Italy
2CIRAD Montpellier, France
3Université d’Auvergne –CERDI Clermont-Ferrand, France
4ETH Zurich, Switzerland
5Climate Economics Chair Paris, France
6Independent Consultant, Paris, France
7IDDRI Paris, France
Email: coline.seyller@gmail.com
SUMMARY
Certification standard bodies in climate governance are assumed to function as independent third parties agencies in transactions, providing
trust and transparency to ensure that the calculation of carbon credits is reliable. This article investigates the validity of this assumption for
the voluntary forest carbon market by analysing the environmental credibility of baseline scenarios of two certified REDD+ projects, in the
Democratic Republic of Congo (the Maï Ndombe REDD+ Project) and in Madagascar (The CAZ REDD+ Project). Authors show that these
two certified REDD+ projects resemble ‘virtual emission reduction machines’ designed to inflate the production of carbon credits and that they
do not structurally change the local economy characteristics which drive deforestation. The design of both REDD+ and certification standards
business models leads almost inevitably to the decision to use a baseline scenario with high deforestation rates and to limited interventions in
the field. The need to deal with the carbon market’s price volatility and to cover the fixed costs of certification exacerbates this trend towards
inflated baselines, which also assists in the reduction of land use conflicts with local populations.
Keywords: REDD+, baselines, additionality, Democratic Republic of Congo, Madagascar
L’‘économie virtuelle’ des projets REDD+: la certification privée des projets REDD+ garantit-elle
leur intégrité environnementale?
C. SEYLLER, S. DESBUREAUX, S. ONGOLO, A. KARSENTY, G. SIMONET, J. FAURE et L. BRIMONT
En matière de gouvernance climatique, les standards de certification sont supposés agir comme un organisme tiers, gage de confiance et de
transparence pour s’assurer de la fiabilité du calcul des crédits carbone. Cet article étudie la validité de cette hypothèse dans le cadre du marché
volontaire du carbone, en analysant la crédibilité environnementale du scénario de référence de deux projets REDD+ certifiés: le projet Maï
Ndombe en République Démocratique du Congo et le projet CAZ à Madagascar. Les auteurs démontrent que ces deux projets s’apparentent
davantage à des ‘machines de réduction d’émissions virtuelles’ visant à gonfler la production de crédits carbone, sans agir structurellement sur
les caractéristiques économiques locales responsables de la déforestation. La structure des modèles économiques de REDD+ et des standards
de certification mène presque systématiquement au choix d’un scénario de référence basé sur des taux de déforestation élevés et à des interven-
tions locales limitées. La nécessité de composer avec les prix volatils du marché du carbone et de couvrir les coûts fixes liés à la certification
renforce cette tendance à gonfler les scénarios de référence, évitant également les conflits avec les populations locales pour l’usage de la terre.
La ‘economía virtual’ des los proyectos REDD+: puede la certificación privada de los proyectos
REDD+ garantizar su integridad ambiental?
C. SEYLLER, S. DESBUREAUX, S. ONGOLO, A. KARSENTY, G. SIMONET, J. FAURE y L. BRIMONT
1 The views expressed in this paper are those of the authors and do not necessarily reflect the views or policies of the organizations they work
for.
Volume 18, Number 2, June 2016, pp. 231-246(16)
2 C. Seyller et al.
countries and the role of conservation, sustainable manage-
ment of forests and enhancement of forest carbon stocks in
developing countries’ (UNFCCC 2011). Initially, REDD+
aimed at creating an international compensation mechanism
to help developing countries reduce greenhouse gas emis-
sions from their forest sector (Pistorius 2012). Finally, a
‘no-liability’ design for the REDD+ mechanism was adopted,
meaning that good environmental ‘performance’ would be
rewarded only if the deforestation rate were lower than the
baseline scenario assigned to the country.
CoP 15 in Copenhagen decided that the mechanism should
progress through three phases: (i) capacity-building, (ii)
policy implementation, including scaling up demonstration
activities and full implementation and (iii) performance-
based payments (UNFCCC 2009). In early 2015, all countries
engaged in a REDD+ process were enrolled in the capacity-
building phase (Phase 1, also referred to as the Readiness
phase). Some countries, such as the Democratic Republic
of Congo (DRC), claimed they had reached Phase 2, but no
demonstration of concrete achievement had been seen thus far
(Brockhaus and Di Gregorio 2014). Although in early 2015
REDD+ was not yet operational as a results-based payment
mechanism, some multilateral or bilateral initiatives, notably
supported by aid and cooperation agencies, had taken steps in
this direction. The best example is the bilateral agreements on
REDD+ implementation concluded between the Norwegian
Agency for Development Cooperation and a set of countries
including the governments of Brazil, Guyana and Indonesia3.
In addition to those bilateral agreements, the World Bank’s
Forest Carbon Partnership Facility (FCPF) has created a
Carbon Fund as a multilateral initiative to reward countries
and REDD+ programs at the jurisdictional level for ‘perfor-
mance’ expressed in removed or sequestered CO2. Defining
a jurisdictional level enables jurisdictions (e.g., autonomous
En lo que respecta a la gobernanza relacionada con el cambio climático, las normas de certificación en el mercado voluntario de carbono actúan
como una tercera instancia que garantiza la confianza y transparencia de los cálculos de los créditos de carbono. El presente artículo examina
la validez de esta hipótesis a través del análisis de la credibilidad de la línea de base de dos proyectos REDD+ certificados: el proyecto
Maï Ndombe en la República Democrática del Congo y el proyecto CAZ en Madagascar. Los autores argumentan que estos dos proyectos se
asemejan a ‘máquinas de reducción virtual de emisiones’ para inflar la producción de créditos de carbono sin actuar estructuralmente en las
dinámicas socio-económicas locales responsables de la deforestación. Los autores argumentan que la estructura de los modelos económicos
de los proyectos REDD+ y de normas de certificación tienden a conducir a la elección de un escenario de referencia basado en altas tasas
de deforestación y alcance limitado de las intervenciones a escala local. La necesidad de hacer frente a los precios volátiles del mercado de
carbono y de cubrir los costes fijos de la certificación refuerza esta tendencia a inflar los escenarios de referencia, lo que, adicionalmente,
también permite evitar conflictos con la población local para el uso de la tierra.
2 Permanence in REDD+ refers to the insecurity of the forest carbon stocks or the ‘backing’ of forest carbon credits obtained from avoided
deforestation and afforestation/reforestation activities. Permanence is related to the question of ensuring that gains will not be lost through
forest destruction in the short or medium term. Additionality refers to the guarantee that a REDD+ initiative will induce additional environ-
mental benefits (such as emission reduction or carbon storage) that would not have happened without this initiative. Leakage refers to the fact
that forest carbon emissions avoided in the territory under REDD+ can create or increase CO2 emissions outside this territory (Santilli et al.
(2005), Angelsen et al. (2009)).
3 In Guyana, the reference level adopted in the bilateral agreement actually allows the country to increase its deforestation rate since the
baseline scenario refers to a global deforestation level for 85 developing countries. This approach has been severely criticized as a way of
institutionalizing ‘hot air’ (Bulkan 2011); this is detailed in Section III.
INTRODUCTION: THE CONTEXT OF CARBON-
RELATED TROPICAL FORESTRY DISCUSSIONS
Although emissions from deforestation are estimated at
around 12% of global CO2 emissions (Van der Werf et al.
2009, Harris et al. 2012), tropical forests have long been mar-
ginal in international negotiations on climate change. During
the formulation of the Kyoto Protocol, the inclusion of forest
activities in the Clean Development Mechanism (CDM) was
heavily debated, especially in the two Conferences of Parties
(CoP), held in The Hague in 2000 and Marrakech in 2001.
Numerous issues emerged from the post-Kyoto climate
regime, among them how a project-based approach could
deal with the non-permanence and non-additionality of forest
carbon credits storage, while addressing the risk of leakage
and carbon market volatility2. Ultimately, only afforestation
and reforestation (AR) activities were included in the CDM,
while other forestry activities (i.e., conservation and forest
management) were excluded.
To deal with the issue of non-permanence for AR projects
in the CDM, an outcome was found through the creation of
‘temporary credits’ (Santilli et al. 2005). However, because of
their complexity, temporary credits were a failure and ended
up being used in very few projects under the CDM. Overall,
the inclusion of AR projects in the CDM scheme was not as
successful as expected; by August 2015, there were only 55
registered AR projects, a mere 0.7% of the total number
of registered CDM projects (UNEP-Risoe 2015). Forest
activities finally reappeared in the post-Kyoto debate with
the emergence of the REDD+ mechanism, adopted under the
Cancun Agreement in 2010.
The United Nations Framework Convention on Climate
Change (UNFCCC) defines REDD+ as ‘reducing emissions
from deforestation and forest degradation in developing
The ‘virtual economy’ of REDD+ projects 3
states and provinces) to set their own baselines and get
remuneration accordingly4.
Within this relatively weak international framework for
performance-based payments in the forestry sector, some
local REDD+ initiatives have emerged on the ground and
self-declared REDD+ projects are blossoming throughout
the world. According to Simonet et al. (2015), as of October
2014 there were 344 REDD+ projects in 57 countries. Of the
two countries studied in this paper, DRC was ranked first
in terms of number of REDD+ projects, with 19 REDD+
projects, while six were implemented in Madagascar. Most
REDD+ projects have been carried out by the private sector,
including Non-Governmental Organizations (NGOs) and
commercial carbon companies that see REDD+ projects as a
new source of financing for forest conservation projects or as
a commercial opportunity to capitalise on the environment.
The purpose of those REDD+ projects in the broader
REDD+ scheme is ambiguous. When REDD+ was initially
proposed as a national mechanism and after the Bali CoP,
REDD+ projects were encouraged as demonstration / pilot
projects meant to test and develop methods to tackle defores-
tation, not to produce and sell carbon credits. However,
‘project promoters’, both conservation NGOs and carbon
investors, did not want to be dependent on uncertain refunds
and remuneration by the government under a national scheme.
Therefore, they referred to REDD+ as a logo or a tag to
develop carbon credit projects, flagged as ‘REDD+ projects,’
and started to sell credits on the so-called ‘voluntary market’.
Project promoters thought the REDD+ credits would eventu-
ally be accepted on compliance markets – not only those
created by the Kyoto Protocol and regulated by UN-appointed
bodies, but also on regional and national compliance markets,
such as the EU Emissions Trading System (ETS).
To take advantage of this private project-scale dynamic,
Pedroni et al. (2009) proposed a ‘nested approach’ for REDD+
aiming at combining the national and project scales. How-
ever, this combination on different scales needs to be closely
monitored to avoid decreases in emissions in some parts of a
given country being nullified by emissions in another part.
Forest carbon credits generated by the private sector are
well represented in the voluntary carbon market. In 2012,
REDD+ projects generated 26% (or 19.3 megatons of CO2eq)
of the volume of offsets transacted on the voluntary market
(Peters-Stanley and Yin 2013). In this unregulated market, the
motivation of buyers – usually speculators, but also public
and private institutions – is often linked to brand image (cor-
porate social responsibility [CSR] and, for some, the rejection
of compliance markets imposing a fixed limit on greenhouse
gases [GHG] emissions). Contrary to the CDM, there is no
legal authority which controls and certifies carbon credits
sold on the voluntary carbon market and transactions are
generally ‘over-the-counter’. However, because of large
information asymmetry between project developers and
carbon credit buyers, the latter often demand some ‘guaran-
tees’ on the quality of the credits they buy. Some voluntary
certification schemes have emerged to add legitimacy and
reliability to the REDD+ project-based approach (see
Table 1). The Verified Carbon Standard –VCS5– is one of the
most representative of these voluntary certification schemes
(Chagas et al. 2013). In 2014, half of REDD+ projects were
certified, mainly by the VCS (Simonet et al. 2015), which
applies the same methodological principles as the CDM.
4 Jurisdictions have been proposed by the Verified Carbon Standard (VCS) as the subnational level recognized in the Cancun agreement
(CoP 16) for the baseline setting (‘as an interim measure’ according to the CoP 16 decision) and therefore, potentially, for ‘performance’
remuneration.
5 Other examples: Plan Vivo, CarbonFix Standard (acquired by The Gold Standard in September 2012).
TABLE 1 Structure of a voluntary independent third-party certification scheme
Entity Role of the entity in a (voluntary) third-party independent
environmental certification scheme Entity for this paper
Scheme owner Develops the global performance standard and related normative
and governance apparatus
VCS – develops REDD+ rules within
(or outside) the structure of Kyoto and
Cancun agreements under UNFCCC
Standards development
group
Adapts the global standard to the national (jurisdictional)
situation
Not applicable
Accreditation agency ISO or ISEAL members who accredit the competence of
conformity assessment bodies to certify compliance with the
global or national standard, with the agreement of the scheme
owner
Not applicable
Conformity assessment
(certification) body
Accredited by the accreditation agency to audit (or verify)
applicant performance against the global or national standard
and to award conditional certificates
Det Norske Veritas (DNV) in DRC
Rainforest Alliance (Madagascar)
Applicant / project
developer / certificate
holder
Applies the rules of the standard to the area to be certified,
contracts with the conformity assessment body for
performance audit (verification)
ERA Inc. and WWC in DRC
Conservation International in
Madagascar
4 C. Seyller et al.
RESEARCH QUESTION AND METHOD
In this paper, we assess whether private certification schemes
used by certification bodies, such as VCS, provide a sufficient
safeguard to ensure that the baseline scenarios adopted
in REDD+ projects are accurate. Our analyses focus on
two case studies: the Maï Ndombe project in DRC and the
Ankeniheny-Zahamena Corridor (CAZ) project in Madagascar.
The choice of these two REDD+ projects is justified by
our access to first-hand data and project documents and
should not be considered as representative. Rather this was an
arbitrary selection corresponding to opportunities we had to
get access to detailed information on REDD+ project method-
ologies used by project developers to design the baseline
scenarios of their projects. Although the analysis of the refer-
ence scenario was not the main objective of the fieldwork, we
paid particular attention to those scenarios settings as we
wanted to investigate the hypothesis of a ‘virtual economy’ of
these REDD+ projects. Our hypothesis is that beyond the
search for environmental performance, project developers
and verifying bodies have converging interests to design a
convenient baseline scenario of future deforestation in order
to increase their own income by generating and selling
as many carbon credits as possible, knowing that the actual
impact of their project on the drivers of deforestation is likely
to be limited due to factors beyond their control.
Relevant data have been collected through fieldwork con-
ducted by the authors, as well as through a literature review
and through spatial analyses using global deforestation
data compiled by Hansen et al. (2013). For the Maï Ndombe
project, most of the information comes from the project
developers’ and certifiers’ documents, and from external
information from third parties. For the CAZ project, field-
work, including face-to-face interviews and meetings, was
conducted from 2010 to 2014.
SETTING A BASELINE IN REDD+ PROJECTS
Understanding the concept of a ‘baseline’
In the REDD+ literature, ‘baseline’, ‘reference level’ or ‘busi-
ness-as-usual’ (BAU) scenarios are often used as synonyms to
describe methods for predicting future deforestation trends,
although the exact definition of each term differs in practice.
Their definitions share a common criteria: they are all one
single scenario that must be a reasonable projection of what
would happen if the REDD+ project were not implemented
(Angelsen 2008, Griscom et al. 2009, Olander and Ebeling
2011). For the sake of consistency, we will primarily use the
term ‘baseline’ in this paper. Baselines are often derived
from BAU scenarios in the sense that they are often based on
historical trends, extrapolated and applied to the future and
corrected by a set of quantitative and qualitative parameters
at the local level (deforestation drivers, economic and
geographic conditions, etc.).
The VCS validates methodologies proposed by project
promoters and accredits private consulting firms to verify and
validate the effectiveness of the purported carbon emission
reductions. Part of this certification process includes the
setting of a so-called ‘baseline scenario’. This baseline
scenario is supposed to reflect what would have happened in
a business-as-usual scenario (BAU) i.e., “How would emis-
sions from deforestation and degradation evolve without the
REDD activity” (Angelsen 2008). Additionality, referring
to the environmental benefits that would not have occurred
without the project, is directly linked to the definition of a
baseline or reference level (Pirard and Karsenty 2008,
Griscom et al. 2009). This link between additionality and
baseline scenario is the main reason why setting a baseline
scenario is crucial and yet why it encompasses many chal-
lenges, as this paper will demonstrate.
The idea of a mechanism centered on a hypothetical
scenario in a climate policy dates back to the CDM in 1997
and has fuelled many debates. The case of the energy sector
provides a good example: although predicting carbon
emissions from industrial plants was initially thought to be
possible thanks to benchmarking6, in-depth analysis of large-
scale renewable and non-renewable CDM energy projects in
India and China revealed a lack of additionality for these proj-
ects (Michaelowa and Purohit 2007, Shishlov and Bellassen
2012) leading to doubts about their environmental integrity
(Erickson et al. 2014). It is interesting to analyse the same
issue regarding REDD+, particularly as it is well known that
the pace of tropical deforestation is much more difficult
to predict than that of industrial emissions (Angelsen and
Kaimowitz 1999, Geist and Lambin 2001, Kanninen et al.
2007).
What can we learn from the study of baseline settings in
REDD+ projects? Does it sufficiently address the issues of
permanence and additionality? More importantly, can certifi-
cation standards provide a legitimate guarantee that chosen
baselines are reliable measures for predicting CO2 emissions’
reductions in the long term? Our findings from two certified
projects in DRC and Madagascar suggest the development of
a ‘virtual economy’ of emission reductions, which could be
defined as artificially generated performance via convenient
reference scenario settings and limited field interventions, to
promote development and conservation activities and to avoid
social conflicts. Such a virtual economy leads to interventions
that are more oriented toward a traditional combination of
development and conservation activities than toward innova-
tive approaches that could be specific to REDD+ as we know
it: remuneration through measurable avoided deforestation.
Thanks to these convenient reference scenario settings and
limited conditionality of development actions, fears of ‘land
and carbon grabbing’ by carbon investors to the detriment
of local communities, which some analysts believe to be
rampant in REDD+ schemes, are prevalent.
6 Technical processes for industrial production are documented and provide the basis for the comparison of projects.
The ‘virtual economy’ of REDD+ projects 5
To overcome the shortage of available data, gain statistical
power and observe future trends, baselines can be determined
using a ‘reference area’, a concept that has already been
defined and suggested to monitor leakage in CDM-AR
projects. A reference area corresponds to “a land unit used to
reflect the baseline land use without the planned activity. It
is applied to determinate the likely future land use for the
project area in a standardized way” (Dutschke, Butzengeiger
and Michaelowa, 2006, p. 96). Its size has to be “5 to 10 times
the size of the [project area]” (idem). The reference area can
thus largely differ from the actual project area. According to
the VCS guidelines, the reference area does not have to be
adjacent to the project area.
Baselines are crucial in REDD+ since the whole mecha-
nism is articulated around them. Indeed, the anticipated envi-
ronmental efficiency of projects in terms of GHG emissions
is determined by the difference between a baseline and a
projected scenario (the development path that is supposed
to occur if the project is implemented). The increase in finan-
cial earnings of the project developers is aligned with this
expected decrease in emissions: the baseline scenario will
serve as a reference against which to measure each ton of
avoided CO2eq emissions, which will then be quantified as a
carbon credit. To cover the risk of non-permanence, a few
tons of carbon credits are not accounted for (Streck and
Costenbader 2012). In the VCS, this ‘reserve’, also called
‘buffer’, represents between 10% and 40% of the credits
expected, depending on the estimated level of risk associated
to the project.
Because of their central importance, the choice of a par-
ticular baseline scenario in a REDD+ project can have major
financial consequences. The most frequent case is the risk of
an overestimation of GHG reduction emissions in which a
project will generate an excessively high number of (virtual)
carbon offsets: this tendency to induce payments based on
artificial emission reductions is called the ‘hot air’ phenome-
non (Angelsen et al. 2009, p. 314). It is tempting for project
developers to design a ‘convenient’ baseline scenario to
generate more credits in order to seek financial profit or,
as currently appears to be the most frequent case, to render a
high-cost REDD+ project financially viable.
From deforestation rate to forest carbon stocks:
the limits of prediction
Baselines generally stem from modelling. It is well known
that deforestation is ‘multicausal’; deforestation models thus
include a large set of parameters, such as demographic trends,
fallow lengths, availability of cultivable land, slopes, roads,
navigable waterways, etc. Some of these parameters are
difficult to predict, e.g., the evolution of agricultural prices,
which is a key variable to determine the opportunity cost of
forest conservation and hence, the probability of land conver-
sion. In addition, even if past research efforts can allow us
to broadly agree upon a set of parameters to include, the
complex interaction between them remains unclear (Angelsen
and Kaimowitz 1999, Geist and Lambin 2001, Kanninen
et al. 2007). When dealing with small-scale agriculture, for
instance, it is difficult to predict the ratio between shorter
fallow periods and increased population density. In addition,
although demographic growth previsions are generally fairly
robust, allowances must be made for the many factors that
influence the decision of rural youths to stay in their villages
or to migrate to the towns and cities (Tollens 2010). Some
other parameters are entirely unpredictable: conflicts (which
will entail unregulated migration), policy and institutional
changes (leading to either the abandonment of controls or,
conversely, to better law enforcement), natural disasters, etc.
In any case, defining a baseline to measure avoided defor-
estation at national and subnational levels is highly challeng-
ing because of the ever-changing environment of tropical
countries. Obtaining a credible baseline involves numerous
challenges, including ensuring the robustness of the method-
ology and limiting the temptation of project developers
to overestimate the dynamics of forest loss. This ‘quest for
credibility’ has led to the emergence of new actors, such as
third party certifiers, who are expected to guarantee a reliable
choice of baselines and better governance of the system, as
we will see in the following section.
PRIVATE GOVERNANCE IN THE REDD+ REGIME:
A BRIEF THEORETICAL BACKGROUND
Civil society and the position of an independent third
party in private governance
The literature gives several definitions of governance. This
paper uses the synthetic, comprehensive definition that
describes governance as “a system of co-production of norms
and public goods where the co-producers are different kinds
of actors” (Bartolini 2011). Governance in this paper is
related to the production of norms and public goods from a
state-centered approach, with its hierarchical, bureaucratic
state management, to a multi-actor approach that includes
non-state actors, such as civil society organizations and
private corporations. Governance here reflects a substantial
change of government with a new process for governing
society and societal issues (Rhodes 2012, Peters 2012).
This new system of co-production of societal norms by
both public and private actors is well perceived under the
REDD+ regime. Indeed, most of the projects sell carbon
credits on the voluntary carbon market, and the majority of
forest carbon credit sellers and buyers claim to be motivated
by environmental and social concerns for reducing deforesta-
tion, such as biodiversity conservation, carbon neutral goals,
climate change mitigation, and the livelihoods of forest-
dependent communities. The need for trust, transparency and
accountability between sellers and buyers in this unregulated
voluntary carbon market has been a springboard for the
emergence of carbon standards to certify the social and envi-
ronmental effectiveness of the carbon credits from REDD+
projects. Carbon standards such as the VCS can function as
a non-profit organisation (NPO) or, through the certifying
bodies, as a third-party in carbon market transactions between
suppliers and buyers by acting as a source of independent
6 C. Seyller et al.
THE ENVIRONMENTAL INTEGRITY OF CERTIFIED
REDD+ PROJECTS: TWO CASE STUDIES
The Maï Ndombe and the CAZ projects were VCS-certified
in 2012 and 2014, respectively (WWC 2012a, Conservation
International 2013), thus removing supposed doubts about
their additionality. However, a closer analysis raises serious
questions about their environmental integrity. The weaknesses
highlighted in the two projects could be linked to inadequa-
cies in the certification process. There is also the intrinsic
difficulty in predicting deforestation, notably in the context
of DRC and Madagascar that opens opportunities to design
convenient baselines.
The ERA-WWC Maï Ndombe REDD+ Project
(DR Congo)
The ERA-WWC Maï Ndombe REDD+ project8 (Figure 1)
in DRC was jointly implemented by Ecosystem Restoration
Associates Inc. and Wildlife Works Carbon Inc. (WWC), two
private companies that have been developing carbon projects
since March 20119. The project is expected to generate carbon
credits for a period of 30 years.
The ERA-WWC project is located in an area originally
planned for logging. However, in 2002 the DRC government
amended the forestry Code to make the concessions alloca-
tion process transparent and competitive and suspended,
through a moratorium, the allocation of new concessions.
However, it has been reported that despite the moratorium, the
ministry in charge of forests still allocated some concessions
up to 2005 (Debroux et al. 2007). The government then
decided to reissue all forest titles to ensure their legality.
Many titles were cancelled, including two in the project area.
In 2010 ERA made an offer to manage two former logging
concessions and turn them into conservation concessions. In
March 2011, a Memorandum of Understanding was signed
between ERA and the Ministry of Environment (MECNT) for
the allocation of 299,654 hectares for the purpose of Reduc-
ing Emissions of Deforestation and Degradation (REDD) and
Improving Forest Management (IFM). The aim was to reduce
emissions by more than 175 million tons CO2eq by 2041.
Activities undertaken were very similar to those one can find
in many conservation-development projects: participatory
mapping, local development planning, support of agriculture
intensification (onions, beans, potatoes. . .), health care
facilities, (C. Reyniers, pers. com.), etc.
In 2013, the ERA-WWC Maï Ndombe project, located
in DRC’s Bandundu province, was declared the first Congo
expertise on the environmental reliability of forest carbon
credits.
In commercial relations between the state and the private
sector, the introduction of carbon standards through NPOs
tends to inspire the confidence and trust of society at large.
NPOs involvement in the governance of societal issues is
usually associated with particular values, such as ethical
behaviour, honesty, integrity and communitarianism (Etzioni
1996). From this standpoint, the virtuous and selfless
behaviour expected of NPOs strengthens their credibility.
Private governance, trust and the issue of credibility in
forest carbon markets
In market governance, certification standards allay uncertain-
ties about the quality of goods and thus facilitate interactions
between buyers and sellers. To deal with the fact that markets
are sometimes difficult to trust, Van-Waarden (2012) suggests
that prospective buyers need to be able to collect, compare
and evaluate market information to make an informed choice
of what they really want, why, where they can get it, and how
to assess the value. In the case of carbon markets, certification
standards could solve the problem related to the lack of trust
by rigorously setting a certification process for both baselines
and carbon credits.
Since the mid-2000s, a range of private certification stan-
dards have emerged within the climate governance sphere
with growing interest in forest carbon. These standards are
primarily designed to certify carbon accounting methods and
ensure that the carbon credits issued correspond to effective
GHG emission reductions (Guigon et al. 2009).
The two case studies we analyse in the following sections
have both received VCS certification. The VCS (Verified
Carbon Standard) ranks high among the forest carbon leaders
in the market. It claims to be “an independent, non-profit
organization” (VCS 2013), aiming to “dramatically reduce
global greenhouse gas emissions”7. The ID-RECCO global
database on REDD+ projects (Simonet and Seyller 2015)
indicates that 40% of REDD+ projects certified or in the pro-
cess of certification are using the VCS standard. The VCS’s
complex technical norms could suggest that the standard
is ‘serious’ about addressing environmental effectiveness.
However, if our hypothesis of a ‘virtual economy’ governing
VCS-certified projects is confirmed, the VCS certification
regime may not be effective enough to provide a guarantee of
additionality and environmental effectiveness.
7 http://www.v-c-s.org/who-we-are/mission-history [accessed on March 2016].
8 Not to be confused with the jurisdictional Maï Ndombe Initiative (Emission Reduction Programme) developed by the World Wildlife Fund
(WWF) and Wildlife Works Carbon (WWC) on a larger area (12 million ha) and integrated into the DRC’s national strategy, although the
smaller scale Maï Ndombe project is expected to receive a share of the credits generated by the larger scale initiative. VCS received a grant
from Norway (2013–2015) to develop and pilot integrated Jurisdictional and Nested REDD+ (JNR) accounting and verification frameworks
at the national level in Costa Rica and at the sub-national levels in the States of Acre (Brazil), San Martín and Madre de Dios (Peru) and in
the future Maï Ndombe Province (DRC).
9 http://www.v-c-s.org/jnr-pilot-program. The “conservation concession” was granted initially to ERA by the DRC government. WWC and
ERA eventually set up a joint venture: ERA Congo SPRL. WWC bought the ERA shares in the joint venture in October 2013.
The ‘virtual economy’ of REDD+ projects 7
FIGURE 1 Maps of the Maï Ndombe and CAZ REDD+ Projects
Source: Authors
Basin REDD+ project to sell carbon credits10. By February
2013, the project had contributed to the reduction of an esti-
mated 2.5 million tons of CO2eq. The project was certified
using two recognized standards: the VCS for carbon (WWC
2012a) and the Climate, Community and Biodiversity
Alliance (CCBA) for social and environmental co-benefits
(WWC 2012b), a multi-criteria standard widely used in
REDD+ projects. Performance in relation to both standards
was audited by the accredited Norwegian company Det
Norske Veritas (DNV).
The Mayombe forest reference area: a dubious choice
The baseline for Maï Ndombe was determined using as a ref-
erence area the Mayombe forest in the Bas-Congo province.
Mayombe was chosen, “Primarily due to the same primary
agent of deforestation acting within both the Project and
Reference Areas – namely, planned commercial harvest – and
is also similar to the project area in terms of landscape con-
figuration, socio-economic drivers [. . .] and is furthermore
equidistant from the main market – and capital of the DRC –
Kinshasa” (WWC 2012a, p. 26). Descriptive statistics in
Table 2 and Figure 2, however, raise some doubts about the
supposed similarities between Maï Ndombe and Mayombe
forests. In terms of forest structure, Maï Ndombe is a dense
and humid forest while Mayombe is a mosaic forest. Similar
doubts can be raised about the similarity of socio-economic
deforestation drivers. Maï Ndombe is about 50% further from
Kinshasa than Mayombe (Schure et al. 2014). Mayombe is
close to major shipping harbours, which means that export
rates and access to local markets differ between the two areas.
Regarding population density, Mayombe and the rest of Bas-
Congo are characterised by high population pressure whereas
the Maï Ndombe district is sparsely inhabited.
In addition to the above-mentioned differences, at least
one more (significant) difference between the two areas
can be highlighted: the fact that the government suspended
all logging activities in the Bas Congo province in 2007,
although artisanal timber and charcoal production have
continued at a high rate, both for export and for local markets.
This ban was not applied to the area of Maï Ndombe. It is,
thus, not surprising to see large differences in deforestation
rates between the two areas. The average annual deforestation
in the Maï Ndombe project area was around 0.17% in the
2000s, close to the national deforestation rate of 0.2% since
10 They were sold to “Forest Carbon Group”, a German company that invests in carbon offsets and forestry projects.
o
n
l
i
n
e
c
o
l
o
u
r
o
n
l
y
8 C. Seyller et al.
11 The choice of the current reference level and its certification by DNV (VCS standard) has also been criticized by other actors, such as the
European Commission’s Joint Research Center (P. Mayaux, pers.com).
12 Approved VCS Methodology VM0009 Version 3.0, 6 June 2014, Sectoral Scope 1.
TABLE 2 Maï Ndombe and Mayombe forests, a basic comparison
Project Area: Maï Ndombe Reference Area: Mayombe forest
Region Maï Ndombe Lake
299,654 ha of forest
Bas-Congo Region
388,192 ha of forest
Forest type Dense & humid forest Mosaic forest – savannah
Distance to
Kinshasa
395 km, unpaved roads in poor conditions and uneasy
access
About 275 km of paved roads, with heavy traffic
Logging conditions
and goals
No direct access to the sea; for local consumption or
Kinshasa market
Access to the sea; for formal export and also export
to Angola of informal sawing
Population density Low (5–25 hab/km2)
(average in DRC = 24 hab/km2)
Relatively high (50 hab/km2)
Source: Debroux et al. 2007; WWC 2012b and CI, 2013.
FIGURE 2 Maï Ndombe and Mayombe, two different forest structures
Source: Authors.
the late 1980s (Defourny et al. 2011) but half the average
(0.34%) of the Bandundu province (FCPF 2014). The defor-
estation rate is twice as high in Mayombe (0.40%) – close to
the figure of 0.6% for the Bas Congo Province according to
UNEP. Looking beyond average rates, differences in yearly
tree losses, highlighted in Figure 3, are even more problem-
atic. Figure 3 shows that deforestation trends themselves are
not similar in the two areas: they do not vary symmetrically.
Such dissimilarities confirm that deforestation drivers in the
two areas are not the same11.
From ‘planned deforestation’ to ‘planned degradation
and unplanned deforestation’: a questionable baseline
scenario
The selected baseline scenario for the ERA-WWC Maï
Ndombe project was, in 2012, an ‘Avoided Planned Defores-
tation (APD)’ type, corresponding to the avoidance of logging
activities. The Project Design Document (PDD) states that,
“In the baseline scenario, it is expected that the primary agent
would clear 5,000 to 6,000 hectares of forest per year [.. .]
over the 25-year period of the logging concession.” (WWC
2012b, p. 31). The choice of such a scenario is surprising.
In DRC, due to high costs of transport and difficulties in
commercialising timber, the harvest intensity per hectare
is one of the lowest in the world: around 4 cubic meters
on average per hectare, less than one tree per 2 hectares
(Debroux et al. 2007). In 2014, the selected baseline changed
from ‘avoided deforestation’ to ‘avoided planned degradation
and unplanned deforestation’, following a new methodology
developed by WWC and ecoPartners and endorsed by VCS
under the name of ‘Avoided Ecosystem Conversion’12. The
idea behind the “avoided planned degradation and unplanned
deforestation” baseline is simple: “Where deforestation is
initiated by the primary agent through legally-sanctioned
commercial harvest and the area is ultimately converted to
The ‘virtual economy’ of REDD+ projects 9
Debroux et al. (2007), including the SOFORMA/BIMPE
AGRO concession. This moratorium is supposed to be main-
tained until some enabling conditions are met, inter alia a
transparent and competitive allocation process for conces-
sions, but also (even if it is implicit rather than explicit) a
zoning plan defining where new concessions can be allocated,
and the capacity to monitor and evaluate management plans
(Debroux et al. 2007). These conditions have probably not
been fulfilled, but it is difficult to predict if the government
will start allocating concessions again. Despite current specu-
lation, no one – including the present government – can
say whether it will happen or not. Therefore, the choice
of an ‘avoided planned deforestation’ scenario remains
questionable.
The Madagascar CAZ project
The CAZ project was initiated by Conservation International
(CI), an international conservation NGO, on a protected
area of 370,000 hectares demarcated in 2004 in the eastern
humid forest of Madagascar. In 2008, a carbon dimension was
added and the REDD project was launched. An Emissions
Reduction Process Agreement (ERPA) was signed between
CI and the World Bank’s BioCarbon Fund , which committed
to purchase about $1.5 million worth of certified carbon
credits corresponding to 430,000 tons of CO2eq ($3.5/t)13
avoided emissions on an area of 80,000 ha of primary forest
within the CAZ over a 30-year period. The PDD was verified
by Rainforest Alliance. Integrated Conservation Develop-
ment Programs have been implemented though the “Node
small grants” program that aims at developing alternative
non-forest by the secondary agent through unplanned
deforestation (e.g. subsistence agriculture), the baseline type
is F-P1.b.” (p. 50). This is presented as the unavoidable fate
of forest concessions in DRC. However, a comprehensive
study on deforestation drivers in DRC by Defourny et al.
(2011, p. 5) provides a very different perspective: “The pres-
ence of a forest concession and mining activities does not
seem to play a role in deforestation/degradation, at least
at the national and sub-national levels we studied (. . .) It is
essentially the size of the resident population that determines
the amount of forestland affected by deforestation and degra-
dation.” Ultimately, the loss of forest cover in DRC depends
on many drivers including commercial or illegal logging,
mining, farming and industrial agriculture. The weight of
each driver on deforestation and forest degradation may
reflect the degree of compliance with the law by logging/
mining/agricultural companies, the local context of poverty
and land tenure, and overall, the capacity of state bureaucra-
cies to implement an efficient command and control system.
It is also unclear whether or not the concession would
have been awarded to a logging company if it had not been
attributed to ERA-WWC, although the DNV validation report
specifies: “At the time of the allocation both the project and
SOFORMA (the previous concession owner) applied for the
concession rights, meaning that if the project had not been
awarded the concession, the concession would have been
allocated to a commercial logging company” (DNV 2012,
p. 15). Since the 2002 moratorium, when the government
decided to stop allocating timber concessions, the allocation
of new concessions has been illegal in DRC, although several
have been allocated despite of the moratorium according to
FIGURE 3 Differences in deforestation trends between the Maï Ndombe and Mayombe areas
Source: authors
13 The average price of REDD carbon credits was about $7.8 in 2012 (Peter-Stanley et al. 2013).
10 C. Seyller et al.
sources of revenue for locals without a conditional explicit
conservation counterpart (Brimont and Karsenty 2015).
Like the Maï Ndombe case, the CAZ project also has to
cope with the problem of a questionable reference area. This
time, the reference area includes the CAZ project in an area
that is 22 times its size (Figure 1). Figure 4 presents differ-
ences in physical features between the project and the refer-
ence area, such as elevations and slopes, population density
(as the reference area is more populated than the CAZ)
and farmers’ use of the land. Indeed, agricultural farmers
in Madagascar have two techniques to crop rice, the staple
food in Madagascar: irrigated farming on the lowlands and
slash-and-burn on the hillsides. There is one major difference
between slash-and-burn and irrigated farming: the former is a
shifting system and thus requires more land than the latter.
The result is that slash-and-burn farming is the principal cause
of deforestation in Madagascar. Figure 4 shows that farmers
in the reference area systematically practice more sedentary
irrigation farming than in the CAZ where farmers rely almost
exclusively on slash-and-burn. At the same time, the reference
area is more populated than the intervention area, and average
elevation, average slope and isolation drastically differ
between the two. Taking everything into account, there are
major differences between the CAZ project area and its
reference area.
As in Maï Ndombe, we observe significant dissimilarities
in deforestation rates between the two areas. For Ramaroson
(2012), annual deforestation in the CAZ was 0.6% between
2001 and 2005 and 0.57% between 2005 and 2009. Hansen
et al. (2013) suggest a slightly lower rate: 0.5% annually
between 2001 and 2012 (Figure 5). For the reference area,
annual deforestation was between 1% (Figure 5) and 1.26%
(CI 2013): a figure about twice as high as either estimation
for the CAZ.
This higher deforestation rate has been chosen as a base-
line scenario:
“We apply the historical average rate of 1.26% deforesta-
tion per year to the entire reference area over the project
timeline (approach ‘a’), with distinct time steps for each
5-year reporting period (CI 2013, p. 89).”
Even more disconcerting, the 1.26% baseline scenario is
hastily assumed to be the historical rate of the project area
(ndlr, while it is actually the historical rate of the reference
area) in the PDD:
“The deforestation rate inside a well-established pro-
tected area is 0.20%/yr, being an 84% reduction of the
historical deforestation rate within CAZ (1,26%/yr)”
(CI 2013, p. 129).
Without any actual change on the ground, the project could,
on paper, reduce deforestation and thus carbon emissions by
50% — this is highly unlikely and could lead to the so-called
‘hot air’ phenomenon.
DISCUSSION: BUSINESS MODELS AND
ENVIRONMENTAL INTEGRITY
The issue of larger reference areas
Conceptually, the problem of using a much larger reference
area is that it leads almost automatically to the selection of a
landscape with current deforestation rates higher than those in
the project boundaries area. Indeed, avoided deforestation
projects tend to concentrate on the most forested areas,
and the related ‘reference area’ is likely to exhibit different
characteristics, such as less forest cover, greater distance to
ports (Maï Ndombe), higher altitude and topography (CAZ),
lower population density (both cases), and other drivers
(agriculture opportunities, road proximity.. .). In some cases,
finding a reference area with the exact same characteristics
and deforestation drivers is not always possible in countries
where geographic, economic and social conditions are hetero-
geneous. Admittedly, some of the deforestation drivers of the
reference area may develop later in the project area in a BAU
scenario, but obviously not all of them. Most importantly,
although some of the aforementioned drivers will affect the
project area in the future, it is not really possible to predict the
timeframe of the evolution of the deforestation. Without
any possibility of verification (the reference scenario will not
happen if the REDD+ project is implemented), this situation
opens large avenues for designing ‘convenient scenarios’.
The issue of baseline scenarios
The case studies and empirical evidence from DRC and
Madagascar analysed in this paper confirm that the baseline
scenarios in REDD+ projects amount to untestable guesses14,
even when applying the certification bodies’ best practices
guidelines. Using ‘reference areas’ much larger than, and
sometimes remote from, the project area (therefore with dif-
ferent characteristics and history) as a predictor for the future
deforestation of the project area makes it possible to increase
the deforestation rate used for setting the reference scenario.
The value of forecasting deforestation trends using uncertain
drivers is also questionable. External unplanned stressors can
arise particularly in ‘fragile States’ (Karsenty and Ongolo
2012) where most REDD+ projects are located. In sum,
several scenarios could be considered for certification, all of
them being legitimate under a given forecast hypothesis (see
Box 1). One example of these unplanned stressors is political
instability, well-illustrated in the CAZ project. In the CAZ
PDD, there is a statement that, if taken seriously, could
trouble any reader concerned with environmental integrity:
“Past experience shows that deforestation in Madagascar
decreases during periods of strong regulation, and then
suddenly increases as soon as the regulations are no
longer enforced. Deforestation is also widely associated
with political uncertainty” (CI 2013, p. 80).
14 They are “non falsifiable” in the Popperian sense, i.e., they cannot be taken as scientific as they cannot be disproved.
The ‘virtual economy’ of REDD+ projects 11
FIGURE 4 Comparison between the CAZ and its reference area, using QQ-plots of deforestation drivers
Source: authors
Madagascar is a politically unstable country which has expe-
rienced many political crises over the past decades, including
a severe one from 2009 to 2014. Predicting deforestation
in such a context is simply illusory, as implicitly admitted
by the authors of the PDD. Nevertheless, a baseline scenario
has been certified and the Carbon Fund of the FCPF will
buy carbon credits, with the condition that these credits are
certified by the VCS.
A narrow vision of ‘leakage’
The baseline issue is not the only weakness of REDD+
projects. Leakage is also a problem which has, thus far, been
poorly addressed. Project developers and certifying organiza-
tions tend to adopt a narrow view of leakage, figuring that the
pressure can be moved outward from the project area, e.g.,
around 2 km outside in the case of the Malagasy CAZ project.
However, in this era of globalization and heavy commercial
pressure on lands, ‘leakage’ is more likely to occur because
investments (notably foreign ones) that could have been made
in the REDD+ project area are ultimately located elsewhere.
Countries often encourage the relocation of investments after
they have, on the one hand, encouraged REDD+ projects in
remote areas that are not very suitable for agriculture and, on
the other hand, encouraged agribusiness investors to develop
o
n
l
i
n
e
c
o
l
o
u
r
o
n
l
y
12 C. Seyller et al.
FIGURE 5 Deforestation rates in the CAZ and its reference area
Source: authors. The level of 75% of tree cover is the threshold that allows to replicate official forest maps in Madagascar for the eastern
eco-region.
Box 1. Predicting deforestation rates in the Congo Basin: diverging views among experts
To anticipate deforestation rates, modellers often consider the impact of shifting agriculture and the potential development of
agribusiness.To anticipate the impact of shifting agriculture, the projects hypothesise population growth rates in the rural area,
annual land areas cleared by households, fallow periods, road networks (existing and planned), river access, etc. They also use
soil maps and data on slopes and distance to urban and rural agglomerations. For agribusiness, modellers tend to produce maps
of the economic potential of lands for a variety of crops (and husbandry); they also consider agricultural development plans at
the national and provincial scale and investment trends at the global scale for supplying world demand. Given growing global
demand and population figures, the natural tendency is to extend existing trends for small-scale agriculture and expect large-
scale investments in agribusiness. However, experts do not all agree on this forecasting method. For the Congo Basin, a study
by Zhang et al. (2002), only considered the dynamics of shifting agriculture and forecasted a deforestation rate of 1.20% in
2030 (against the actual 0.30% rate today). They adopted the following hypotheses (i) Population growth above 3% in
rural areas, (ii) No rural migration to urban areas, (iii) Shifting cultivation with 15 year fallows and population densities of
25 inhabitants per km2.
Eric Tollens (2010), a senior agronomist with extensive knowledge of the DRC and the Congo Basin, considered these
forecasts unrealistic. Tollens felt that (i) with the increases in population density, fallow time would be shortened and finally
fade out completely, (ii) with declining soil fertility and crop yields, young people would out-migrate to the cities, and the rural
population density would level at around 6–10 persons/km2. For Tollens, the deforestation rate stemming from small-scale
farming would likely remain the same in the coming years. Tollens was also sceptical about the largely-held belief of the
imminent, massive development of industrial plantations (notably palm oil) in the Congo basin and especially in DRC—
expressed, for instance, in a 2009 McKinsey report on the DRC that was endorsed by the government (MECNT 2009) and used
in preparing the REDD+ strategy. He pointed to the following difficulties: (i) The bad investment climate and the very weak
state of infrastructures are limiting factors that result in lack of competitiveness for these productions in most of the Congo
Basin countries, except perhaps Cameroon, (ii) The land tenure issue is a compounding difficulty for developing large-scale
plantations, (iii) The ‘Dutch disease’ phenomenon (growth under the dependence of extractive industries) will lead to an
increase of food imports and inflation, thus hampering competitiveness.
The ‘virtual economy’ of REDD+ projects 13
palm oil plantations on non-protected (or degazetted) forest-
lands in the forest frontier area.
A business model favouring the convergence of interests
In the VCS system, a project approved by a ‘Validation/Veri-
fication Body’ (VVB, i.e., private auditors) is automatically
certified without an external review. Auditors operate under
strong pressure from project developers: “There is pressure
on auditors to approve their clients’ methodologies in order
to maintain a good relationship and not compromise future
work opportunities. As has been shown in the CDM (. ..) this
design flaw in carbon markets is difficult to address as long
as the project developer pays for and can choose the auditor.”
(Kollmuss et al. 2008, p. 62). This potential conflict of inter-
ests in the ‘business-to-business’ model is not specific to
carbon project certification. It has often been pointed out with
respect to auditing in general (Goldman and Barlev 1974,
Moore et al. 2006). However, with REDD+ projects there is a
kind of irreducible uncertainty regarding what the ‘right refer-
ence scenario’ should be. Our case studies show that only
small differences in baseline scenarios – whether designed
intentionally or not – can have severe financial (positive for
business actors) and environmental (negative for the climate)
consequences. The interest of the project developers is obvi-
ous: as the market price of carbon credits falls, the financial
viability of a project (that relies on the carbon market for
financing) declines15. ‘Optimizing’ the parameters, notably
those related to baseline settings, seems to be the only way
to maintain the viability of a project’s business model. It is
important to remember that the financial health of the VCS
depends on the number of carbon credits it certifies: in
Madagascar, according to the CAZ Project Design Document
(PDD), the VCS charges USD 0.10 per certified VCU
(Verified Carbon Unit) while the CCBA charges USD 0.3. In
addition, exerting higher demands than other certifiers could
harm the VCU’s business model.
CONCLUSION: ARE REDD+ PROJECTS A THREAT
OF ‘CARBON LAND GRABBING’ OR SYMBOLS OF
THE VIRTUAL ECONOMY OF CARBON PROJECTS?
REDD+ projects tend to be more professional than they were
at the beginning, when several cases of ‘carbon cow-boys’
were reported. These so-called ‘cow-boys’ negotiated with
local communities to “buy their carbon” and attempted to sell
back these self-declared ‘offsets’ to companies. The profes-
sionalization of REDD+ projects has been achieved thanks
to a certification process that requires highly specialized
and skilled private expertise for applying the ‘methodologies’
required by the certification organizations and verified by a
mutually agreed upon body. REDD+ certification procedures
are supposed to ensure the quality of the carbon credits that
the projects wish to sell in order to cover their costs and, if
possible, generate benefits. The process of establishing ‘meth-
odologies’ was inspired by the UN’s CDM and its regulating
body, the CDM Executive Board. The baseline/additionality
issue, including the problem of ‘inflated’ baselines, is not
specific to REDD+ projects. It also affects the CDM and
individual countries as seen in Guyana’s efforts to formulate
‘inflated’ baselines in 2008 (Karsenty et al. 2014). The
examination of long technical project documents brings out
evidence of the decision to deliberately select the hypotheses
that will most probably allow the project to be financially
sustainable, even (or especially) if carbon credit prices are
low.
We want to avoid the term ‘manipulation’, since there is
a large range of possible future trends of deforestation in
each specific area given the ‘multicausal’ dimension of the
phenomenon, the complexity of deforestation drivers and the
frequent divergence of experts’ views. When purchasing
carbon credits an intangible good, the buyer is never in a posi-
tion to determine its quality and has to rely on information
from the project developer or the verifier. However, we must
not forget the convergence of interests between the project
promoters, the verifying body and the certification organiza-
tion. It is in the interest of all of these stakeholders to select
the variables that support the prediction of a high scenario of
deforestation in order to generate numerous credits – the basis
of remuneration for all of these stakeholders. Actually, the
convergence of interest is likely to go beyond this first circle
of actors. On the voluntary market, corporations buying
REDD+ credits to fulfil commitments they have adopted
for reasons of ‘social and environmental responsibility’ have
neither the will nor the expertise to verify the ‘quality’ of the
credits they purchase; instead they count on the private
labelling system. It is difficult to expect the market to regulate
itself in the absence of a financially-neutral regulatory body.
In the case of a governmental or intergovernmental body like
the CDM, verifiers are somewhat limited by the risk that the
CDM Executive Board might revoke their license, which it
has temporarily done for all major verifiers involved in the
CDM16. The oversight capacity of the CDM is limited.
The structural problems are similar in the case of voluntary
standards; in fact they are likely to be worse as the oversight
capacity of the voluntary standard bodies is even more limited
and as voluntary standards are financially dependent on the
volume of credits they verify.
Our research shows that some promoters of REDD+
projects, including conservation NGOs, consider the REDD+
scheme as a financial opportunity to fund their usual activities
in biodiversity and natural habitat conservation. Thus, focus-
ing on carbon credits becomes a convenient way to overcome
the shortfall of funds for forest conservation. Even if the
creation of new protected areas may have negative social
15 According to a DRC official, of 2.5 million tons of CO2eq validated by VCS for the 2011–2012 period, only 700,000 tons were sold. This
convinced the project to stop certification operations in 2013.
16 See http://uk.reuters.com/article/2010/03/26/us-carbon-un-suspensions-idUKTRE62P5E420100326 (accessed 26 January 2015).
14 C. Seyller et al.
effects, the positive environmental effects should not be over-
looked as they contribute to the mitigation of global climate
change. However, promoters of REDD+ projects seem to
place their interests for new forest conservation funding
opportunities above their compliance with basic requirements
ensuring the environmental effectiveness of their projects.
The issues of additionality, risks of leakage and the non-
permanence of emission reductions, which are the conditions
for the global and long-term environmental benefits of
REDD+ projects, are often neglected or simply ignored.
Finally, one of the likely impacts of private governance of
projects through certification might well be the precautions
taken vis-à-vis local populations. As observed by the Forest
Stewardship Council (FSC) for forest management certifica-
tion, interaction between project developers and FSC-
certified concessions seems to have a positive effect on the
social conditions of workers and local populations (Cerutti
et al. 2014). Negative social impacts are much easier to assess
than positive impacts, such as the sustainability of manage-
ment practices, whose assessment would require two or three
felling cycles (a felling cycle is 25–30 years or more) (Karsenty
and Gourlet-Fleury 2006). Furthermore, certified companies
are afraid of negative campaigns waged by NGOs. For
REDD+ projects, VCS certification is often associated with
the CCBA17, which focuses on the social dimension, an
important criteria to satisfy possible buyer concerns. Some
analysts warn that REDD initiatives and projects might lead
to ‘carbon land grabbing’ (RRI 2014) that can dispossess
local communities and indigenous people of their land use
rights. Undeniably, a certain number of REDD+ projects have
required the creation of new protected areas, thereby impos-
ing restrictions on land use, restrictions which depend on land
category as determined by IUCN. In reality, it appears much
more convenient for REDD+ project promoters to design a
baseline that produces virtual ‘emission reductions’ rather
than to spend a lot of time and money preventing small farm-
ers from practicing slash-and-burn, which, moreover, could
trigger conflicts that could jeopardize their certification. This
is what we call the ‘virtual economy’ of REDD+ projects.
REFERENCES
ANGELSEN, A. 2008. How do we set the Reference Levels
for REDD Payments? In ANGELSEN, A. (ed.), Moving
Ahead with REDD: Issues, Options and Implications.
Bogor: CIFOR. pp. 53–64.
ANGELSEN, A. and KAIMOWITZ, D. 1999. Rethinking
the Causes of Deforestation: Lessons from Economic
Models. World Bank Research Observer 14(1): 73–98.
ANGELSEN, A., BROCKHAUS, M., KANNINEN, M.,
SILLS, E. and SUNDERLIN, W. 2009. Realising REDD+:
National Strategy and Policy Options. Bogor: CIFOR.
BARTOLINI, S. 2011. New Modes of Governance: An Intro-
duction. In A. HERITIER and M. RHODES (eds.),
New Modes of Governance in Europe: Governing in the
Shadow of Hierarchy. Houndmills: Palgrave Macmillan.
pp. 1–18.
BRIMONT, L. and KARSENTY, A. 2015. Between Incen-
tives and Coercion: the Thwarted Implementation of
PES Schemes in Madagascar’s Dense Forests. Ecosystem
Services 14(2015): 113–121.
BROCKHAUS, M. and M. DI GREGORIO. 2014. National
REDD+ Policy Networks: From Cooperation to Conflict.
Ecology and Society 19(4): 14.
BULKAN, J. 2011. Challenges for Norwegian Financial
Support to Guyana. CFA Newsletter.
CERUTTI, P.O., LESCUYER, G., TSANGA, R., KASSA,
S.N., MAPANGOU, P.R., MENDOULA, E.E., …
YEMBE R.Y. 2014. Impacts Sociaux de la Certification
du Forest Stewardship Council: Évaluation dans le Bassin
du Congo (Occasional Paper). Bogor: CIFOR.
CHAGAS, T., COSTENBADER, J., STRECK, C. and ROE,
S. 2013. Reference Levels: Concepts, Functions, and
Application in REDD+ and Forest Carbon Standards
(January, 2013), Amsterdam: Climate Focus.
CONSERVATION INTERNATIONAL [CI]. 2013. Reduced
Emissions from Deforestation in the Ankeniheny-Zahamena
Corridor, Madagascar (Version 03). Published by Con-
servation International and Verified Carbon Standard.
DEBROUX, L., HART, T., KAIMOWITZ, D., KARSENTY,
A. and TOPA, G. 2007. Forests in Post-Conflict Demo-
cratic Republic of Congo: Analysis of a Priority Agenda.
Bogor: World Bank/Cirad and Cifor.
DEFOURNY, P., DELHAGE, C. and KIBAMBE J.-P. 2011.
Analyse Quantitative des Causes de la Déforestation et de
la Dégradation des Forêts en République Démocratique
du Congo, Louvain: Université Catholique de Louvain.
DNV 2012. Validation Report for The Maï Ndombe REDD+
Project. Published by Det Norske Veritas Inc.
DUTSCHKE, M., BUTZENGEIGER, S. and MI-
CHAELOWA, A. 2006. A Spatial Approach to Baseline
and Leakage in CDM Forest Carbon Sinks Projects.
Climate Policy 5(5): 517–530.
ERICKSON, P., LAZARUS, M. and R. SPALDING-
FECHER. 2014. Net Climate Change Mitigation of the
Clean Development Mechanism. Energy Policy 72(2014):
146–154.
ETZIONI, A. 1996. A Moderate Communitarian Proposal.
Political Theory 24(2): 155–171.
FOREST CARBON PARTNERSHIP FACILITY [FCPF].
2014. Emissions Reduction Program Idea Note (ER-PIN)
of the REDD+ Maï Ndombe Project (version of 04 April,
2014), Democratic Republic of Congo: Forest Carbon
Partnership Facility, World Bank (FCPF).
17 Using the ID-RECCO database (Simonet and Seyller 2015), we find that 55% of the REDD+ projects certified by the VCS are also certified
by the CCBA. Moreover, 83% of the avoided deforestation projects certified by the VCS are also certified by the CCBA.
The ‘virtual economy’ of REDD+ projects 15
GEIST, H.J. and LAMBIN, E.F. 2001. What Drives Tropical
Deforestation? A Meta-analysis of Proximate and Under-
lying Causes of Deforestation Based on Subnational Case
Study Evidence (LUCC Report Series No. 4). Louvain-la-
Neuve: LUCC International Project Office, University of
Louvain.
GOLDMAN A. and BARLEV B. 1974. The Auditor-Firm
Conflict of Interests: Its Implications for Independence,
The Accounting Review 49(4): 707–718.
GRISCOM, B., SHOCH, D., STANLEY, B., CORTEZA, R.
and VIRGILIO, N. 2009. Sensitivity of Amounts and
Distribution of Tropical Forest Carbon Credits Depending
on Baseline Rules. Environmental Science & Policy 12(7):
897–911.
GUIGON P., BELLASSEN, V. and AMBROSI, P. 2009.
Voluntary Carbon Markets: What the Standards Say....
(Mission Climat working paper no. 2009-04). Paris:
Mission Climat, Caisse des Dépots.
HANSEN, M.C., POTAPOV, P.V, MOORE, R., HANCHER,
M., TURUBANOVA, S.A., TYUKAVINA, A., (...),
TOWNSHEND, R.G. 2013. High-Resolution Global
Maps of 21st-Century Forest Cover Change. Science
342(6160): 850–853.
HARRIS, N. L., BROWN, S., HAGEN, S. C., SAATCHI, S.
S., PETROVA, S., SALAS, W. (...) and LOTSCH, A.
2012. Baseline Map of Carbon Emissions from Deforesta-
tion in Tropical Regions. Science 336(6088): 1573–1576.
KANNINEN, M., MURDIYARSO, D., SEYMOUR, F.,
ANGELSEN, A., WUNDER, S. and GERMAN, L. 2007.
Do Trees Grow on Money? The Implications of Deforesta-
tion Research for Policies to Promote REDD, Forest
Perspectives 4. Bogor: CIFOR.
KARSENTY, A. and GOURLET-FLEURY, S. 2006. Assess-
ing Sustainability of Logging Practices in the Congo
Basin’s Managed Forests: the Issue of Commercial
Species Recovery. Ecology & Society 11(1): 26. http://
www.ecologyandsociety.org/vol11/iss1/art26/
KARSENTY, A. and ONGOLO, S. 2012. Can ‘fragile states’
decide to reduce their deforestation? The inappropriate
use of the theory of incentives with respect to the REDD
mechanism. Forest Policy and Economics 18 (May 2012),
38–45.
KARSENTY, A., VOGEL, A. and CASTELL, F. 2014.
Carbon rights, REDD+ and payments for environmental
services. Environment Science & Policy 35: 20–29.
KOLLMUSS, A., ZINK, H. and POLYCARP, C. 2008. Mak-
ing Sense of the Voluntary Carbon Market: A Comparison
of Carbon Offset Standards, WWF Germany.
MECNT. 2009. Potentiel REDD+ de la RDC, Kinshasa,
DRC: Ministère de l’Environnement, de la Conservation
de la Nature et du Tourisme.
MICHAELOWA, A. and PUROHIT, P. 2007. Additionality
Determination of Indian CDM Projects: Can Indian CDM
Project Developers Outwit the CDM Executive Board?
(Discussion Paper CDM-1). Zürich: University of Zurich,
Institute for Political Science.
MOORE, D.A., TETLOCK, P.E., TANLU, L. and BAZER-
MAN, M.H. 2006. Conflicts of Interest and the Case
of Auditor Independence: Moral Seduction and Strategic
Issue Cycling. Academy of Management Review 31(1):
10–26.
OLANDER, J. and EBELING, J. 2011. Building Forest
Carbon Projects: Step-by-Step Overview and Guide (ver-
sion 2.0). Washington, DC: Forest Trends/EcoDecisión.
PEDRONI, L., DUTSCHKE, M., STRECK C. and ESTRA-
DA PORRUA, M. 2009. Creating Incentives for Avoiding
Further Deforestation: The Nested Approach. Climate
Policy 9: 207–220
PETERS, G. 2012. Governance as political theory. In D.
LEVI-FAUR (ed.), The Oxford Handbook of Governance.
NewYork: Oxford University Press: 19–32.
PETERS-STANLEY, M. and YIN, D. 2013. Maneuvering
the Mosaic. State of the Voluntary Carbon Markets 2013.
Washington, DC, 126: A Report by Forest Trends’ Ecosys-
tem Marketplace & Bloomberg New Energy Finance.
PIRARD, R. and KARSENTY, A. 2008. Climate Change
Mitigation: Should Avoided Deforestation (REDD) Be
Rewarded? Journal of Sustainable Forestry 28(3): 434–
455.
PISTORIUS, T. 2012. From RED to REDD+: the Evolution
of a Forest-based Mitigation Approach for Developing
Countries. Current Opinion in Environmental Sustainabi-
lity 4(6): 638–645.
RAMAROSON, N. 2012. Analyse Historique de la Défores-
tation par Télédétection et Modélisation de la Déforesta-
tion à Madagascar: Cas du Corridor Ankeniheny-Zahamena,
Université de la Réunion.
RHODES, R. A. W. 2012. Waves of Governance. In D. Levi-
Faur (ed.), The Oxford Handbook of Governance. New
York: Oxford University Press: 33–48.
RIGHTS AND RESOURCES INITIATIVE [RRI] 2014. Sta-
tus of Forest Carbon Rights and Implications for Com-
munities, the Carbon Trade, and REDD+ Investments.
http://www.rightsandresources.org/documents/files/
doc_6594.pdf
SANTILLI, M., MOUTINHO, P., SCHWARTZMAN, S.,
NEPSTAD, D., CURRAN, L. and NOBRE, C. 2005.
Tropical Deforestation and the Kyoto Protocol. Climate
Change 71(3): 267–276.
SCHURE, J., LEVANG, P. and WIERSUM, K.F. 2014.
Producing Woodfuel for Urban Centers in the Democratic
Republic of Congo: A Path Out of Poverty for Rural
Households? World Development 64(1): 80–90.
SHISHLOV, I. and BELLASSEN, V. 2012. Dix Enseigne-
ments pour les Dix Ans du MDP (Étude Climat n°37/
October 2012). Paris: CdC, Climat Recherche.
SIMONET, G. and SEYLLER, C. 2015. ID-RECCO, a New
Collaborative Work Tool to Improve Knowledge on
REDD+ Projects, Climate Economics Chair Working
Paper n° 2015-08.
SIMONET, G., KARSENTY, A., DE PERTHUIS, C., NEW-
TON, P. and SCHAAP, B. 2015. REDD+ projects in 2014:
an overview based on a new database and typology,
Cahiers de la Chaire Economie du Climat – Information
and Debates Series 32. Working paper.
STRECK, C. and COSTENBADER, J. 2012. Standards for
Result-based REDD Finance: Overview and Design
Parameters (November 2012). Amsterdam: ClimateFocus.
16 C. Seyller et al.
TOLLENS, E. 2010. Potential Impacts of Agriculture Devel-
opment on the Forest Cover in the Congo Basin. Washint-
gon, DC: The World Bank.
UNEP-RISOE. 2015. CDM Pipeline Overview. Retrieved
from http://cdmpipeline.org/ (last accessed in March
2016).
UNFCCC. 2009. Copenhagen Accord. Document: FCCC/
CP/2009/L.7, 18 December 2009, from the United Nations
Climate Change Conference.
UNFCCC 2011. Cancun Accord. Document: FCCC/
CP/2010/7/Add.1, 15 March 2011, from the United
Nations Climate Change Conference.
VAN DER WERF, G.R., MORTON, D.C., DEFRIES, R.S.,
OLIVIER, J.G., KASIBHATLA, P.S., JACKSON, R.B.,
... and RANDERSON, J.T. 2009. CO2 Emissions from
Forest Loss. Nature Geoscience 2(11): 737–738.
VAN-WAARDEN, F. 2012. The Governance of Markets: on
Generating Trust in Transactions. In D. LEVI-FAUR (ed.),
The Oxford Handbook of Governance. New York: Oxford
University Press: 355–37.
VERIFIED CARBON STANDARD [VCS]. 2013. VCS
Missions. Retrieved from http://www.v-c-s.org/who-we-
are/mission-history, accessed on 24 April 2014.
WWC. 2012a. The Maï Ndombe REDD+ Project, Democratic
Republic of Congo, DRC: A joint project of ERA & Wild-
life Works (version 1.63). Published by Wildlife Works
Carbon and Verified Carbon Standard.
WWC. 2012b. Maï Ndombe REDD+: A Joint Project of ERA
& Wildlife Works Project Design Document, Validated
to the Climate, Community, and Biodiversity Standards
(2nd edition). Published by Wildlife Works Carbon and
Climate, Community, and Biodiversity Standards.
ZHANG, Q., JUSTICE, C.O. and DESANKER, P.V. 2002.
Impacts of Simulated Shifting Cultivation on Deforesta-
tion and the Carbon Stocks of the Forests of Central
Africa. Agriculture, ecosystems & environment 90(2):
203–209.
