Costs Benefits of Reducing Carbon Emissions

Description
This report presents a conceptual framework for estimating the costs to tropical nations of implementing REDD programs and applies this framework to the Brazilian Amazon region.

Reducing
Emissions from
Deforestation & Forest
Degradation
REDD
REDD
©
Te Woods Hole Research Center
United Nations Framework Convention on Climate Change (UNFCCC)
Conference of the Parties (COP), Tirteenth session
3-14 December 2007
Bali, Indonesia
Reducing Emissions from Deforestation and Forest Degradation
T
H
E

C
O
S
T
S

A
N
D

B
E
N
E
F
I
T
S

O
F

R
E
D
U
C
I
N
G

C
A
R
B
O
N

E
M
I
S
S
I
O
N
S

F
R
O
M

D
E
F
O
R
E
S
T
A
T
I
O
N

A
N
D

F
O
R
E
S
T

D
E
G
R
A
D
A
T
I
O
N

I
N

T
H
E

B
R
A
Z
I
L
I
A
N

A
M
A
Z
O
N
For more information, please contact
Daniel Nepstad
Senior Scientist
Te Woods Hole Research Center
149 Woods Hole Road
Falmouth, MA 02540
USA
508 540 9900, x131
[email protected]
Te Woods Hole Research Center uses recycled paper and vegetable-based inks.
“Te designations employed and the presentation of the material in this publication,
including delimitation of borders and frontiers, do not imply the expression of any opinion
whatsoever on the part of those organizations supporting this research. Te sponsors and
supporters of this research do not warrant that the information contained in this publica-
tion is complete and correct, and shall not be liable for any damages incurred as a result of
its use. Te views expressed by the authors or editors do not necessarily represent the deci-
sions or the stated policy of the sponsors and supporters of this research.”
Our thanks to Karen Schwalbe, Paul Lefebvre,
Tracy Johns, Wendy Kingerlee, Claudia Stickler, David
McGrath, and Richard Houghton in reviewing, editing,
and preparing this report.
Cover and report design by
Michael Ernst and Elizabeth Braun.
©
Te Woods Hole Research Center
149 Woods Hole Road
Falmouth, MA 02540
USA
whrc.org
THE COSTS AND BENEFITS OF REDUCING CARBON EMISSIONS FROM
DEFORESTATION AND FOREST DEGRADATION IN THE BRAZILIAN AMAZON
Authors:
Daniel Nepstad (WHRC, IPAM), Britaldo Soares-Filho (Univ. Federal Minas Gerais),
Frank Merry (WHRC), Paulo Moutinho (IPAM, WHRC),
Hermann Oliveira Rodrigues (UFMG), Maria Bowman (WHRC)
Steve Schwartzman (ED), Oriana Almeida (Univ. Federal Para), Sergio Rivero (UFPa),
Sponsors of this report:
Te Linden Trust for Conservation,
Joseph H. Gleberman,
Roger and Victoria Sant,
and the William and Flora Hewlett Foundation
Support for the research presented in this report:
Te Gordon and Betty Moore Foundation,
Te David and Lucile Packard Foundation
US National Science Foundation,
US Agency for International Development,
Blue Moon Fund
and the Global Opportunities Fund (GOF)
T~niv ov Cox:vx:s
¡xccutivc Summary

!ntroduction
Can dclorcstation in thc 8razilian Amazon bc rcduccd to zcro:
Prcmiscs
Tc conccptual lramcwork ol a 8razilian Amazon R¡ÐÐ program
A spatial map ol opportunity costs
Tc Public Forcst Stcwardship Fund
Tc Privatc Forcst Stcwardship Fund
Tc Govcrnmcnt Fund
Tc costs ol R¡ÐÐ in thc 8razilian Amazon ovcr 30 ycars
Cobcnc?ts ol R¡ÐÐ
How will it work:

Conclusion

Litcraturc
Lis: ov T~nivs ~xu Ficuvvs
Figurc 1. Tc 8razilian Amazon in 2030
Figurc 2. Tc lorcsts in thc 8razilian Amazon
Figurc 3. Potcntial nct prcscnt valuc ol soy production
Figurc 4. Potcntial nct prcscnt valuc ol cattlc production
Tablc 1: Òpportunity Costs ol lorcst maintcnancc
Tablc 2. Carbon stocks and opportunity cost
Figurc 5. Potcntial nct prcscnt valuc ol sustainablc timbcr production
Figurc 6. Forcst carbon stocks ol thc 8razilian Amazon
Figurc 7. Nct opportunity cost ol lorcst protcction in thc 8razilian Amazon
Figurc 8. Marginal opportunity cost ol rcductions in carbon cmissions
Figurc 9. Tirtyycar trajcctory ol an illustrativc, hypothctical R¡ÐÐ program.
Tablc 3. Summary ol costs ol 8razilian Amazon R¡ÐÐ program in Ycar 10
Figurc 10. ¡xamplc ol thc cstimatcd costs
¡xccutivc Summary 1
!ntroduction 6
Can dclorcstation in thc 8razilian Amazon bc rcduccd to zcro: 7
Prcmiscs 8
Tc conccptual lramcwork ol a 8razilian Amazon R¡ÐÐ program 9
A spatial map ol opportunity costs 10
A dclorcstation rcduction schcdulc and lorcst allocation 14
Tc Public Forcst Stcwardship Fund 15
Tc Privatc Forcst Stcwardship Fund 19
Tc Govcrnmcnt Fund 20
Tc costs ol R¡ÐÐ in thc 8razilian Amazon ovcr 30 ycars 20
Cobcnc?ts ol R¡ÐÐ 21
How will it work: 21
Conclusion 22
Footnotcs 23
Litcraturc 24
List ol Tablcs and Figurcs
Figurc 1. Tc 8razilian Amazon in 2030 6
Figurc 2. Tc lorcsts in thc 8razilian Amazon 11
Figurc 3. Potcntial nct prcscnt valuc ol soy production 12
Figurc 4. Potcntial nct prcscnt valuc ol cattlc production 13
Tablc 1: Òpportunity Costs ol lorcst maintcnancc 13
Tablc 2. Carbon stocks and opportunity cost 13
Figurc 5. Potcntial nct prcscnt valuc ol sustainablc timbcr production 14
Figurc 6. Forcst carbon stocks ol thc 8razilian Amazon 15
Figurc 7. Nct opportunity cost ol lorcst protcction in thc 8razilian Amazon 16
Figurc 8. Marginal opportunity cost ol rcductions in carbon cmissions 17
Figurc 9. Tirtyycar trajcctory ol an illustrativc, hypothctical R¡ÐÐ program 18
Tablc 3. Summary ol costs ol 8razilian Amazon R¡ÐÐ program in Ycar 10 19
Figurc 10. ¡xamplc ol thc cstimatcd costs 22
1
¡xvcu:ivv Su::~vv:
• Tropical dclorcstation and lorcst dcgradation contributcd 7 to 28º ol worldwidc, humaninduccd carbon
cmissions to thc atmosphcrc in thc 1990s (0.5 to 2.4 billion tons) climbing to morc than 3.0 billion tons
during ycars ol scvcrc drought and lorcst ?rc. An important ncw carbon crcdit rcgimc is undcr ncgotiation
within thc UN Framcwork Convcntion on Climatc Changc (UNFCCC) lor thc post2012 pcriod that
could compcnsatc tropical countrics lor thcir nationwidc rcduction in cmissions lrom dclorcstation and
lorcst dcgradation (R¡ÐÐ).
• 8razil is a primc candidatc lor a R¡ÐÐ program bccausc ol its groundbrcaking succcsscs in rcducing
and monitoring dclorcstation and lorcst dcgradation in thc Amazon rcgion, whcrc most ol its cmissions
occur (~70º). 8razil contains morc carbon in tropical lorcst trccs than any othcr country—47±9 billion
tons in 3.3 million squarc kilomctcrs ol lorcst in thc Amazon alonc. 8ut thcrc is considcrablc dcbatc
and discussion ovcr how R¡ÐÐ programs will work, and how much thcy will cost to thc implcmcnting
nations.
• Tis rcport prcscnts a conccptual lramcwork lor cstimating thc costs to tropical nations ol implcmcnting
R¡ÐÐ programs and applics this lramcwork to thc 8razilian Amazon rcgion. Vc cstimatc thc opportunity
costs ol lorcst conscrvation and, scparatcly, calculatc thc annual and 30ycar costs ol rcducing carbon
cmissions lrom dclorcstation and lorcst dcgradation in thc 8razilian Amazon to closc to zcro ovcr a tcn
ycar pcriod. Vc cnd with an initial asscssmcnt ol thc bcnc?ts ol thcsc rcductions to 8razilian socicty and
clcmcnts ol thc institutional arrangcmcnts that will bc ncccssary to managc this ncw markct.
• R¡ÐÐ programs will bc complcx and must bc rc?ncd through dialoguc, dcbatc, and cxchangc ol idcas and
approachcs among a divcrsc, intcrnational group ol stakcholdcrs. Tis rcport is dcsigncd to stimulatc this
dialoguc and “dcmystily” kcy challcngcs ol R¡ÐÐ by proposing a practical conccptual lramcwork and an
initial cstimatc ol how much R¡ÐÐ would cost in thc 8razilian Amazon.
Te purpose of this report is to demystify the key challenges of REDD, and to stimulate
dialogue and innovation toward solving these challenges.
2
P???????
• #1. Te costs vs. the value of reduced emissions. Our goal is to estimate the cost of developing a REDD
program in the Brazilian Amazon, not the value of such a program. Te value of Amazon forest conservation
far exceeds the costs of protecting it, although these values are di? cult to monetize. Te ultimate price of
REDD carbon credits and, hence, the ?ow of money into REDD, will be determined by the size of the
world carbon market which is, in turn, de?ned by the emissions reduction targets that developed countries
commit to.
•
• #2. National opportunity costs. Te costs of REDD programs should be bounded by the nation-
wide opportunity costs of forgone pro?ts from forest-replacing agricultural and livestock production
systems applied to forest lands and potentially forested lands less pro?ts/savings associated with forest
maintenance.
• #3. All forest lands. Opportunity costs should be estimated for all forest lands since parks, forest concessions,
and forest law can be undone to permit forest-replacing agriculture. Te portion of these opportunity costs
that are recovered through carbon payments may vary by land category (e.g. protected areas vs. private
forests).
• #4. Compensating forest stewards. Forest-based cultures, including indigenous groups, traditional societies,
and some smallholder farmer communities, should be compensated for their historical and ongoing role—
or potential role—as forest stewards.
• #5. Current government budget outlays continue. Payments to the government are for costs above and
beyond current budget outlays for the management and protection of forests.
• #6. Carbon payments for governance. Within REDD, payments for the ecosystem services of carbon
retention in forest biomass are applied to the entire REDD program, including payments to forest stewards
and to government.
• #7. Te deforestation scenario. We assume that the REDD program reduces deforestation in the Brazilian
Amazon to approximately zero over a ten year period from a current baseline of 20,000 km
2
per year.
• #8. A century-long payment schedule. Brazil should receive REDD payments at a rate that is commensurate
with the rate of reductions in emissions. At current rates, it would take more than a century to clear the
forests of the Brazilian Amazon, hence payments should continue over this period.
3
Tnvvv Fuxus
• Òur approach cnvisions thrcc major componcnts ol a R¡ÐÐ program:
(a) a Public Forcst Stcwardship Fund,
(b) a Privatc Forcst Stcwardship Fund, and
(c) a Govcrnmcnt Fund.
!n this rcport, wc prcscnt onc sccnario ol illustrativc cstimatcs ol thc costs ol cach lund. Morc dctailcd
prcscntation ol thc mcthods can bc lound at http://whrc.org/8razilcarbonsupplcmcnt/
Fovvs: ~iioc~:iox
• !n thc sccnario prcscntcd hcrc, thc cvcntual allocation ol thc 3.3 million km
2
ol lorcst rcmaining in thc
rcgion would bc: 40º “Social” Rcscrvcs, 30º “8iological/¡cological” and “Production” Rcscrvcs, and
30º privatc land rcscrvcs.
Òvvov:uxi:v cos:s
• Vc cstimatc thc opportunity costs (ÒCs) ol complctc lorcst conscrvation as an initial uppcr bcnchmark
lor asscssing thc cost ol R¡ÐÐ. ÒCs arc calculatcd using spatiallycxplicit modcls ol potcntial rcnts
lrom soy, cattlc and timbcr production. For cach lorcstcd parccl (4 km
2
), rcnts lor cach compcting land
usc (soy, cattlc, timbcr) arc summcd lor 30 ycars assuming a 5º discount ratc and a prcdctcrmincd
schcdulc ol highway paving. Considcring only thc maximum opportunity cost ol lorgonc pro?ts lrom
soy vs. cattlc ranching, thc ÒC ol prcscrving thc rcmaining lorcsts ol thc 8razilian Amazon (3.3 million
km
2
and 47 billion tons ol carbon) is 85.5 pcr ton ol carbon, and a total ol 8257 billion. Tis cost
dcclincs to 82.8 pcr ton ol carbon and 8123 billion ovcrall il lorcst convcrsion to soy and cattlc ranching
is pcrmittcd on thc 6º ol rcmaining lorcstcd lands that havc thc highcst opportunity costs (370,000
km
2
ol lorcst containing 3 billion tons ol carbon). Ònc lourth ol this high potcntial lorcst land would
bc clcarcd during thc ?rst tcn ycars ol thc program. Tc subtraction ol potcntial rcvcnucs lrom timbcr
managcmcnt rcduccs opportunity costs by only 4º.
4
Puniic Fovvs: S:vw~vus Cos:s
• !ndigcnous groups, rubbcr tappcrs, and othcr lorcstbascd populations dclcnd public lorcsts—or could
potcntially bccomc lorcst dclcndcrs—but havc rarcly rcccivcd compcnsation to do so. Tcy control
26º ol thc lorcsts ol thc 8razilian Amazon, and wc assumc will cvcntually control 40º through thc
crcation ol ncw rcscrvcs. Tc Public Forcst Stcwardship Fund would compcnsatc thcsc populations
with thc goal ol incrcasing thc viability ol lorcstbascd livclihoods and strcngthcning thcir rolc as lorcst
stcwards. Paymcnts would bc ticd to pcrlormancc. To providc thc annual cquivalcnt ol a 1/2 minimum
salary (81,200 pcr ycar) to all ca. 150,000 lorcst stcward lamilics living in “social” rcscrvcs (indigcnous
lands, cxtractivc rcscrvcs, sustainablc dcvclopmcnt rcscrvcs) would cost 8180 million pcr ycar. Anothcr
813 million would bc nccdcd to support thcsc groups in pcrimctcr patrol ol thcir rcscrvcs. Annual
compcnsation cquivalcnt to onc hall ol a minimum salary would cnablc an additional 50,000 smallholdcr
lamilics (860 million pcr ycar) living in govcrnmcnt agricultural scttlcmcnts to rcstorc lorcsts on
dcgradcd land as thcy shilt to highcarbon, stablc production systcms. Paymcnts would dcclinc ovcr
timc as lorcst stcwards shilt to lorcstbascd cconomics.
Pviv~:v Fovvs: S:vw~vus Cos:s
• Privatc lorcst stcwards in thc 8razilian Amazon arc privatc landholdcrs with lcgal titlcs to thcir land. Tcy
arc currcntly rcquircd to maintain 80º ol thcir land in lorcst, but compliancc is low and rcpcal ol this law
is lrcqucntly thrcatcncd. Vc assumc that currcnt lcgal
1
privatc landholdcrs rcccivc partial compcnsation
(20º) ol thc opportunity costs ol thcir privatc land lorcst rcscrvcs that arc rcquircd lor compliancc with
thc law, and highcr compcnsation (100º) ol thc opportunity costs ol thcir privatc land lorcst rcscrvcs in
cxccss ol thc lcgal rcquircmcnt. !l wc also assumc that hall ol thc lorcsts that arc clcarcd cach ycar in thc
8razilian Amazon arc privatcly, lcgally owncd, annual compcnsation ol privatc lorcst stcwards would bcgin
at 89 million, climbing to a maximum ol 890 million altcr tcn ycars. Tcrcaltcr, paymcnts to privatc lorcst
stcwards would dcclinc as thc pool ol lcgally owncd, uncompcnsatcd privatc lorcst land diminishcs. Tosc
who acquirc thcir privatc lorcsts altcr a cuto? datc do not qualily lor thc compcnsation ol privatc lorcst
opportunity costs, sincc thcsc costs should bc rc?cctcd in thc land salc pricc.
Govvvx:vx: Cos:s
• Tc govcrnmcnts ol 8razil (lcdcral and statc) will incur addcd costs to achicvc lasting rcductions ol
carbon cmissions. Vc cstimatc thc annual addcd costs ol monitoring, protccting, and managing cxisting
public lorcsts at 825 million, with an additional 88 million pcr ycar to cstablish ncw public lorcsts. Tc
dcvclopmcnt ol a privatc lorcst monitoring and liccnsing systcm would cost 816 million pcr ycar to cstablish
and implcmcnt. Additional scrviccs to lorcst stcward lamilics bcyond currcnt lcvcls ol support (an addcd
8700 pcr lamily pcr ycar lor improvcd cducation, hcalth, justicc, and tcchnical assistancc scrviccs) would
cost an additional 8140 million pcr ycar lor 200,000 rural lamilics. Total, additional, annual govcrnmcnt
lund outlays would bc a maximum ol 8190 million pcr ycar.
5
A :niv:vvv~v cos: ~xu v:issioxs :v~jvcvv
• Òvcr thc ?rst 10 ycars ol a 8razilian R¡ÐÐ program, annual costs to 8razil would climb lrom 872 million
pcr ycar to 8530 million pcr ycar as annual cmissions lall lrom thc 250 million ton carbon basclinc to
roughly zcro. Òngoing costs altcr ycar 10 dcclinc as public lorcst stcwards shilt to lorcstbascd cconomics,
thc pool ol uncompcnsatcd privatc lorcst dcclincs, and govcrnmcnt costs dcclinc through grcatcr c? cicncy
and tax rcvcnucs. Òvcr thc 30ycar pcriod, carbon cmissions would bc approximatcly 6 billion tons bclow
thc historical basclinc and 13 billion tons bclow projcctcd cmissions at a cost ol 88 billion. Full paymcnt
ol thc opportunity costs ol thcsc rcduccd cmissions would bc approximatcly 818 billion. Tcrc is thcrclorc
a margin lor adjusting thc thrcc cost catcgorics upward. Carbon cmission rcductions would climb lrom
~25 million tons in ycar 1 to ~250 million tons in ycar 10 and bcyond.
Auui:iox~i nvxvvi:s
• Substantial cobcnc?ts ol this program includc: thc doubling ol incomc ol 200,000 rural lorcstbascd
lamilics, a rcduction in ?rcbascd costs to socicty (rcspiratory illncss, dcaths, agricultural and lorcstry
damagcs) ol 810 to 880 million pcr ycar, and protcction ol thc rainlall systcm that may supply much ol
thc 8razilian grain bclt and hydroclcctric cncrgy production ol thc industrial southwcst ol thc country.
Substantial nonmonctizcd bcnc?ts includc biodivcrsity conscrvation, such as avoidancc ol thc ncar
climination ol ?vc ccorcgions.
• ¡missions rcductions lrom dclorcstation or lossil lucls combustion madc today may always bc canccllcd
tomorrow, il a country or ?rm that has tradcd rcductions latcr cmits bcyond its targct. Any cmissions
trading rcgimc nccds mcchanisms to insurc against such lailurcs. !n thc casc ol R¡ÐÐ, a carbon crcdit
insurancc rcscrvc could bc crcatcd, such that somc ol thc rcductions achicvcd and dcmonstratcd would bc
hcld in rcscrvc as carbon insurancc in casc ol luturc incrcascs in dclorcstation or ?rcs.
C~vnox cvvui: ixsuv~xcv vvsvvvv
6
1. Introduction
Tropical deforestation and forest degradation released 0.5 to 2.4 billion tons of carbon each year during
the 1990s (Houghton 2005), and was therefore 0.8 to 2.8% of the annual worldwide human-induced emission
of carbon to the atmosphere. During El Niño episodes, when severe drought a?ects large areas of tropical
forests in the Amazon, SE Asia, and elsewhere, emissions can double through ?res that burn forests and tropical
peat soils (Page et al. 2002, Alencar et al. 2006). Tropical deforestation emissions may increase in the coming
decade as rising worldwide demand for animal ration, meat, and biofuel places new pressures on potential
agricultural lands in the tropics (Soares-Filho et al. 2006, Nepstad et al. 2006c, Nepstad et al. in press). We
estimate that in a business-as-usual scenario, 55% of the forests of the Brazilian Amazon will be cleared, logged,
or damaged by drought by the year 2030, releasing 20±5 billion tons of carbon to the atmosphere (Figure 1).
Tese predictions do not include the e?ects of regional or global climate change.
Figure 1. Te Brazilian Amazon in 2030, showing drought-damaged, logged, and cleared forests. Tis map assumes that
deforestation rates of 1997-2003 continue into the future, and that the climatic conditions of the last 10 years are repeated
into the future. From Soares-Filho et al. 2006, Nepstad et al. 2004, 2007, Nepstad and Stickler in press, Merry et al. in
review. (See Supplemental Online Material for description of methods at http://whrc.org/Brazilcarbonsupplement
7
Although greenhouse gas emissions from deforestation were excluded from the UN Framework
Convention on Climate Change (UNFCCC) negotiations of the Kyoto Protocol (Fearnside 2001, Moutinho
and Schwartzman 2005, Gullison et al. 2007, Schlamadinger et al. 2007a), such a system is part of the current
negotiations focused on the post-Kyoto (post-2012) period (Schlamadinger et al. 2007a). A proposal to
compensate tropical countries for nation-wide reductions in greenhouse gas emissions from deforestation and
forest degradation (referred to here as “REDD”), ?rst presented at the Milan Conference of the Parties in 2003
(Santilli et al. 2005). A similar proposal was advanced by Papua New Guinea, Costa Rica, and other tropical
nations at the Montreal COP in 2005 (Silva-Chavez and Petsonk 2006, Schlamadinger et al. 2007b, Skutsch et
al 2007, Sedjo and Sohngen 2007). Brazil endorsed a similar “tropical forest fund” at the Nairobi COP, but did
not support a market mechanism for supplying this fund (Government of Brazil 2006, Gri? ths 2007). SBSTA
2

negotiations on REDD will conclude with recommendations to COP 13 in Bali.
Te Brazilian government’s opposition to the carbon market-funded compensation of reductions in
carbon emissions from deforestation is surprising since it is superbly positioned to bene?t from a REDD
program. Roughly two thirds of Brazil’s annual carbon emissions come from deforestation, mostly in the
Amazon (Moutinho and Schwartzman 2005), and Brazil has been a world leader in developing innovative and
successful approaches to forest conservation, as described below.
One of the obstacles to the eventual approval of a REDD mechanism within the UNFCCC process is
uncertainty about how REDD would work, how much it would cost, and how much carbon would potentially
come into the carbon market at what price. In this report, we provide a conceptual framework for the
development of a REDD program for the Brazilian Amazon, an initial estimate of the cost of implementing this
program over a thirty year period, and the amount of carbon that could enter the carbon market. We complete
the report with a preliminary assessment of the co-bene?ts of a Brazilian Amazon REDD program.
Te purpose of this report is to help move discussions of REDD forward by providing a practical
framework for assessing costs and volumes of carbon at stake. Te actual costs of developing and implementing
a REDD program in Brazil would depend upon several premises and re?nements of cost analyses.
2. Can deforestation in the Brazilian Amazon be reduced to zero?
One of the biggest questions of the REDD dialogue is: can it be done? Brazil has provided several
important examples that illustrate the feasibility of lowering deforestation. For example, from January 2004
through December 2006, 23 million hectares of public forest reserves in the Brazilian Amazon were created,
including large forest reserves at the edge of the active agricultural frontier (Campos and Nepstad 2006,
Nepstad et al. 2006a). Brazil’s Mato Grosso state has a sophisticated system of private forest reserve monitoring
(Fearnside 2003, Chomitz and Wertz-Kanounniko? 2005, Lima et al. 2005) and one of the world’s most
advanced systems of rainforest monitoring (INPE 2007). An ambitious federal government program to reduce
Amazon deforestation succeeded in cutting rates in half from 2004 to 2006, (aided by the plummeting prices
of soy and beef ). More recently, the “National Pact for Valuing the Amazon forest and Ending Deforestation”
3
,
with political support from the Federal Government, four Amazon state governors, the environmental NGO
community, and segments of the private sector, has proposed a seven-year target to reduce deforestation to zero.
Among the Pact’s supporters is Blairo Maggi, Governor of the state of Mato Grosso State, which emits more
greenhouse gases from deforestation than any other state during most years. Te Brazilian Congress has also
developed legislation proposals that would establish national deforestation emission reduction targets.
8
3. Premises
• #1. Te costs vs. the value of reduced emissions. Our goal is to estimate the cost of developing a REDD
program in the Brazilian Amazon, not the value of such a program. (Te value of Amazon forest
conservation far exceeds the costs of protecting it, although these values are di? cult to monetize.) We
assume that nations estimate the acceptable carbon price for their REDD programs in a way that is
commensurate with the cost of achieving these reductions less the economic bene?ts that accrue to that
nation through forest conservation. Tis report contributes to discussions on the amount of carbon that
could come into the carbon market from tropical forests, and at what minimum price. Te ultimate price
of REDD carbon credits and, hence, the ?ow of money into REDD, will be determined by the size of the
world carbon market which is, in turn, de?ned by the emissions reduction targets that developed countries
commit to.
• #2. National opportunity costs. Te maximum costs of REDD programs should be constrained by the
nation-wide opportunity costs of forgoing forest clearing and thinning less pro?ts from low-emissions
forest-based economic activities. Tese costs include forgone pro?ts from forest-replacing agricultural and
livestock production systems applied to forest lands and potentially forested lands. Tese costs are o?set
by revenues from forest-based economic activities, such as timber production, and other local and national
bene?ts that are often more di? cult to monetize, such as reduced economic damages from ?re. Tis report
considers opportunity costs incurred over a 30-year time horizon.
• #3. All forest lands. Opportunity costs should be estimated for all forest lands and potentially-forested
lands, not just those that are privately held. Parks and forest concessions can be undone to permit forest-
replacing agriculture. Land laws can be modi?ed to liberate landholders to clear their forests. Ongoing
positive economic incentives are needed to keep forests standing.
• #4. Compensating forest stewards. Forest-based cultures, including indigenous groups, traditional societies,
and some smallholder farmer communities, should be compensated for their historical and ongoing role—
or potential role—as forest stewards (Nepstad et al. 2006b, Gri? ths 2007). Tis compensation should be
designed to foster the development of forest-based livelihoods, maximizing the social and environmental
bene?ts of the REDD program. Similarly, REDD must provide positive economic incentives to agricultural
and livestock producers who hold legal
1
titles to their land and demonstrate their commitment to sound
forest stewardship and compliance with the law.
• #5. Current government budget outlays continue. Payments to the government are for costs above and
beyond current budget outlays for the management and protection of forests. We assume that governments
maintain current investments, thereby increasing the additionality of the REDD program. Tis premise
carries the moral hazard of rewarding countries that have invested little in natural resource conservation in
the past.
• #6. Carbon payments for governance. Within REDD, payment for the ecosystem service of carbon
retention in forest biomass is applied to the entire REDD program, including payments to forest stewards
and to the government. Tis expanded concept of payments for ecosystem services is necessary since
REDD is a nation-wide program.
9
• #7. Te deforestation scenario and forest allocation. We estimate that a well-designed REDD program
could reduce deforestation in the Brazilian Amazon to approximately zero over a ten year period from
a current baseline of 20,000 km
2
and approximately 250 millions tons of carbon emissions per year. Te
allocation of forest land at the end of the REDD program would be 40% social reserves, 40% biological
and production forest reserves, and 30% private property, from the current distribution of 26%, 31%, and
20%, respectively. (Te remaining forest land is undesignated.) Te allocation of forested land de?nes the
cost estimates, since costs vary among social forests (and their inhabitants), biological reserves, production
reserves, and private forested lands. Tese premises will be the subject of considerable analysis and debate.
We use a 20,000 km
2
/year deforestation baseline since average deforestation from 1997 through 2006 was
19,200 km
2
/year (INPE 2007) and deforestation is projected to increase in the future (Soares-Filho et al.
2006).
• #8. A century-long payment schedule. Brazil should receive REDD payments at a rate that is commensurate
with the rate of reductions in emissions. At current rates (20,000 km
2
/yr), it would take more than a
century to clear the forests of the Brazilian Amazon (3.3 million km
2
). Tis simple premise creates a long-
term incentive for tropical countries to invest in maintenance of their forest carbon stock, and it reduces
the risk that a large ?ow of REDD carbon credits would dilute the carbon market.
4. Te conceptual framework of a Brazilian Amazon REDD program
Most e?orts to quantify the costs of reducing greenhouse gas emissions from tropical deforestation
and forest degradation have focused on estimating the opportunity costs associated with forgone pro?ts
from agriculture and livestock production that are incurred when restrictions to forest clearing are imposed.
Tese analyses have employed equilibrium and partial equilibrium global economic models to estimate these
opportunity costs and have had to make simplifying assumptions about potential rents from agriculture and
livestock on tropical forest lands (Kremen et al. 2000, Sedjo et al. 2001, Sathaye et al. 2006, Obersteiner et al.
2006, Sohngen and Sedjo 2006, Kindermann et al. 2006). We are unaware of published analyses that estimate
the opportunity costs of REDD programs from the ground up, beginning with the biophysical, climatic, and
infrastructure constraints to agriculture and livestock expansion in tropical forest regions, and then re?ning
these costs through analysis of a REDD program framework. In this report, we present results of a model
of opportunity costs of forest maintenance estimated using spatially-explicit rent models for high-carbon
(timber) and low-carbon (agriculture, ranching) uses of Brazilian Amazon forests. (Additional details on the
methodology used to make these estimates can be found on-line at http://whrc.org/Brazilcarbonsupplement)
We estimate opportunity costs of forgone pro?ts from non-forest land uses as an upper limit benchmark
to the cost of REDD programs. Te actual costs of REDD programs should be considerably lower than full
compensation of these opportunity costs since Brazilian society has already taken steps to remove much of
these forests from the agricultural/livestock land market through the creation of formal forest reserves. Te
cost of REDD should also be lower than full compensation of opportunity costs because of the bene?ts of
forest protection that accrue to Brazilian society. For example, there is some evidence that the rainfall system
of central and southwestern Brazil is partially dependent upon moisture coming from the Amazon region and
that this moisture is, in turn, dependent upon Amazon forest evapotranspiration (Clement and Higuchi 2006).
Hence, the rains that feed Brazil’s grain belt and extensive hydroelectric reservoir network appear to depend, in
part, upon Amazon forests.
10
Te institutional steps to achieve lasting reductions in carbon emissions from tropical deforestation and
forest degradation are also in need of a clarifying conceptual framework. REDD programs will depend upon
e?ective governance of remote forest regions and an equitable, e? cient system of channeling these incentives
to the people who control tropical forests. We propose three general targets of REDD funding to help meet
these goals. First, a “Public Forest Stewardship” fund would compensate those people who have defended forests
against forest-replacing economic activities, or who could potentially defend forests. Tis funding targets forest-
based indigenous groups, traditional rural populations (such as rubber tappers, Brazil-nut gatherers, and others),
and some smallholder populations that are taking steps towards stable land-use systems that maintain or expand
carbon stocks in forest vegetation.
A “Private Forest Stewardship Fund” would compensate those legal private landholders who retain
forest on their properties. (Tis fund is complicated by the di? culty of de?ning land ownership in the Brazilian
Amazon.) We propose a di?erential rate of compensation for forest conservation on private land, with lower
compensation going to forest reserves that are legally required, and higher compensation going to reserves that
are above and beyond this legal requirement.
A “Government Fund” would compensate government programs and expenditures that are necessary
for REDD above and beyond current budget outlays. Tese expenditures include heightened monitoring and
management of public forests, expansion of the protected area and indigenous land network of public forests,
improved provision of services (education, health, technical assistance) to rural populations, and the expansion
of existing systems for environmental licensing and monitoring of private land forests to the entire Brazilian
Amazon region.
5. A spatial map of opportunity costs
Te opportunity costs of maintaining the forests of the Brazilian Amazon (Figure 2) was mapped using
spatially-explicit models of potential rents for soy, cattle, and timber production. Tese models were developed
as part of the “Amazon Scenarios” program of the Woods Hole Research Center, the Universidade Federal de
Minas Gerais, and the Instituto de Pesquisa Ambiental da Amazonia. Te soy model integrates a biophysical
yield model, a transportation model, and a production cost model in estimating the economic returns to soy
production for the Brazilian Amazon (Vera Diaz et al. in press). Soy expansion is constrained by a soil and
climate suitability map that is applied as a ?lter (http://whrc.org/Brazilcarbonsupplement). Soy rents are
positive only in areas where suitability is high. Te cattle ranching model integrates a herd development model,
a production cost function (that includes land purchase, herd establishment, and periodic pasture reformation),
and a transportation cost model (Merry et al. in preparation). Te timber model integrates a transportation
model, a harvesting and processing cost model, and simulates the expansion, contraction, initiation, and
extinction of timber processing centers depending upon each center’s neighborhood of timber stocks that
could be pro?tably harvested (Merry et al. in review). (See online supplemental information for more details
(http://whrc.org/Brazilcarbonsupplement)
Tese three rent-based models are integrated within the “SimAmazonia” modeling system (Soares-
Filho et al. 2006). In this report, the net present value of each of the three competing land uses is estimated
over a 30-year time period by summing rents into the future for each forested pixel of the Brazilian Amazon
(Figure 3-5). Future rents are discounted at a 5% annual rate. All three models are highly sensitive to changes
in transportation costs. We therefore developed a schedule of highway paving based upon an analysis of current
policies and capital availability (Soares-Filho et al. 2006). Hence, the rent of each forested pixel changes
di?erentially through time for each competing land use depending upon expansion of the paved highway
network as prescribed.
11
We estimate the opportunity cost of maintaining forest for each 4-km
2
forest “pixel” as the maximum
net present value of deforestation-dependent land use (the maximum, discounted, 30-year rent of soy vs. cattle
ranching). We also estimate the “net” opportunity cost, in which the net present value of timber production is
subtracted from that of soy or cattle, since timber maintains most of the carbon stock of forests. In this report,
we “force” the timber industry into a sustainable mode by limiting annual harvest for each processing center
to 1/30
th
of the total timber volume around each processing center that could be pro?tably harvested. (Tis
assumes that each forested pixel can be harvested every thirty years because of tree growth.) Tis opportunity
cost is divided by the carbon stock for each forested pixel using the forest carbon map developed by Saatchi et
al. (2007, Figure 6), to estimate the payment per ton of carbon that would fully compensate the opportunity
costs of forest maintenance (Figure 6, 8). Te net opportunity cost is calculated by dividing the di?erence in net
present value (soy or cattle minus timber) by the di?erence in carbon stock of agriculture/livestock vs. timber
4
.”
Figure 2. Te forests in the Brazilian Amazon. Tis 5-million square kilometer region has 3.3 million square
kilometers of forest, with roughly half (49%) in public forests, including indigenous reserves, biological reserves
and parks, “sustainable use” (community development forests and production forests), and military reserves.
Source: (http://whrc.org/Brazilcarbonsupplement)
12
Figure 3. Te potential net present value (2007 through 2037) of soy production on the forested lands of the
Brazilian Amazon. (http://whrc.org/Brazilcarbonsupplement, Vera Diaz et al. 2007.)
Te 3.3 million square kilometers of forests in the Brazilian Amazon contain 47±9 billion tons of forest
carbon (excluding soil carbon) (Saatchi et al. 2007, Soares-Filho et al. 2006). Te opportunity cost of protecting
this forest all at once, in 2007 dollars, is $257 billion and $5.5 per ton of carbon. Only 6% of the forests of the
region have opportunity costs of more than $10 per ton carbon, however. If these forests are removed from our
estimate, the cost of fully compensating OCs declines to $123 billion and the per-ton cost to $2.8 (Tables 1, 2).
Outside of protected areas, there are 24 billion tons of carbon in forests with opportunity costs of $137 billion
($6.05 per ton carbon). By excluding the high-rent forest parcels (representing 6% of total forest area outside
of protected areas), it would be possible to fully compensate OCs of 22.2 billion tons of forest carbon for $56
billion ($2.75 per ton C) (Tables 1, 2).
Tese surprisingly low per ton values for carbon are attributable to the low pro?tability of cattle ranching
in the Amazon (Figure 4). Te animal grazing density of Amazon cattle pastures averages 0.8 animal units per
hectare, and yields pro?ts that are generally well below $50 per hectare per year (Arima et al. 2006, Margulis
2003, Mattos and Uhl 1994). Te opportunity costs of forgone pro?ts from soy production (Figure 3) represent
the steep part of the carbon supply curve in the ?nal 6% of the forest carbon stock (Figure 8). Tese OCs
decline by 4% if pro?ts from sustainable timber management (Figure 5), which can retain at least 85% of forest
carbon stocks, are subtracted from the OC estimate (Table 1).
Figure 4. Potential net present value of cattle production (2007-2037) on the forested lands of the Brazilian
Amazon. (http://whrc.org/Brazilcarbonsupplement).
Table 1. Opportunity Costs of
forest maintenance outside of
protected areas, inside of PAs, and
for the entire Brazilian Amazon.
With Timber
Rents
($B)
Without
Timber Rents
($B)
Percent
Reduction
Outside protected areas 137.5 143.4 4.1
Outside protected areas, <$10/ton 56.3 61.5 8.5
Inside protected areas 120.8 121.6 0.7
Inside protected areas <$10/ton 60.4 61.1 1.1
Total 247.3 257.1 3.8
Total <$10/ton 114.6 123.3 7.1
Table 2. Carbon stocks and
opportunity cost per ton C outside
of PAs, inside of PAs, and for the
entire Brazilian Amazon.
Carbon Stocks $ per ton C
Outside protected areas 23.8 6.03
Outside protected areas, <$10/ton 22.2 2.75
Inside protected areas 23.1 5.26
Inside protected areas <$10/ton 21.7 2.81
Total 47.1 5.65
Total <$10/ton 44.1 1.56
13
14
6. A deforestation reduction schedule and forest allocation
Our analysis is based upon a ten-year timetable for lowering deforestation to ~zero kilometers per year
from an historical baseline of 20,000 km
2
per year (Fig. 8). We use a 20,000 km
2
per year rate as our baseline
since deforestation for the last 10 years was 19,200 km
2
but reached an average of 24,000 km
2
per year during
the 2002-2004 period (INPE 2007). Deforestation is projected to increase in the future under business-as-usual
assumptions (Soares-Filho et al. 2006). Deforestation is assumed to be reduced by 2,000 km
2
per year until year
ten, when deforestation is reduced to ~0 km
2
per year. Te deforestation reduction schedule is presented for 30
years, which is the time period for which opportunity costs were estimated. In practice, compensation would
continue into the future at a rate that is commensurate with ongoing emissions reductions. During the 30-year
period, the deforested area would be reduced by 490,000 km
2
below the baseline and carbon emissions would
be reduced by 6.3 billion tons. If the 90,000 km2 of deforestation that takes place during the ?rst ten years of
this period is on forested lands with high opportunity costs of forest maintenance, then the remaining area of
potentially high-pro?t forest declines to 280,000 km
2
.
Our calculations also depend upon the ultimate allocation of forest land. Roughly one third of Brazilian
Amazon forests today are without formal designation (called “terra devoluta”, Lentini et al. 2003). Tirty-one
percent of forests are public forest reserves (26% of these being “social” reserves, including indigenous lands,
extractive reserves, and sustainable development reserves). Te remainder of the land is private. We assume
that remaining forests of the Brazilian Amazon will be allocated as: 40% social forests (where the public forest
stewardship fund applies), 30% biological and production forest, and 30% private land.
Figure 5. Potential net present value of sustainable timber production (2007-2037) for the forests of the Brazilian
Amazon. Processing centers in this run of the timber rent model are restricted to annual harvests of 1/30
th
of the
pro?tably harvestable timber stocks, thereby “forcing” the industry into sustainable, 30-year rotations. See http://.whrc.
org/Brazilcarbonsupplement for model description.
15
Figure 6. Forest carbon stocks of the Brazilian Amazon. Aboveground and roots. (Assumes that root biomass is 21% of
live aboveground biomass and that dead biomass is 9% of live aboveground biomass.) Source: Saatchi et al. 2007.
7. Te Public Forest Stewardship Fund
Indigenous communities inhibit deforestation at the same level as biological reserves and parks (Nepstad
et al. 2006b), providing an important rationale for strengthening their role as stewards of these public forests.
Tis rationale is further supported by the fact that 25% of current Brazilian Amazon forests are allocated to
some form of “social forest” use (indigenous land, extractive reserve, sustainable development reserve), and these
social reserves are much more common in active deforestation frontiers than are biological reserves and parks
(Nepstad et al. 2006b). Te “Aliança dos Povos da Floresta” (the Forest Peoples’ Alliance) has de?ned several
forms of compensation that it expects from a REDD program
5
. Tese forms of compensation include economic
incentives for forest-based livelihoods, improved health, education, technical assistance services, and payments
for patrolling reserve perimeters, and are described in greater detail in supplemental online information (http://
whrc.org/Brazilcarbonsupplement).
We estimate the cost of providing incentives for forest-based livelihood on a per-family basis. We
simplify this calculation by assuming that a payment of one-half of a minimum salary ($1,200 per year) would
be su? cient to provide a strong incentive to stabilize agricultural systems (through a shift to swidden fallow
that does not depend upon primary forest clearing) and to develop forest-based economies (e.g. McGrath et al.
2006). Te exact form of compensating forest stewards will depend upon a deeper analysis, and may include
16
price subsidies for non-timber forest products such as have already been established in Acre and Amazon
states for native rubber. Direct payments to forest families also have a precedent in the Amazon through the
Proambiente program and, more recently, through the Amazonas state “bolsa ?orestal” program. In the case
of Proambiente, payments of $50 per month (half of our estimate) were su? cient to foster changes in farmer
agricultural strategies. In Amazonas, payments are $25 per month. A payment of $1,200 per year for all 50,000
indigenous families, all 50,000 extractivist families, and for 50,000 forest-margin smallholder families would
cost $180 million per year (Table 3). We assume that it would take ten years of linearly increasing payments to
reach all families contemplated.
We estimate the cost of perimeter control based at $10 per square kilometer upon estimates from the
Aliança dos Povos da Floresta at $10 per square kilometer. Te 1.3 million square kilometers of social reserves
would require $13 million per year to be monitored by their residents (Table 3).
An additional incentive is included for those smallholder families that are in public settlement projects
that hold potential for forest restoration and a shift to stable agricultural systems. Sixty million dollars per year
would be necessary to compensate 50,000 smallholder families (out of a total of 650,000 smallholder families
across the Brazilian Amazon) (Table 3).
Figure 7. Net opportunity cost of forest protection in the Brazilian Amazon. Calculated as maximum net present value
of soy or cattle production minus NPV of timber. Te value was then divided by forest carbon stocks (Figure 6).
17
Figure 8. Marginal opportunity cost of reductions in carbon emissions for the Brazilian Amazon. Tis graph plots the
opportunity cost per ton of carbon, as described in Figure 7, from the cheapest to the most expensive emissions reductions.
Ninety percent of the opportunity costs are less than $5 and 94% are less than $10. Te total opportunity cost to
maintain the entire forest is $257 billion (if paid all at once in 2007 dollars) for 47 B tons of C; the cost of compensating
94% of the “cheapest” forests is $115 billion, with carbon stocks of 44 billion tons C.
18
Figure 9. Trajectory of deforestation, reduced deforestation, the opportunity costs of this reduction, and an initial
estimate of the cost of achieving the reduction (the sum of the Public Forest Stewardship, Private Forest Stewardship,
and Government funds) in the Brazilian Amazon for a thirty-year period. Te values of each fund are found in Figure
10.
19
Table 3. Summary of costs of Brazilian Amazon REDD program in Year 10
Public Forest Stewardship Fund (Forest People)
a. Forest steward compensation
Annual payment per family $1,200
100,000 indigenous and extractivist families $120,000,000
50,000 qualifying forest margin smallholders $60,000,000
b. Forest monitoring, protection, management
Average annual cost per square kilometer $10
1,000,000 km
2
indigenous reserves $10,000,000
200,000 km
2
extractive reserves $2,000,000
100,000 km
2
community reserves $1,000,000
c. Forest settlement restoration
Average annual cost per family $1,200
50,000 smallholder families $60,000,000
d. Total annual forest people payments $253,000,000
Private Forest Stewardship Fund
Opportunity costs compensation, extensive ranching $90,000,000
Te Government Fund
a. Public forest protection, management, creation
Monitoring: average annual cost per square kilometer $20
Maintenance of current public forests $24,800,000
Cost to create new protected area ($/km
2
) $50
Creation of new protected areas (10%/yr) $7,800,000
b. Private forest registration, monitoring
Env’l registration system establishment (10%/yr) $10,000,000
Cost to register private lands ($/km
2
) $50
Property registration (10% per year, $200 per km
2
) $6,000,000
c. Services (health, education, justice, technical support)
Annual payment per family $700
Annual payments for forest peoples $140,000,000
d. Total Government Fund $188,600,000
Total cost of all funds in year 10 $531,600,000
8. Te Private Forest Stewardship Fund
It is very di? cult to quantify the area of Amazon forests that are legally owned or that could be legalized
without rewarding ?agrant fraud (Alston et al. 1999). Antiquated titling processes, competing land claims, and
sophisticated illegal land grabbing operations make it virtually impossible to map legal land claims. For the
purpose of this report, we assume that one half of the forests cleared each year are on private properties that are
legally held or that will eventually be legalized. Tose who purchase forest lands in the future do not qualify for
compensation of their opportunity costs, since these costs should be re?ected in the sale price of the land. (If
we assume that Brazil will enter a regime of forcefully lowering deforestation rates, land prices should decline
as the possibility of forest conversion to agriculture or livestock declines.) Landholders are legally required to
maintain 80% of their property as private forest reserve. However, there are frequent attempts to turn back this
legislation and compensation of these legally-mandated forest reserves is therefore appropriate. We estimate
compensation of 20% of the opportunity costs of forest maintenance for these legally-mandated private forests.
Compensation of opportunity costs should be higher for forests held in excess of this 80% requirement, but the
number of properties with more than 80% forest cover is too small to a?ect these estimates. We estimate that
compensation of private forest stewards increases linearly until year 5, when these payments would equal $90
million per year (Table 3).
20
9. Te Government Fund
Te cost of government monitoring and management of existing public forests is estimated at $20
per km
2
and would cost an additional $28 million per year to be accomplished successfully. We assume
that protected area expansion would take place over 10 years to achieve the ?nal land allocation of 40% in
social reserves and 30% in biological and production reserves, adding 36,000 km
2
each year. If protected area
creation costs an additional $50 per km
2
, this cost would be $7.8 million per year. (Te added burden on the
government of an expanding protected network is counterbalanced by the growing capacity of public forest
stewards to defend and manage these areas.) Development of state-run private land environmental licensing
and monitoring systems, similar to Mato Grosso State’s “Sistema de Licensiamento Ambiental de Propriedades
Rurais” (Rural Property Environmental Licensing System) (Fearnside 2003, Lima et al. 2005, Chomitz and
Werth-Kanounniko? 2005), would cost $10 million per year for ten years, with an additional $50 per km
2
to
bring new private properties into the system ($6 million) (Table 3, http://whrc.org/Brazilcarbonsupplement).
Te largest governmental cost would be enhancement of its services provided to forest stewards.
Additional investments in and improvements to public health, education, and technical support programs are
estimated at $700 per family, for a total of $140 million per year (Table 3, http://whrc.org/Brazilcarbonsupple
ment). Tese additional funds would be channeled through existing institutions, such as the “Sistema Única de
Saúde”, in the case of health for non-indigenous families.
10. Te costs of REDD in the Brazilian Amazon over 30 years
We estimate the costs to Brazil of carrying out this REDD program over 30 years, which is the period
for which opportunity costs were calculated (Figure 7, 8). We assume that the Public and Private Forest
Stewardship Fund increases linearly over ten years to their maximum values presented in Table 3 (Figure 10).
Government costs must build up more rapidly to provide necessary law enforcement early in the program. We
assume that the Government Fund builds up linearly over a ?ve-year period. First year combined expenditures
of $72 million climb to $530 million in year 10 as deforestation declines from 20,000 km
2
to ~0 km
2
and
emissions decline from ~250 million to ~0 tons of carbon per year. After the initial ten-year period, ongoing
costs are incurred as Brazil continues to compensate remaining private land forest stewards, and for protecting/
managing the 2.3 million km
2
public forest estate. Tese ongoing payments are theoretically justi?ed as the
continuing, partial compensation of opportunity costs that will end >100 years into the future. Tis long time
horizon is necessary to fully compensate these opportunity costs because compensation is commensurate with
emissions reductions, which are determined by the 20,000- km
2
per year baseline. Tis long payment schedule
also provides an ongoing incentive to Brazil to continue its forest governance. We assume that the cost of
achieving forest governance declines over time as institutional e? ciency increases, and as the tax base of the
government expands through a thriving timber industry.
Over the thirty-year period, $8.2 billion are expended to reduce emissions of carbon by 6.3 billion tons.
In other words, for a bit more than a dollar per ton of carbon, emissions of carbon to the atmosphere could
be reduced by an amount equivalent to about seven months of worldwide emissions (which, in 2006, passed
10 billion tons per year, Canadell et al. 2007) while conserving the world’s largest tropical rainforest. Te full
opportunity cost of avoiding the emission of 6 billion tons of carbon would be $3 per ton, or $18 billion, if we
assume that the highest 6% of opportunity costs are not compensated (Table 2, Figure 8). Part of the di?erence
between these two estimates of REDD costs ($8 vs. $18B) is diminished by the bene?ts to Brazilian society of a
REDD program. In other words, there are substantial bene?ts to Brazilian society of protecting Amazon forests
that should be counted against opportunity costs as the real cost of a REDD program is estimated.
21
11. Co-bene?ts of REDD
Te proposed REDD program would have direct impacts on the livelihoods of 200,000 low-income
rural families, including all of the indigenous and traditional families of the Brazilian Amazon. Tese families
would more than double their incomes as they shift to forest-based economic activities. Tey would also receive
$700 per family per year in added educational, health, and technical support services. Te program would reduce
the likelihood of deforestation-driven reductions in rainfall in the Brazilian grain belt (Clement and Higuchi
2007), and would also reduce the likelihood of drought-driven energy shortages, such as the one that crippled
the Brazilian economy in 2003 when hydroelectric reservoirs dried up. By reducing the incidence of ?re, the
program would avoid $11 to 83 million dollars per year in ?re-related costs associated with respiratory ailments
and deaths, agricultural damages, and damages to timber if we assume that the incidence of ?re in the region
will decline together with the reductions in emissions (Mendonça et al. 2004 and http://whrc.org/Brazilcarbon
supplement). Te slowing of deforestation would also prevent the devastation of at least ?ve ecoregions whose
ranges would decrease by at least 85%. Tese ecoregions include the Maranhão babaçu forest, the Marañon dry
forest, and the Tumbes/Piura dry forest (Soares-Filho et al. 2006).
12. How will it work?
Detailed analysis of the mechanics of a Brazilian REDD program is beyond the scope of this report. Instead,
we propose a few key characteristics of the REDD program that would make it more likely to succeed.
• Carbon credit insurance reserve. Emissions reductions achieved today for deforestation or fossil fuel may
always be cancelled tomorrow if a country or ?rm that has traded reductions later emits beyond its target.
Tis problem is particularly important for REDD because of the risk of forest ?re. Any emissions trading
regime needs mechanisms to insure against such failures. In the case of REDD, a carbon credit insurance
reserve could be created, such that some of the reductions achieved and demonstrated would be held in
reserve as carbon insurance in case of future increases in deforestation or ?res. Contractual liability rules
should be established as part of the REDD negotiation to determine whether the seller, the buyers, or both
are responsible for the insurance reserve. If Brazil were to assume responsibility for a very conservative ratio
of insurance reserve to marketable reductions, of 1:1, this would in e?ect double the cost of implementing
REDD. Te greater the seller’s willingness to provide such insurance, the more competitive its reductions
would be in the market.
• Transparency and oversight. Te Brazilian Amazon REDD program will depend upon major strides in
improving the e? ciency of government institutions. Te success of the program will depend upon the design
of e? cient, transparent systems for managing REDD funds, for issuing and implementing deforestation
permits during the ?rst ten years of the program, for managing the timber sector, for developing programs
that support a transition to forest-based economies among public forest stewards, and for determining the
fair compensation cost to private forest stewards will be central to the success of the program.
• Monitoring and validation. Brazil has developed the world’s most successful system of rainforest monitoring
(INPE 2007). Tis system could become even better as it begins to incorporate recent innovations in the
mapping of Amazon forest degradation (Asner et al. 2005, Oliveira et al. 2007) and cloud-free mapping
of land cover and biomass using new radar sensors, such as ALOS/PALSAR (Kellndorfer et al. 2007,
companion report). In the near term, Brazil’s “PRODES” monitoring program could be supplemented
with annual mapping of the entire Brazilian Amazon forest formation, with no interference from clouds
and with biomass estimates for a large portion of cleared lands, for a price that is well below optical sensor
methods.

Figure 10. Example of the estimated costs of the Public Forest Stewardship Fund (Forest People), the Private Forest
Stewardship Fund, and the Government Fund over a thirty-year period using the premises set forth in this report.
13. Conclusion
Tis analysis indicates that carbon emissions from the Brazilian Amazon might decrease by six billion tons over
a thirty-year period through a fairly modest ?ow of funding into the region—about $8 billion. Tis estimate
is lower than previous estimates of REDD (Sathaye et al. 2006, Obersteiner et al. 2006, Sohngen and Sedjo
2006, Stern 2006), largely because opportunity costs are not fully compensated, and spatially-explicit modeling
of land use rents demonstrates that most carbon emissions carry very low opportunity costs. A REDD program
that compensates at a level that is less than opportunity cost is justi?able given the very substantial bene?ts that
this program would provide to Brazilian society. Tese include the doubling of income and improved health,
education, and technical assistance services for 200,000 forest-dwelling families. Te bene?ts also include a
more secure rainfall system for central and southern Brazil, and the avoidance of $11 to 83 million per year
in ?re-related damages to the Amazon economy. Te successful reduction of emissions to nearly zero over a
decade is a daunting task, and will depend upon innovation and major strides in the development of e? cient,
transparent institutions.
22
F????????
1
Landholders are “legal” if they have clear title to their property or have been issued legal declarations (“Termo de Ajuste de
Conduto”) through which they commit to take the necessary steps to legalize their properties.
2
Subsidiary Body for Scienti?c and Technological Advice
3
Te Pact was launched on October 3
rd
2007 within the National Congress (Committee of Environment and Sustainable
Development). Tis launching session was attended by the Minister of Environment, Marina Silva, two state governors (Mato Grosso
and Amapá), Secretaries of two others states (Amazonas and Acre) and the main environmental relevant congressmen. Te Pact
establishes an agreement among di?erent sectors (State Amazon governors, Federal governor, representatives from the rural producers,
from the agribusiness industries, socio-environmental organizations, social movements, indigenous and traditional population living in
the forests) to acknowledge the value of the standing forest and eliminate the deforestation in Amazonia over the next seven years.
4
We assume that logging decreases carbon stocks by 15% (Asner et al. 2005) while soy and pasture reduces stocks by 85% (Fearnside
1997). Carbon emission reduction is taken as the di?erence between these two for a given forest pixel that is not cleared.
4
We assume that logging decreases carbon stocks by 15% (Asner et al. 2005) while soy and pasture reduces stocks by 85% (Fearnside
1997). Carbon emission reduction is taken as the di?erence between these two for a given forest pixel that is not cleared.
5
Tis report is not an o? cial representation of the expectation of the Aliaca dos Povos da Floresta, but is
informed by discussions with its members.
23
LITERATURE:
Alencar, A., D. C. Nepstad, and M. d. C. Vera Diaz (2006), Forest understory ?re in the Brazilian Amazon in
ENSO and non-ENSO Years: Area burned and committed carbon emissions, Earth Interactions, 10(Art.
No. 6).
Alston, L. J., and G. D. Libecap (1999), Titles, Con?ict and Land Use: Te Development of Property Rights on the
Brazilian Amazon Frontier, University of Michigan Press, Ann Arbor, MI.
Arima, E., P. Barreto, and M. Brito (2006), Cattle ranching in the Amazon: trends and implications for
environmental conservation, IMAZON, Belem.
Asner, G. P., D. E. Knapp, E. N. Broadbent, P. J. C. Oliveira, M. Keller, and J. N. Silva (2005), Selective logging
in the Brazilian Amazon, Science, 310(5747), 480-482.
Campos, M. T., and D. C. Nepstad (2006), Smallholders, the Amazon’s new conservationists, Conservation
Biology, 20(5), 1553-1556.
Canadell, J. G., C. Le Querec, M. R. Raupacha, C. B. Field, E. T. Buitenhuisc, P. Ciaisf, T. J. Conway, N. P.
Gillett, R. A. Houghton, and G. C. Marland (2007), Contributions to accelerating atmospheric CO
2

growth from economic activity, carbon intensity, and e? ciency of natural sinks, Proc. Nat. Acad. Sci.
Chomitz, K. M., and S. Wertz-Kanounniko? (2005), Measuring the initial impacts on deforestation of Mato
Grosso’s program for environmental control, Te World Bank Group. World Bank Working Paper No.
WPS3762, Washington, D.C.
Clement, C. R., and N. Higuchi (2006), A ?oresta Amazonica e o futuro do Brasil. Ciencia e Cultura 58(3).
Fearnside, P. M. (1997), Greenhouse gases from deforestation in Brazilian Amazonia: net committed emissions,
Climatic Change, 35, 321-360.
Fearnside, P. M. (2001), Saving tropical forests as a global warming countermeasure: an issue that divides the
environmental movement, Ecological Economics, 39, 167-184.
Fearnside, P. (2003), Deforestation control in Mato Grosso: a new model for slowing the loss of Brazil’s Amazon
forest, Ambio, 32(5), 343-345.
Government of Brazil (2006), Positive incentives for voluntary action in developing countries to address climate
change: Brazilian perspective on Reducing Emissions from Deforestation Paper presented to UNFCCC
COP 12, Nairobi, Kenya, November 2006., http://unfccc.int/?les/meetings/dialogue/.
Gri? ths, T. (2007), Seeing “RED”? “Avoided deforestation” and the rights of Indigenous Peoples and local communities.
, Forest Peoples Programme., Moreton-in-Marsh, UK.
Gullison, R. E., P. C. Frumho?, J. G. Canadell, C. B. Field, D. C. Nepstad, K. Hayhoe, R. Avissar, L. M.
Curran, P. Friedlingstein, C. D. Jones, and C. Nobre (2007), Tropical forests and climate policy, Science,
316(5827), 985-986.
Houghton, R. A. (2005), Tropical deforestation as a source of greenhouse gas emissions in Tropical Deforestation
and Climate Change, edited by P. Moutinho and S. Schwartzman, pp. 13-21, Amazon Institute for
Environmental Research, Belém, Pará, Brazil.
Instituto Nacional de Pesquisa Espacial - INPE (2007), Estimativas Annuais de Desmatamento. Projecto
PRODES monitoramento da ?oresta Amazônica Brasileira por satélite, Available at http://www.obt.
inpe.br/prodes/.
Kellndorfer, J., W. S. Walker, D. Nepstad, P. Brando, C. Stickler, P. Lefebvre, and K. Kirsch (2007), ALOS
Radar image data for deforestation assessment in the Xingu Watershed, Mato Grosso, Brazil: a WHRC
publication for COP 13, 3-14 December, Bali, Indonesia.
24
Kindermann, G., M. Obersteiner, E. Rametsteiner, and I. McCallum (2006), Predicting the deforestation-trend
under di?erent carbon-prices Carbon Balance Management, 1, 15, http://www.pubmedcentral.nih.gov/
articlerender.fcgi?artid=1766350&rendertype.
Kremen, C., J. O. Niles, M. G. Dalton, G. C. Daily, P. R. Ehrlich, J. P. Fay, D. Grewal, and R. P. Guillery (2000),
Economic incentives for rain forest conservation across scales, Science, 288(5472), 1828-1832.
Lentini, M., A. Veríssimo, and L. Sobral (2003), Fatos Florestais da Amazônia 2003, Imazon, Available at: http://
www.imazon.org.br/upload/im_livros_002.pdf, Belém, Brasil. .
Lima, A., C. T. Irigaray, R. T. Silva, S. Guimaraes, and S. Araujo (2005), Sistema de Licenciamento Ambiental
em Propriedades Rurais do Estado de Mato Grosso: Análise de Lições na Sua Implementação (Relatório
Final), Ministério do Meio Ambiente/Secretaria de Coordenação da Amazônia/Programa Piloto para
a Proteção das Florestas Tropicais do Brasil/Projeto de Apoio ao Monitoramento e Análise (AMA).
(Projeto PNUD: BRA 98/0005), Brasilia.
Margulis, S. (2003), Causas do desmatamento da Amazônia Brasileira - 1a edição, Banco Mundial, Brasília.
Mattos, M. M. d., and C. Uhl (1994), Economic and ecological perspectives on ranching in the eastern Amazon,
World Development, 22(2), 145-158.
McGrath, D. G., and M. D. C. Vera Diaz (2006), Soja na Amazônia: impactos ambientais e estratégias de
mitigação., Ciência e Ambiente, Janeiro.
Mendonça, M. J. C., M. d. C. V. Diaz, D. C. Nepstad, R. S. d. Motta, A. A. Alencar, J. C. Gomes, and R. A.
Ortiz (2004), Te economic costs of the use of ?re in the Amazon, Ecological Economics, 49(1), 89-105.
Merry, F. D., B. S. Soares-Filho, D. C. Nepstad, G. Amacher, and H. Rodrigues (in review), Amazon logging as a
valuation benchmark and conservation catalyst, Proceedings of the National Academy of Sciences.
Moutinho, P., and S. Schwartzman (2005), Tropical deforestation and climate change, Environmental Defense
Fund, Washington, DC.
Nepstad, D. C., P. A. Lefebvre, U. L. Silva Jr, J. Tomasella, P. Schlesinger, L. Solorzano, P. R. d. S. Moutinho, and
D. G. Ray (2004), Amazon drought and its implications for forest ?ammability and tree growth: a basin-
wide analysis, Global Change Biology, 10, 704-717.
Nepstad, D., P. Moutinho, and B. Soares-Filho (2006a), Te Amazon in a Changing Climate: Large-Scale
Reductions of Carbon Emissions from Deforestation and Forest Impoverishment, IPAM, WHRC, and
UFMG, Belem, Para, Brazil.
Nepstad, D. C., S. Schwartzman, B. Bamberger, M. Santilli, D. G. Ray, P. Schlesinger, P. A. Lefebvre, A. Alencar,
E. Prins, G. Fiske, and A. Rolla (2006b), Inhibition of Amazon deforestation and ?re by parks and
indigenous lands, Conservation Biology, 20(1), 65-73.
Nepstad, D. C., C. M. Stickler, and O. T. Almeida (2006c), Globalization of the Amazon soy and beef
industries: opportunities for conservation, Conservation Biology, 20(6), 1595-1603.
Nepstad, D., I. Tohver, D. Ray, P. Moutinho, and G. Cardinot (2007), Long-term experimental drought e?ects
on stem mortality, forest structure, and dead biomass pools in an Eastern-Central Amazonian forest,
Ecology.
Nepstad, D. C. and C. M. Stickler (in press), Managing the tropical agriculture revolution, Journal of Sustainable
Forestry, 27(1 & 2).
Obersteiner, M., G. Alexandrov, P. C. Benítez, I. McCallum, F. Kraxner, K. Riahi, D. Rokityanskiy, and Y.
Yamagata (2006), Global supply of biomass for energy and carbon sequestration from a?orestation/
reforestation activities, Mitigation and Adaptation Strategies for Global Change, 11, 1003-1021.
Oliveira, P. J. C., G. P. Asner, D. E. Knapp, A. Almeyda, R. Galvan-Gildemeister, S. Keene, R. F. Raybin, and R.
C. Smith (2007), Land-Use Allocation Protects the Peruvian Amazon, Science, 317(5842), 1233-1236.
Page, S. E., F. Siegert, J. O. Rieley, H. D. V. Boehm, A. Jaya, and S. Limin (2002), Te amount of carbon released
during peat and forest ?res in Indonesia during 1997, Nature, 420, 61-65.
25
Saatchi, S. S., R. A. Houghton, R. C. Dos Santos Alvara, J. V. Soares-Filho, and Y. Yu (2007), Distribution of
aboveground live biomass in the Amazon basin. , Global Change Biology, 13(4), 816-837.
Santilli, M. P., P. Moutinho, S. Schwartzman, D. C. Nepstad, L. Curran, and C. Nobre (2005), Tropical
deforestation and the Kyoto Protocol: an editorial essay, Climatic Change, 71, 267-276.
Sathaye, J., W. Makundi, L. Dale, P. Chan, and K. Andrasko (2006), GHG Mitigation Potential, Costs and
Bene?ts in Global Forests: A Dynamic Partial Equilibrium Approach., Energy Journal, 27, 127-162.
Schlamadinger, B., N. Bird, T. Johns, S. Brown, J. Canadell, and L. Ciccarese (2007a), A synopsis of land-
use, land-use change and forestry (LULUCF) under the Kyoto Protocol and Marrakech Accords,
Environmental Science and Policy, 10 27, 1-282.
Schlamadinger, B., T. Johns, L. Ciccarese, M. Braun, A. Sato, A. Senyaz, P. Stephens, M. Takahashi, and X.
Zhan (2007b), Options for including land use in a climate agreement post-2012: improving the Kyoto
Protocol approach, Environmental Science and Policy, 10, 295-305.
Sedjo, R., B. Sohngen, and R. Mendelsohn (2001), Estimating Carbon Supply Curves for Global Forests and Other
Land Uses. Discussion Paper 01–19, Resources for the Future. , Washington, D.C.
Sedjo, R.A., and B. Sohngen (2007), Carbon credits for avoided deforestation. Washington, D.C.: Resources for the
Future.
Silva-Chavez, G. and Petsonk, A. (2006), Rainforest credits. Carbon Finance 6:18.
Skutsch, M., N. Bird, E. Trines, M. Dutschke, P. Frumho?, B. H. J. de Jong, P. van Laake, O. Masera, and D.
Murdiyarso (2007), Clearing the way for reducing emissions from tropical deforestation, Environmental
Science and Policy 10, 322-334.
Soares-Filho, B. S., D. C. Nepstad, L. M. Curran, G. C. Cerqueira, R. A. Garcia, C. A. Ramos, E. Voll, A.
McDonald, P. Lefebvre, and P. Schlesinger (2006), Modelling conservation in the Amazon basin, Nature,
440(7083), 520-523.
Sohngen, B., and R. Sedjo (2006), Carbon sequestration in global forests under di?erent carbon price regimes,
Energy Journal, 109-126.
Stern, N. (2006), Stern review on the economics of climate change, Cambridge University Press, Cambridge,
England.
Vera-Diaz, M. del C., R. K. Kaufmann, D. C. Nepstad, P. Schlesinger. 2007. An interdisciplinary model of
soybean yield in the Amazon Basin: the climatic, edaphic, and economic determinants. Ecological
Economics doi:10.1016/j.ecolecon.2007.07.015
26

doc_637793396.pdf