Sunde lin, William D. e al.
A icle — Published Ve sion
Why o es s a e impo an o global po e y alle ia ion: A
spa ial explana ion
Ecology and Socie y
P o ided in Coope a ion wi h:
Leibniz Ins i u e o Ag icul u al De elopmen in T ansi ion Economies (IAMO), Halle (Saale)
Sugges ed Ci a ion: Sunde lin, William D. e al. (2008) : Why o es s a e impo an o global po e y
alle ia ion: A spa ial explana ion, Ecology and Socie y, ISSN 1708-3087, Resilience Alliance, Wol ille,
No a Sco ia, Vol. 13, Iss. 2,
h ps://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/216812
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by-nc/4.0/
Copy igh © 2008 by he au ho (s). Published he e unde license by he Resilience Alliance.
Sunde lin, W. D., S. Dewi, A. Pun odewo, D. Mülle , A. Angelsen, and M. Epp ech . 2008. Why o es s a e
impo an o global po e y alle ia ion: a spa ial explana ion. Ecology and Socie y 13(2): 24. [online]
URL: h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Resea ch
Why Fo es s A e Impo an o Global Po e y Alle ia ion: a Spa ial
Explana ion
William D. Sunde lin 1,2, Sonya Dewi 3, A ie Pun odewo 2, Daniel Mülle 4,5, A ild Angelsen 2,6, and
Michael Epp ech 7
ABSTRACT. Fo es s ha e been decla ed impo an o he well-being o he poo because o he kinds o
goods and se ices ha hey p o ide. We asked whe he o es s a e impo an o he poo no only because
o he kinds o goods and se ices hey p o ide, bu also because hey end o be loca ed whe e he poo
a e. We conduc ed a spa ial analysis o asce ain he deg ee o spa ial associa ion be ween po e y and
o es s in se en coun ies: B azil, Hondu as, Malawi, Mozambique, Uganda, Indonesia, and Vie nam. Fo
mos o hese coun ies, he e was a signi ican posi i e co ela ion be ween high na u al o es co e and
high po e y a e ( he pe cen age o he popula ion ha is poo ) and be ween high o es co e and low
po e y densi y ( he numbe o poo pe uni a ea). We explain he indings and discuss policy implica ions
and opics o u u e esea ch.
Key Wo ds: de o es a ion; o es ; po e y; spa ial analysis
INTRODUCTION
The poo a e said o ely on o es s o main ain hei
well-being (By on and A nold 1999, Angelsen and
Wunde 2003, Vedeld e al. 2004) and in some cases
as a sou ce o income imp o emen (e.g., Ruíz-
Pé ez e al. 2004). Fo es s p o ide a “subsidy om
na u e” o ee en i onmen al goods and se ices
(Ande son e al. 1991). These goods and se ices
include a wide di e si y o p oduc s o home
consump ion and sale, new ag icul u al lands,
es o a ion o soil e ili y on allow lands used in
swidden cul i a ion cycles, and access o esh wa e
h ough he wa e shed unc ion o o es s (Sunde lin
e al. 2005). Fo es s can ul ill a ious oles in he
li elihoods o he u al poo . They p o ide a sou ce
o egula subsis ence o people who li e in and
nea o es s in he o m o ood, uel, o age,
building ma e ials, and medicines, among o he
p oduc s (By on and A nold 1999). The
di e si ica ion o income sou ces is impo an o he
u al poo as a means o minimize isk exposu e
(Ellis 2000). Fo es esou ces p o ide a sou ce o
gap- ille income in be ween ag icul u e ha es s
and se e as a sa e y ne in he e en o p i a e
calami y such as a dea h in he amily o communi y
calami y such as ailed ha es , amine, lood,
hu icane, o wa (Angelsen and Wunde 2003,
Takasaki e al. 2004). Fo es s can unc ion as a o m
o sa ings (Chambe s e al. 1993) and as a o m o
na u al insu ance (Pa anayak and Sills 2001,
McSweeney 2004). The open-access cha ac e o
wide a eas o public o es s in de eloping coun ies
assu e ha he poo a e o en ela i ely unhinde ed
in making use o o es esou ces (Sunde lin e al.
2005). All o hese documen ed bene i s sha e a ai
in common: They all conce n wha o es s a e o
poo people.
He e, we ask whe he o es s a e c i ical o
suppo ing he well-being o he poo , no jus
because o wha hey a e, bu also because o whe e
hey a e. S a ed di e en ly, we ask i he p o-poo
cha ac e is ics o o es s a e meaning ul because he
poo and o es s end o sha e he same space.
Wha kind o e idence is he e ha he poo and
o es s occupy he same space? Va ious coun y-
le el obse a ions ha e been made o he endency
o he u al poo o be loca ed in o nea by o es s.
1Righ s and Resou ces G oup, 2Cen e o In e na ional Fo es y Resea ch, 3Wo ld Ag o o es y Cen e, 4Leibniz Ins i u e o Ag icul u al De elopmen in
Cen al and Eas e n Eu ope (IAMO), 5Humbold -Uni e si ä zu Be lin, 6No wegian Uni e si y o Li e Sciences, 7Swiss Na ional Cen e o Compe ence in
Resea ch (NCCR) "No h-Sou h", Uni e si y o Be ne
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
In China, he e is an obse ed o e lap be ween
coun ies ca ego ized as se e ely poo and coun ies
wi h abundan o es esou ces (Zhou and Veeck
1999). In India, Po enbe ge e al. (1996, as ci ed
in Kha e e al. 2000) ound ha he e is a s ong
associa ion be ween he loca ion o ibal people,
who end o be among he poo es people, and he
loca ion o o es s. App oxima ely 275 million
people in India’s u al a eas depend on o es s o a
po ion o hei income; o es dwelle s, who a e
disp opo iona ely ibal, a e among he poo es and
mos ulne able people in India (Wo ld Bank 2006).
App oxima ely hal o India’s 350 million poo
people a e concen a ed in h ee s a es in which
na u al, physical, social, and human capi al a e low
and he g ea es po e y is expe ienced by people
in o es -based economies; 84% o India’s ibal
e hnic mino i ies li e in o es ed a eas (Meh a and
Shah 2003). Shah and Gu u (2004:8) explain ha
he “incidence o po e y, e lec ed by head coun
a io, is highe han he all-India es ima es o he
majo i y o o es based s a es. Compa ed o his,
he incidence o po e y is subs an ially lowe
among d yland s a es excep Maha ash a. The
pa e n is mo e o less he same du ing 1993–1994
and 1999–2000.”
Fi e pieces o ecen esea ch deepen he
unde s anding o he spa ial coincidence o po e y
and o es s by quan i ying he pa e ns,
dis inguishing po e y by ype (i.e., a e and
densi y), and discussing he ac o o dis ance om
u ban a eas. The po e y a e is he p opo ion o
people who a e poo in a gi en a ea; he po e y
densi y is he absolu e numbe o poo pe uni a ea
(indi iduals/km²) in a gi en a ea. In a case s udy o
Nica agua, Chomi z (K. C. Chomi z unpublished
manusc ip ) ound ha “ emo eness om owns and
ma ke s is associa ed wi h high po e y a es, high
o es co e , and low popula ion densi y.”
Simila ly, in a case s udy o Vie nam, Mülle e al.
(2006) demons a ed, h ough bi a ia e local
indica o s o spa ial associa ion analysis, a s ong
s a is ical co ela ion among he loca ion o low
po e y densi y, high po e y a e and se e i y, and
high o es co e in emo e a eas whe e e hnic
mino i ies end o be dominan . Con e sely, hey
ound high co ela ion among a eas o high po e y
densi y, low po e y a e, and low o es co e in
u ban and pe i-u ban a eas and in he Mekong and
Red Ri e del as. Dasgup a e al. (2005) ound ha
po e y densi y in Cambodia ends o be high in
a eas o low o es co e and low in a eas o high
o es co e . D awing on Sunde lin e al.’s (2007)
case s udy wo k, Chomi z (2007) s eng hens he
case ha emo eness and o es s a e posi i ely
ela ed o po e y a e and in e sely ela ed o
po e y densi y.
We add essed he ollowing ques ions. Is he e a
clea associa ion be ween po e y and o es s in
de eloping coun ies on h ee con inen s? I so,
wha a e he cha ac e is ics o ha associa ion? To
answe hese ques ions, we conduc ed spa ial
analysis o se en coun y cases: B azil, Hondu as,
Malawi, Mozambique, Uganda, Indonesia, and
Vie nam. We hen discuss he policy implica ions
o he indings, gi ing a en ion o he way he
indings can be implemen ed in na ional po e y
educ ion s a egies. These a e impo an objec i es
because o es esou ces a e o en gi en li le, i any,
a en ion in na ional s a egies o po e y educ ion
(e.g., Oksanen and Me smann 2003).
METHODS
C i e ia o he selec ion o he se en case
s udies
Ini ially, we sough o analyze as many de eloping
coun y cases as possible ha co e ed he la ges
possible numbe o people li ing in and nea o es s
and he la ges a ea o o es s. In his way, ou
sample would be b oadly ep esen a i e o o es s
o de eloping coun ies as a whole. We sough
coun y cases ha me he ollowing minimum
equi emen s: he o es co e was non-negligible;
ela i ely ecen and eliable popula ion and po e y
da a we e a ailable a a ela i ely ine le el o
disagg ega ion; and he popula ion and po e y da a
we e speci ied a he le el o he dis ic and could
be a ibu ed o a geo- e e enced dis ic map.
In seeking coun y cases ha me hese minimum
da a equi emen s, we lea ned ha he e we e e y
ew such coun ies. Because o una ailable da a o
da a ha did no mee ou speci ica ions, i was no
possible o include ce ain coun ies ha we would
ha e liked o include such as he impo an o es
coun ies o Mexico, Came oon, Democ a ic
Republic o Congo, and Papua New Guinea and
coun ies ha had la ge popula ions such as China
and India. Howe e , based on ou c i e ia, we chose
o analysis se en coun ies om h ee con inen s:
B azil, Hondu as, Malawi, Mozambique, Uganda,
Indonesia, and Vie nam.
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Po e y and popula ion da a
The po e y and popula ion da a we e based la gely
on o icial na ional censuses and su ey da a. Fou
o he se en da a se s (Malawi, Mozambique,
Uganda, and Vie nam) we e based on small a ea
es ima es (Benson 2002, Mino e al. 2003,
In e na ional Li es ock Resea ch Ins i u e 2005,
Simle e al. 2005). This app oach in ol es indi ec
es ima ion p ocedu es ha combine he spa ial
p ecision o census da a wi h he subs an i e dep h
o su eys. Some o he da a se s we e assembled
by na ional ins i u ions (Badan Pusa S a is ik 2000,
Robles 2003; Ins i u o de Pesquisa Econômica
Aplicada na ional po e y da a o he yea 2000:
h p://ipeada a.go .b /), and o he s we e c ea ed by
na ional go e nmen ins i u ions collabo a ing wi h
in e na ional esea ch o ganiza ions (e.g., he
In e na ional Food Policy Resea ch Ins i u e in
Malawi and Vie nam and he In e na ional
Li es ock Resea ch Ins i u e in Uganda). In all
cases, he numbe s o poo we e de e mined by
iden i ying households ha we e below a coun y-
speci ic po e y line. The e is a wide di e si y o
app oaches in de e mining po e y lines in hese
a ious coun ies.
Fo es co e da a
Fo es co e da a we e ob ained om he
Con inuous Fields 1 km T ee Co e map p oduced
by DeF ies e al. (2000a,b) om Ad anced Ve y
High Resolu ion Radiome e (AVHRR) images o
he pe iod o 1992–1993. This map p o ides
es ima es o o es canopy densi y as a pe cen age
o a ea wi hin 1-km² cells. Following he Food and
Ag icul u e O ganiza ion’s (2001) model o o es
co e classi ica ion, we p oduced a new map
acco ding o he ollowing ange o alues: 0–10%
ee c own co e = non o es ; 10–40% = open o es ;
and 40–100% = closed o es . The coa se- esolu ion
T ee Co e map was used despi e some limi a ions
because i p o ides o es co e da a o all o he
coun y cases and because he uni o mi y o he da a
collec ion me hod and o he o es class de ini ions
allows a deg ee o compa abili y among he
coun ies s udied.
Pea son co ela ion, Mo an’s I, and local
indica o s o spa ial associa ion
We used Pea son co ela ion o see i he e we e any
gene al, na ional pa e ns o ela ionships be ween
po e y and o es s. This app oach add essed spa ial
a ia ion, bu no spa ial dependence. Thus, we used
Mo an’s I o measu e spa ial au oco ela ion and
iden i y spa ial clus e s in he da a. The uni a ia e
global Mo an’s I measu es he linea associa ion o
a a iable wi h i sel in space and was used o es
he s eng h o global spa ial au oco ela ion o
po e y and o o es . Mo an’s I anges be ween −1
and +1. A posi i e Mo an’s I indica es spa ial
clus e ing o ei he high o low alues, whe eas a
Mo an’s I nea ze o implies no spa ial
au oco ela ion, o spa ial andomness. A
checke boa d pa e n o dissimila alues esul s in
a Mo an’s I close o −1. Mo an’s I is weigh ed by a
ma ix ha e lec s he spa ial p oximi y used o
de ine he neighbo hood (Anselin 1988).
We explo ed he ela ionships be ween po e y and
o es a he na ional le el aking in o accoun spa ial
dependence using bi a ia e global Mo an’s I.
Fu he , local Mo an’s I, also called local indica o s
o spa ial associa ion (LISA) analysis, was
conduc ed o iden i y egions wi hin a coun y
whe e he e we e di e en pa e ns o ela ionships
be ween po e y and o es .
The bi a ia e local spa ial au oco ela ion measu e
using Mo an’s I s a is ic, de i ed om Anselin’s
o mula (1995), is w i en as:
(1)
whe e x and y a e he po e y and o es a iables
o dis ic i and he neighbo ing dis ic j,
espec i ely; and zx and zy a e he s anda dized z-
sco es o a iables x and y, espec i ely. The spa ial
weigh ma ix wij can be de ined as a bina y
con igui y ma ix ha p o ides he spa ial s uc u e
o he loca ions ha a e included in he calcula ion
o he local Mo an’s I. In he case o i s -o de
con igui y weigh s, all obse a ions ha sha e a
common bo de ha e wij = 1, o he wise wij = 0. A
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
i s -o de queen con igui y ma ix de ined he
spa ial ex en o he obse a ions ha we used. The
queen c i e ion conside s all su ounding neighbo s
(dis ic s) ha ha e a common bo de o common
e ices wi h he obse a ion o in e es . The ma ix
choice is based on logical insigh s and he e o e
in ol es an a bi a y componen . The diagonal
elemen s in he ma ix a e ze o and he o -diagonal
elemen s ep esen he speci ied neighbo hood
a ound obse a ion i. The a e age o all local
Mo an’s I indices o e all obse a ions (dis ic s) N
is p opo ional o he global Mo an’s I (Anselin
1995).
In o he wo ds, Eq. 1 indica es spa ial clus e s o
high (o low) alues o one a iable x su ounded
by high (o low) alues o ano he a iable y. High
and low alues in his con ex a e de ined as alues
ha a e g ea e han and less han he mean o he
espec i e a iable, espec i ely. This p o ides a
measu e o he spa ial associa ion be ween wo
a iables in a neighbo hood ha is de ined by he
spa ial weigh s ma ix.
All LISA indica o s we e calcula ed using GeoDaTM
(Anselin 2005). To c ea e he maps, a andomiza ion
app oach o 999 pe mu a ions was used o a oid
la ge sensi i i ies in he esul s. The andomiza ion
assumes he loca ion o he alues and hei spa ial
a angemen o be i ele an . Based on he
andomiza ion, di e en heo e ical s anda d
de ia ions o Mo an’s I a e ob ained, each yielding
a di e en p- alue as a pseudo-signi icance. The
h eshold alue o p = 0.05 de ines signi icance. The
p- alues ollow an asymp o ically s anda d-no mal
dis ibu ion ha allows he e alua ion o hei
signi icance le el by compa ing hem o a e e ence
dis ibu ion (Anselin 1995).
Applica ion o local indica o s o spa ial
associa ion
We sough o unde s and how a ying magni udes
o po e y (i.e., low o high po e y a e o densi y)
a e ela ed spa ially o a ying p opo ions o o es
co e (i.e., low o high) a he le el o he dis ic
wi hin he se en case s udy coun ies. Again, he
po e y a e is he p opo ion o people who a e poo
in a gi en a ea, and he po e y densi y is he
absolu e numbe o poo people in a gi en a ea.
Fo es co e is he p opo ion o closed o es (40–
100% canopy co e ) pe uni dis ic a ea.
Measu es o po e y and o es co e each ha e he
po en ial o spa ial au oco ela ion. Fo example,
uni a ia e Mo an’s I calcula ions sugges ed ha
simila pe cen ages o o es co e end o be
clus e ed in a ea uni s ha a e close o one ano he .
In applying he bi a ia e global Mo an’s I s a is ic
o he case s udy da a, au oco ela ions o po e y
and o es da a we e disce ned a he coun y le el.
In applying he bi a ia e LISA me hod, spa ial
clus e s we e p oduced ha isualized he
ela ionships be ween wo a iables: he co ela ion
o po e y a e and o es co e in he su ounding
a ea; and he co ela ion o po e y densi y and
o es co e in he su ounding a ea.
The bounda ies be ween high (H) and low (L)
po e y a e and be ween high (H) and low (L)
po e y densi y a e de e mined by he LISA analysis
in each coun y calcula ion. So, o example, o
dis ic s in which co ela ions o o es co e and
po e y a e a e signi ican , he LISA clus e maps
display signi ican co ela ions using he ollowing
labels: HIGH-HIGH (HH) = high o es co e and
high po e y a e; LOW-LOW (LL) = low o es
co e and low po e y a e; HIGH-LOW (HL) =
high o es co e and low po e y a e; LOW-HIGH
(LH) = low o es co e and high po e y a e.
In he LISA analysis, he ca ego iza ion in o hese
ou g oups was conduc ed based on h ee c i e ia.
Fo example, o dis ic i o be classi ied as HL, he
ollowing h ee condi ions mus hold: he bi a ia e
local Mo an’s Ii is signi ican ly di e en om ze o
and nega i e, o , i i is HH o LL, hen i is
signi ican ly di e en om ze o and posi i e; he
o es co e o dis ic i is g ea e han he na ional
mean o es co e ; and he mean po e y o he
dis ic s neighbo ing dis ic i is less han he
na ional mean po e y. These ules apply simila ly
whe he he po e y measu e is a e o densi y.
Each o he igu es ha displays he indings shows
ou maps: o es co e and majo u ban a eas;
popula ion densi y; LISA analysis o o es co e
and po e y a e; and LISA analysis o o es co e
and po e y densi y. The jux aposi ion o he maps
showing u ban a eas and popula ion densi y (maps
1 and 2) wi h he maps showing he LISA esul s
(maps 3 and 4) is impo an because i helps o show,
albei only isually, he ela ionship be ween
emo eness, de ined simply as dis ance om u ban
cen e s, and he pa e ns o associa ion be ween
po e y and o es s as indica ed by he LISA
analysis.

Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
S eng hs and weaknesses o he app oach
The main jus i ica ions o measu ing and es ing
global and local spa ial au oco ela ion in s udying
he ela ionship be ween po e y and o es co e
a e as ollows. The global Mo an’s I echnique
allows he e alua ion o he hypo hesis ha spa ial
au oco ela ion is impo an in assessing he
coun y-le el ela ionship be ween po e y and
o es s. We assume no only ha posi i e spa ial
au oco ela ion will be disce nible in each
uni a ia e case o po e y and o es co e (i.e., ha
bo h po e y and o es co e a e spa ially clus e ed
in some places), bu also ha he e is spa ial
a ia ion in he ela ionship be ween o es co e
and po e y. The LISA me hod p o ides an e icien
way o de ec and isualize co ela ions a he local
le el ha a e no eadily disce nible when pooling
da a a he na ional le el.
Al hough we hink he me hods we used a e
app op ia e o ou objec i es, he e a e impo an
d awbacks and limi a ions ha should be
conside ed. Fi s , he examina ion o jus wo
a iables, i.e., po e y and o es , uns he isk o
p oducing o e simpli ied and pe haps e en
misleading conclusions abou hei ela ionship o
one ano he . The e a e nume ous a iables ha can
a ec he ela ionship be ween po e y and o es
co e , among hem: he deg ee o emo eness om
u ban a eas; he p esence o absence o oads and
ma ke s; opog aphy; and soil quali y. Ne e heless,
i is use ul o measu e he s eng h o he associa ion
be ween po e y and o es s o es and e i y
anecdo al in o ma ion abou hei spa ial ela ionship
and o se he s age o s udies ha use a mul i a ia e
app oach. We chose o a oid mul i a ia e analyses
a his s age because we wan ed o know o wha
ex en he use o jus wo a iables could explain
he la gely anecdo al analysis o he po e y– o es
spa ial link o da e. Mo eo e , i would ha e been
challenging o conduc mul i a ia e analyses
because o he lack o da a laye s ha could be
applied ac oss all case coun ies.
Second, because o he lack o ime se ies da a on
po e y and o es co e , we es ic ed ou sel es o
a s a ic measu e o he ela ionship be ween he wo
a iables. This es ic s ou abili y o d aw de ini i e
conclusions abou cause and e ec pa e ns be ween
he wo, and also es ic s us o educa ed guesses
abou whe e he ela ionship is heading o e ime.
Thi d, AVHRR da a a e coa se, isking inaccu acy
in cap u ing ege a ion e lec ance. Depending on
he landscape pa e ns, he unde - and o e es ima ion
o ege a ion densi y can be signi ican . This can be
especially p oblema ic in a eas o he case coun ies
ha ha e ample o es –sa annah ansi ion zones o
highly agmen ed mosaic- ype landscapes.
Fou h, he e is a empo al gap be ween he AVHRR
da a (1992–1993) and he socioeconomic da a
(oldes 1998, newes 2001). In some coun ies such
as Indonesia, he e has been ne de o es a ion du ing
his empo al gap, whe eas in o he s such as
Vie nam, ne o es co e has ei he been s able o
has inc eased sligh ly.
The i h p oblem conce ns he a bi a iness o he
cu -o poin s in de ining o es as closed o es , ha
is, 40–100% c own co e , as pe he Food and
Ag icul u e O ganiza ion’s (2001) sys em o
classi ica ion. Ou de ini ion excludes wo ypes o
land co e ha a e o po en ial ele ance. Fi s , his
na ow de ini ion o o es excludes open o es
(10–40% c own co e ). Thus, in d y and sub opical
coun ies (e.g., he Miombo woodlands coun ies
such as Angola, Bu undi, Democ a ic Republic o
Congo, Malawi, Mozambique, Tanzania, Zambia,
and sou he n Zimbabwe), la ge a eas o spa se
woodlands ha can be p ope ly called o es s and
upon which la ge segmen s o he popula ion
depend o hei li elihoods will no be conside ed
i he c own co e is < 40%. Second, his na ow
de ini ion o o es omi s om he o es ca ego y
la ge a eas o sh ubs and o he ege a i e co e in
he humid opics ha may no be conside ed as
o es s in he opical con ex , al hough hey may
play simila oles o o es s in local li elihoods. Ou
de ini ion o o es mo e s ongly equa es wi h he
co e a ibu es o emo eness (i.e., dis ance om
u ban a eas and oads, p esence o indigenous
people, lack o ma ke access), making some o ou
obse a ions abou he link be ween emo eness,
po e y a e, and o es s s and ou in bolde elie .
We expe imen ed wi h he app oach o de ining
o es ei he na owly o widely. In he end, we had
o choose one de ini ion o he o he o a oid he
awkwa dness o applying di e en de ini ions o
o es ac oss he se en case coun ies and concluded
ha he bene i s o he na ow de ini ion ou weighed
he disad an ages.
Finally, we mus acknowledge ha he compa isons
and con as s among he case coun ies a e c ude.
Al hough he use o AVHRR da a helps o
s anda dize o es co e measu es among coun ies,
he sho comings o AVHRR da a end o weaken
he basis o compa ison. The basis o compa ison
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
is also weakened by he ac ha me hods o
measu ing po e y di e among he coun ies.
Ne e heless, i one assumes ha all coun ies use
b oadly simila assump ions abou wha cons i u es
adequa e o inadequa e human well-being, he
compa isons a e a leas minimally legi ima e and
use ul.
RESULTS
Pea son co ela ion and Mo an’s I
The Pea son co ela ions and Mo an’s I sugges an
ambiguous ela ionship be ween po e y and
o es s. The Pea son co ela ions and bi a ia e
Mo an’s I indica e ha dis ic -le el co ela ions
be ween po e y a e and o es co e we e low a
he na ional le el, wi h small co ela ion
coe icien s o mos cases, e en hough hey we e
signi ican a p ≤ 0.05 in i e cases (Table 1). In
hese calcula ions, he B azil and Uganda cases
show an unexpec ed nega i e co ela ion. In he case
o B azil, his is because he dis ic s classi ied as
high po e y-low o es co e (HL) and low
po e y-high o es co e (LH; 1772 dis ic s)
ou numbe he dis ic s classi ied as HH and LL
(1280 dis ic s); howe e , he HH and LL dis ic s
compose a much la ge a ea han do he HL and LH
dis ic s. The po e y a e global Mo an’s I shows
medium o s ong co ela ion and signi icance a p
= 0.01 o all coun ies. The o es co e global
Mo an’s I shows medium o s ong co ela ion o
all case coun ies, and signi icance a p ≤ 0.05 o
all se en coun ies.
Maps o local indica o s o spa ial associa ion
By and la ge, he maps o local indica o s o spa ial
associa ion (LISA) ended o con i m ea lie case
s udies ha demons a ed a spa ial associa ion
be ween po e y and o es s (Figs. 1–7). Howe e ,
some o he se en case coun ies con o med o his
pa e n, whe eas o he s did no . The pa e n is
disce ned isually by compa ing he a ea o high
o es co e o po e y a e and po e y densi y.
In mos o he se en case coun ies, he e was a
b oad endency o high o es co e o be associa ed
wi h high po e y a e, indica ed by ed shading in
he o es s. po e y a e map (Figs. 1–7). This a ea
o ed shading ends o con o m o he a ea o closed
o es in he o es and majo u ban a eas map. This
associa ion is pa icula ly s ong in he case o B azil
and Vie nam; e iden , bu weake in he case o
Malawi, Mozambique, and Indonesia; and no
e iden in he case o Hondu as and Uganda. We
no e, howe e , ha one eason why he e ec
appea s pa icula ly s ong in B azil is because he
o es ed dis ic s a e qui e la ge and he e o e end
o domina e he image.
The e is also a b oad endency o high o es co e
o be associa ed wi h low po e y densi y, indica ed
by pink shading in he o es s. po e y densi y map
(Figs. 1–7). This associa ion is disce nible o
app oxima ely h ee-qua e s o he o es a ea in
B azil, Indonesia, and Vie nam; app oxima ely hal
o he o es a ea in Hondu as, Malawi,
Mozambique; and only a small a ea o he o es in
Uganda.
Gi en ha la ge u ban a eas in de eloping coun ies
will o en ha e ela i ely li le o es co e and ha
some a e some imes cen e s o economic
dynamism, one can expec a eas o low o es co e
and low po e y a e in some cases. We see such
a eas, indica ed by da k blue shading in he o es
s. po e y a e map (Figs. 1–7), in B azil (B asilia),
Malawi (Lilongwe), Mozambique (Mapu o),
Indonesia (Jaka a), and Vie nam (Hanoi and Ho
Chi Minh Ci y). Mos ci ies, howe e , ei he de ia e
om his pa e n o show no signi ican associa ion.
Ou indings e eal he impo ance o emo eness:
high popula ion densi y ends o be cen e ed in and
a ound u ban a eas, and o es co e ends o be he
mi o image o popula ion densi y. Mozambique
and Uganda a e excep ions, wi h signi ican
o e laps o high popula ion densi y and high o es
co e . Simila ly, a eas o high o es co e and high
po e y a e and a eas o high o es co e and low
po e y densi y end o be loca ed away om big
ci ies and in emo e places.
O e all indings
We used he da a om he o es s. po e y a e
map o each coun y (Figs. 1–7) and classi ied hem
acco ding o he pe cen age a ea o closed o es
(Fig. 8) and he pe cen age o he o al numbe o
poo people (Fig. 9) in each coun y. This allowed
us o de e mine he ele ance o he indings in e ms
o ou key objec s o conce n: he a ea o closed
o es and he numbe o poo people.
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Table 1. Pea son co ela ions and bi a ia e and uni a ia e global Mo an’s I calcula ions o po e y a e
and o es co e da a om se en case s udy coun ies.
Coun y No. dis ic s Mean dis ic
a ea (km²) SD o
dis ic a ea
(km²)
Pea son co el-
a ion coe icien
Bi a ia e g-
lobal Mo an’s
I
Po e y a e
global Mo an’s
I
Fo es co e
global Mo an’s
I
B azil 5507 1553.34 5746.82 −0.1139** −0.0983** 0.4842** 0.8554**
Hondu as 292 383.94 914.27 0.1250* 0.0588 0.4586** 0.3902**
Indonesia 351 5424.34 10652.99 0.2780** 0.1268** 0.5015** 0.6609**
Malawi 368 260.22 360.68 0.2142** 0.0895** 0.3658** 0.3928**
Mozambique 146 5359.85 4078.16 0.0428 0.0664 0.3510** 0.4809**
Uganda 34 1409.10 1013.26 −0.3641 −0.2236* 0.4415** 0.3093*
Vie nam 601 554.51 559.34 0.5710** 0.4365** 0.6843** 0.7538**
*p < 0.05, **p < 0.01.
In he case o B azil and Vie nam, sligh ly > 70%
o he dis ic s ha con ained high o es co e also
had a high a e o po e y (Fig. 8, da k ed); he
emainde o he dis ic s wi h closed o es a ea (<
30%) comp ised all o he clus e s. Con e sely,
Hondu as showed only a small po ion o closed
o es a ea (~10%) in which high o es co e was
associa ed wi h a high po e y a e; ~40% o i s
closed o es is cha ac e ized by no s a is ical
co ela ion be ween o es co e and po e y a e a
he dis ic le el. Mos case coun ies we e be ween
hese wo ex emes. On a e age, ~40% o he o al
closed o es a ea in he se en case coun ies was
associa ed wi h a high po e y a e. On a e age,
almos hal o he o es a ea showed no signi ican
ela ionship be ween high o es co e and po e y
a e (Fig. 8, yellow).
Only a small pe cen age o he coun ies’ poo
people li e in a eas cha ac e ized wi h high o es
co e and high po e y a e (Fig. 9, da k ed). The
ange was om a low o ~3% o Uganda and
Indonesia o ~20% o Vie nam. Fo all coun ies,
wi h he excep ion o Mozambique, he e we e as
many o mo e poo people in he low o es co e -
high po e y a e zone (Fig. 9, ligh blue) as he e
we e in he high o es co e -high po e y a e zone
(Fig. 9, da k ed). Fo somewha mo e han hal o
all poo , he e is no signi ican ela ionship be ween
po e y a e and o es co e (Fig. 9, yellow).
Reasons o he coincidence o po e y and
o es s
Why a e many o he people who li e in o es ed
a eas poo , and why do he a es o po e y in
o es ed a eas end o be high? One eason is ha
po e y a es in de eloping coun ies end o be
highe in u al han in u ban a eas (De con 2008)
and o es s end o be mo e abundan in u al han
in u ban a eas. Does his mean ha he supposed
po e y– o es link is no hing mo e han a disguised
po e y– u ali y link? We hink no . The e is a
spa ial dimension ha mus be aken in o accoun .
Ch onic po e y and po e y a es end o be highe
in emo e u al a eas, which a e some imes e e ed
o as spa ial po e y aps (Jalan and Ra allion 1997,
Bi d e al. 2002, Kanbu and Venables 2003), and
o es s, pa icula ly p ima y o es s, end o be
dense in emo e u al a eas. Why would o es s, as
a componen o emo e u al a eas, end o ha e high
po e y a es? We see se en easons:
1. P imo dial low income in o es ed a eas.
Many cen u ies ago, s anda ds o li ing we e
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Fig. 1. B azil. Maps o o es and majo u ban a eas, popula ion densi y, local indica o s o spa ial
associa ion (LISA) analysis o o es co e by po e y a e, and LISA analysis o o es co e by po e y
densi y. Dis ic bounda ies a e no displayed o show he colo pa e ns mo e clea ly.
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Fig. 8. Dis ibu ion o local indica o s o spa ial associa ion o es s. po e y a e indings by
pe cen age o closed o es a ea.
p obably no be among he poo es o he poo
because i equi es sa ings o unde ake
mig a ion. Some easons o mig a ing o he
o es a e, howe e , no economic. Fo es s
ha e a high p esence o poo people in some
coun ies because hey ha e se ed as a
e uge o people leeing pe secu ion,
con lic , and wa ;
7. Low in es men in emo e a eas. In some
de eloping coun ies, he a e o go e nmen
in es men in emo e u al a eas is low (e.g.,
he lowe Amazon: Fu emma and B ondízio
2003). One likely explana ion is ha he e a e
low e u ns o in es men in emo e a eas. Fo
example, a gi en amoun o in es men
se es many ewe people in u al a eas han
in nea -u ban a eas. Ano he explana ion
ela es o he powe lessness o people in
o es ed a eas. Because o hei ela i e
powe lessness, o es dwelle s o en do no
ha e he ba gaining powe equi ed o ge a
a o able decision om people in ci ies who
make decisions on public in es men s.

Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Fig. 9. Dis ibu ion o local indica o s o spa ial associa ion o es s. po e y a e indings by
pe cen age o poo people.
DISCUSSION
We hink ha much mo e a en ion o o es s is
dese ed in na ional po e y alle ia ion s a egies
o wo easons. Fi s , despi e he compa a i ely low
po e y densi ies ha we iden i ied, he e a e many
poo people who li e in o es en i onmen s.
Second, he e a e nuances o o es po e y ha
equi e speci ic a en ion o be add essed
sa is ac o ily.
Many poo people li e in o es en i onmen s
Supe icially, he indings seem o say ha na ional
po e y educ ion s a egies a e jus i ied in gi ing
scan a en ion o o es s. A e all, on a na ional
scale, a small p opo ion o he poo end o li e in
o es s. Howe e , he e is one impo an way in
which he local indica o s o spa ial associa ion
esul s se iously unde s a e he spa ial associa ion
be ween he numbe s o poo people and nea by
o es s.
The e a e places in de eloping coun ies whe e
o es co e is low (< 40% canopy co e ) and ye
o es esou ces can be a signi ican sou ce o
household income in hese a eas. I any hing, o es
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
esou ces can be seen as mo e impo an o po e y
alle ia ion in some o hese a eas han in dense
o es s because a la ge numbe o people li e in such
o es s and o es s in hese open o es a eas a e
concei ably supplying basic needs o a la ge
numbe o people. Such places include he sa annah
and woodlands o he Ce ado in eas e n B azil (Fig.
1), he Miombo Woodlands in a ious pa s o
Malawi and Mozambique (Figs. 3 and 4), and some
d y o es a eas in eas e n Indonesia (Fig. 6).
Ou side o ou case s udy coun ies, his
conside a ion is impo an in o he Miombo
coun ies, i.e., Angola, Bu undi, Democ a ic
Republic o Congo, Tanzania, Zambia, and
Zimbabwe.
How many poo people li e in all o es
en i onmen s? As ye he e a e no eliable
es ima es. Al hough he p opo ion o all poo
people li ing in a eas o high o es co e may be
low (Fig. 9), he absolu e numbe s a e high in some
coun ies. I only 10% o all poo in China and India
li e in a eas o high o es co e (no o men ion
a eas o low o es co e ), his would amoun o >
50 million people.
Nuances o o es po e y ha equi e speci ic
a en ion
A high po e y a e is o en linked wi h a high dep h
and se e i y o po e y. The Fos e , G ee , and
Tho becke (FGT) class o po e y measu es (Fos e
e al. 1984) de ines he dep h o po e y (α = 1) as
he a e age dis ance o he poo below he po e y
line (i.e., he income gap). The se e i y o po e y
(α = 2) is he a e age o he squa e o he dep h o
po e y, i.e., gi ing mo e weigh o he e y poo es .
Va ious s udies show ha he a e, dep h, and
se e i y o po e y a e o en highe in u al han in
u ban a eas (Odu o and A yee 2003, Shimeles and
Thoenen 2005). Likewise, ch onic po e y has been
obse ed o be a co ela e o high po e y a e,
dep h, and se e i y, bu his endency canno be
assumed in all cases and emains an empi ical
ques ion in need o e i ica ion (Hulme 2003,
Hulme and Shephe d 2003). Ch onic po e y is
de ined as occu ing “...when an indi idual
expe iences signi ican capabili y dep i a ions o
a pe iod o i e yea s o mo e” (Hulme and Shephe d
2003:404–405). Ch onic po e y ends o be
p onounced in emo e u al a eas and moun ainous
a eas (McKay and Lawson 2003). This may explain
he obse ed co ela ion be ween high po e y a e,
dep h, and se e i y on he one hand, and ch onic
po e y on he o he . Thus, i he e is a na ional
p og am o e adica e he nodes o po e y ha a e
he mos di icul o add ess, hen i may make sense
o a ge a eas o high o es co e , among o he
emo e a eas, because hei inhabi an s a e likely o
exhibi hese kinds o po e y.
The high po e y a e cha ac e is ic o wide a eas
o o es in some coun ies (e.g., B azil and
Vie nam; Fig. 8) can lend i sel o some e iciency
gains in add essing po e y. Al hough i is
undoub edly challenging o educe po e y in
emo e o es ed a eas, wi h a high cos o public
in es men pe capi a among o he p oblems, he e
will a leas be low a es o leakage (i.e., bene i s
abso bed by he nonpoo ) in such a eas.
Wha kind o po e y alle ia ion policies should be
pu sued in ela ion o o es esou ces? I is no
app op ia e o add ess his ques ion in dep h he e,
bo h because i is unwise o gene alize om ou
indings o pa icula coun ies and because he e is
al eady ample li e a u e on his opic. Ne e heless,
we pu o wa d some hough s on his subjec .
The e is likely a ela i ely high dependence on
o es s o li elihoods in a eas o high o es co e
and high po e y, so i may make sense o p omo e
he use o o es esou ces as a pa o he po e y
alle ia ion s a egy in hese places. P omising
oppo uni ies o his kind occu in places whe e
non o es li elihood al e na i es a e ew, e.g.,
mig a ion is di icul o no sough o he ag icul u al
land quali y is low, and whe e he sus ainable use
o o es esou ces is possible, among o he
p econdi ions.
Howe e , i is impo an o bea in mind ha e en
lands wi hou o es s a e po en ially impo an o
ee-based po e y alle ia ion s a egies. This is
because some a eas wi hou o es , especially pe i-
u ban zones, a e o en wood de ici a eas whe e he
poo can base hei li elihoods on supplying he
need o imbe , i ewood, cha coal, and wood c a s
h ough smallholde plan a ions and small-scale
indus ies. The poo who li e in a eas o low o es
co e and high po e y a e (Fig. 9, ligh blue) a e
po en ial bene icia ies o in es men s in such
s a egies.
We men ioned ha high a es o po e y in o es ed
a eas a e in pa he p oduc o geog aphical
emo eness. In hese places, i is app op ia e o
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
pu sue po e y alle ia ion s a egies ha a e
speci ically gea ed o o e come he disad an ages
imposed by emo eness, o example, he
cons uc ion o oads and b idges, in oduc ion o
ma ke s, and acili a ion o ou mig a ion i i is
sough . Ne e heless, hese policies may be
insu icien i hey do no ake in o accoun o es
esou ce dependence and he compa a i e
powe lessness o o es dwelle s in emo e a eas.
To assis in o e coming his powe lessness,
go e nmen s can ei he do so passi ely, e.g., by no
s anding in he way o he g ass oo s mo emen s o
esou ce con ol ha a e unde way, o ac i ely, e.
g., by wo king o escind laws and egula ions ha
ha e pe pe ua ed disad an ages.
CONCLUSION
We sough o e alua e he p oposi ion ha o es s
a e impo an o he poo in de eloping coun ies
no jus because o wha hey o e , bu jus as
impo an ly, because o whe e hey a e loca ed in
ela ion o he places whe e he poo li e. Ou
indings, based on case s udies in B azil, Hondu as,
Malawi, Mozambique, Uganda, Indonesia, and
Vie nam, indica e ha he e is a disce nible
endency o a eas o high o es co e o be
cha ac e ized by a high po e y a e. Howe e , high
o es co e also ends o be associa ed wi h low
po e y densi y. This makes sense because mos
dense o es s a e no places ha can suppo dense
human habi a ion.
We a gue ha na ional po e y alle ia ion s a egies
should gi e g ea e a en ion o o es esou ces han
hey do cu en ly because he e a e la ge numbe s
o poo who li e in o es s o all kinds, bo h closed
and open, and because he high po e y a e ha is
cha ac e is ically ound in a eas o high o es co e
is o en associa ed wi h deep, se e e, and ch onic
po e y. We obse e ha o es -based po e y
alle ia ion s a egies can be ele an ega dless o
whe he he e a e o es s in he landscape. Po e y
alle ia ion s a egies ha a e cus oma ily applied o
emo e a eas should be complemen ed by suppo
o ini ia i es ha a e aimed a o e coming he
poli ical powe lessness o people who li e in emo e
o es ed egions.
Ou indings a e impo an because hey pu pas
assump ions abou he spa ial link be ween po e y
and o es s on a s onge empi ical oo ing and
because hey make a s ong case o a cou se
co ec ion in some na ional po e y alle ia ion
s a egies. Ne e heless, u u e esea ch o his kind
conduc ed h ough addi ional coun y case s udies
would be aluable and can be done when da a
become a ailable. New s udies can ein o ce he
need o ake o es y mo e se iously in de elopmen
planning and policies. The maps p oduced h ough
such esea ch can be a aluable ool o spa ial
a ge ing o u al de elopmen p og ams and
policies. Mo eo e , mul i a ia e and ime se ies
analyses will allow a mo e ho ough unde s anding
o he causali y o he associa ions ound he e.
Responses o his a icle can be ead online a :
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/ esponses/
Acknowledgmen s:
We hank se e al people who p o ided excellen
commen s on a longe epo om which his a icle
was de i ed: Ken Chomi z, Pe e Dewees, Da id
Kaimowi z, B uce McKenney, Daniel Mülle , Je
Saye , and Luca Tacconi. And ee Ekadina a and
Agus Salim p o ided aluable help in echecking
he da a and e ising he maps. We alone a e
esponsible o any e o s ha may emain in he
inished p oduc .
LITERATURE CITED
Ande son, A. B., P. H. May, and M. J. Balick.
1991. The subsidy om na u e: palm o es s,
peasan y, and de elopmen on an Amazon on ie .
Columbia Uni e si y P ess, New Yo k, New Yo k,
USA.
Angelsen, A., and S. Wunde . 2003. Explo ing he
o es –po e y link: key concep s, issues and
esea ch implica ions. CIFOR Occasional Pape
Numbe 40. Cen e o In e na ional Fo es y
Resea ch, Bogo , Indonesia. A ailable online a :
h p://www.ci o .cgia .o g/Publica ions/Pape s/.
Anselin, L. 1988. Spa ial econome ics: me hods
and models. Kluwe Academic Publishe s,
Do d ech , The Ne he lands.
Anselin, L. 1995. Local indica o s o spa ial
associa ion—LISA. Geog aphical Analysis 27
(2):93-115.
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Anselin, L. 2005. Explo ing spa ial da a wi h
GeoDaTM: a wo kbook. Spa ial Analysis Labo a o y,
Depa men o Geog aphy, Uni e si y o Illinois a
U bana-Champaign, U bana, Illinois, USA.
A ailable online a :
www.sal.uiuc.edu/s u /s u -sum/
pd /geodawo kbook.pd .
Badan Pusa S a is ik. 2000. SUSENAS (na ional
socioeconomic su ey) 2000. Badan Pusa S a is ik,
Jaka a, Indonesia.
Ba bie i, A. F., D. L. Ca , and R. E. Bilsbo ow.
2008. Mig a ion wi hin he on ie : he second
gene a ion coloniza ion in he Ecuado ian Amazon.
Popula ion Resea ch and Policy Re iew in p ess.
Benson, T. 2002. Malawi: an a las o social
s a is ics. Na ional S a is ical O ice, Go e nmen
o Malawi, Zomba, Malawi, and In e na ional Food
Policy Resea ch ins i u e, Washing on, D.C, USA.
A ailable online a :
h p://www.i p i.o g/pubs/cp/m
alawia las.h m.
Bi d, K., D. Hulme, K. Moo e, and A. Shephe d.
2002. Ch onic po e y and emo e u al a eas.
CPRC Wo king Pape Numbe 13. Ch onic Po e y
Resea ch Cen e, Uni e si y o Bi mingham,
Bi mingham, UK. A ailable online a :
www.ch on
icpo e y.o g/pd s/13Bi d_e _al.pd .
By on, R. N., and J. E. M. A nold. 1999. Wha
u u es o he people o he opical o es s? Wo ld
De elopmen 27(5):789-805.
Ca , D. L. 2006. A ale o wo oads: land enu e,
po e y, and poli ics on he Gua emalan on ie .
Geo o um 37(1):94-103.
Chambe s, R., M. Leach, and C. Con oy. 1993.
T ees as sa ings and secu i y o he u al poo .
Ga ekeepe Se ies Numbe 3. In e na ional Ins i u e
o En i onmen and De elopmen , London, UK.
A ailable online a :
h p://www.iied.o g/pubs/pd s/
6025IIED.pd .
Chomi z, K. M. 2007. A logge heads: ag icul u al
expansion, po e y educ ion, and en i onmen in
he opical o es s. Wo ld Bank, Washing on, D.
C., USA.
Dasgup a, S., U. Deichmann, C. Meisne , and D.
Wheele . 2005. Whe e is he po e y–en i onmen
nexus? E idence om Cambodia, Lao PDR, and
Vie nam. Wo ld De elopmen 33(4):617-638.
DeF ies, R., M. Hansen, J. R. G. Townshend, A.
C. Jane os, and T. R. Lo eland. 2000a. 1 kilome e
ee co e con inuous ields. Ve sion 1.0.
Depa men o Geog aphy, Uni e si y o Ma yland,
College Pa k, Ma yland, USA. A ailable online a :
h p://www.landco e .o g/da a/ eeco e /.
DeF ies, R., M. Hansen, J. R. G. Townshend, A.
C. Jane os, and T. R. Lo eland. 2000b. A new
global 1-km da ase o pe cen ee co e de i ed
om emo e sensing. Global Change Biology 6
(2):247-254.
Deininge , K. 2003. Land policies o g ow h and
po e y educ ion. Wo ld Bank, Washing on, D.C.,
USA.
Depa men o In e na ional De elopmen .
2007. Land: be e access and secu e igh s o poo
people. Depa men o In e na ional De elopmen ,
London, UK. A ailable online a :
h p://www.d id.
go .uk/pubs/ iles/LandPape 2007.pd .
De con, S. 2008. Ru al po e y: old challenges in
new con ex s. GPRG-WPS 072. Global Po e y
Resea ch G oup, Economic and Social Resea ch
Council, Ox o d, UK. A ailable online a :
www.gp
g.o g/pubs/wo kingpape s/pd s/gp g-wps-072.pd
.
Ellis, F. 2000. Ru al li elihoods and di e si y in
de eloping coun ies. Ox o d Uni e si y P ess,
Ox o d, UK.
Food and Ag icul u e O ganiza ion. 2001. Global
o es esou ces assessmen 2000: main epo .
FAO Fo es y Pape 140. Food and Ag icul u e
O ganiza ion, Rome, I aly. A ailable online a : h
p://www. ao.o g/ o es y/ a2000 epo /en/.
Food and Ag icul u e O ganiza ion. 2002. Land
enu e and u al de elopmen . FAO Land Tenu e
S udies 3. Food and Ag icul u e O ganiza ion,
Rome, I aly. A ailable online a :
h p://www. ao.o g/
DOCREP/005/Y4307E/y4307e00.h m.
Fos e , J., J. G ee , and E. Tho becke. 1984. A
class o decomposable po e y measu es.
Econome ica 52(3):761-766.
Fu emma, C., and E. S. B ondízio. 2003. Land
e o m and land-use changes in he lowe Amazon:
implica ions o ag icul u al in ensi ica ion. Human
Ecology 31(3):369-402.
Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Hulme, D. 2003. Ch onic po e y and de elopmen
policy: an in oduc ion. Wo ld De elopmen 31
(3):399-402.
Hulme, D., and A. Shephe d. 2003. Concep ualizing
ch onic po e y. Wo ld De elopmen 31(3):403-423.
In e na ional Li es ock Resea ch Ins i u e.
2005. Po e y a ge ing ool: s a egic analysis and
knowledge suppo sys em (SAKSS). CD-ROM,
BETA e sion. In e na ional Li es ock Resea ch
Ins i u e, Nai obi, Kenya.
Ra allion, M., and J. Jalan. 1997. Spa ial po e y
aps? Wo ld Bank Policy Resea ch Wo king Pape
Numbe 1862. Wo ld Bank, Washing on, D.C.,
USA.
Kanbu , R., and A. J. Venables. 2005. Spa ial
inequali y and de elopmen . Pages 3-14 in R.
Kanbu and A. J. Venables, edi o s. Spa ial
inequali y and de elopmen . Ox o d Uni e si y
P ess, Ox o d, UK.
Kha e, A., M. Sa in, N. C. Saxena, S. Pali , S.
Ba hla, F. Vania, and M. Sa yana ayana. 2000.
Join o es managemen : policy, p ac ice, and
p ospec s. Policy ha Wo ks o Fo es and People
Se ies Numbe 3. Wo ld Wide Fund o Na u e,
New Delhi, India, and In e na ional Ins i u e o
En i onmen and De elopmen , London, UK.
McKay, A., and D. Lawson. 2003. Assessing he
ex en and na u e o ch onic po e y in low income
coun ies: issues and e idence. Wo ld De elopmen
31(3):425-439.
McSweeney, K. 2004. Fo es p oduc sale as na u al
insu ance: he e ec s o household cha ac e is ics
and he na u e o shock in eas e n Hondu as. Socie y
and Na u al Resou ces 17(1):39-56.
Meh a, A. K., and A. Shah. 2003. Ch onic po e y
in India: incidence, causes and policies. Wo ld
De elopmen 31(3):491-511.
Mino , N., B. Baulch, and M. Epp ech . 2003.
Po e y and inequali y in Vie nam: spa ial pa e ns
and geog aphic de e minan s. In e na ional Food
Policy Resea ch Ins i u e, Washing on, D.C., USA,
and Ins i u e o De elopmen S udies, B igh on,
UK. A ailable online a :
h p://www.i p i.o g/di s/
MTID/DR/d 200312mapping.h m.
Mülle , D., M. Epp ech , and W. D. Sunde lin.
2006. Whe e a e he poo and whe e a e he ees?
Ta ge ing o po e y educ ion and o es
conse a ion in Vie nam. CIFOR Wo king Pape
Numbe 34. Cen e o In e na ional Fo es y
Resea ch, Bogo , Indonesia. A ailable online a :
h p://www.ci o .cgia .o g/Publica ions/Pape s/wkpape s/
.
Odu o, A. D., and I. A yee. 2003. In es iga ing
ch onic po e y in Wes A ica. Ch onic Po e y
Resea ch Cen e Wo king Pape Numbe 28. Cen e
o Policy Analysis, Uni e si y o Manches e ,
Manches e , UK. A ailable online a :
h p://www.c
h onicpo e y.o g/pd s/28Odu o_A yee.pd .
Oksanen, T., and C. Me smann. 2003. Fo es s in
po e y educ ion s a egies: an assessmen o PRSP
p ocess in sub-Saha an A ica. Pages 121-155 in T.
Oksanen, B. Paja i, and T. Tuomasjukka, edi o s.
Fo es s in po e y educ ion s a egies: cap u ing
he po en ial. EFI P oceedings Numbe 47.
Eu opean Fo es Ins i u e, Joensuu, Finland.
A ailable online a :
h p://www.e i.in /po al/ i ua
l_lib a y/publica ions/p oceedings/47/.
Pa anayak, S. K., and E. O. Sills. 2001. Do
opical o es s p o ide na u al insu ance? The
mic oeconomics o non- imbe o es p oduc
collec ion in he B azilian Amazon. Land
Economics 77(4):595-612.
Robles, M. 2003. Es imación de indicado es de
pob eza y desigualdad a ni el municipal en
Hondu as. BID/MECOVI–Ins i u o Nacional de
Es adís ica, Tegucigalpa, Hondu as.
Ruíz Pé ez, M., B. Belche , M. Fu, and X. Yang.
2004. Looking h ough he bamboo cu ain: an
analysis o he changing ole o o es and a m
income in u al li elihoods in China. In e na ional
Fo es y Re iew 6(3-4):306-316.
Shah, A., and B. Gu u. 2004. Po e y in emo e
u al a eas in India: a e iew o e idence and issues.
CPRC-IIPA Wo king Pape Numbe 21. Indian
Ins i u e o Public Adminis a ion, New Delhi,
India. A ailable online a :
h p://www.ch onicpo e y.
o g/p/308/publica ion-de ails.php.
Shimeles, A., and R. Thoenen. 2005. Po e y
p o iles: a me hodological no e on measu ing
po e y. Wo king Pape ESPD/NRP/04/05. Uni ed
Na ions Economic Commission o A ica, Addis

Ecology and Socie y 13(2): 24
h p://www.ecologyandsocie y.o g/ ol13/iss2/a 24/
Ababa, E hiopia. A ailable online a :
www.uneca.o
g/espd/publica ions/Po e y_p o iles_a_me hodol
ogical_no e_on_measu ing_po e.pd .
Simle , K. R., and V. Nha e. 2005. Po e y,
inequali y, and geog aphic a ge ing : e idence om
small-a ea es ima es in Mozambique. FCND
Discussion Pape Numbe 192. In e na ional Food
Policy Resea ch Ins i u e, Washing on, D.C., USA.
A ailable online a :
h p://ideas. epec.o g/p/ p / cn
ddp/192.h ml.
Sunde lin, W. D., A. Angelsen, B. Belche , P.
Bu ge s, R. Nasi, L. San oso, and S. Wunde .
2005. Li elihoods, o es s, and conse a ion in
de eloping coun ies: an o e iew. Wo ld
De elopmen 33(9):1383-1402.
Sunde lin, W. D., S. Dewi, and A. Pun odewo.
2007. Po e y and o es s: mul i-coun y analysis o
spa ial associa ion and p oposed policy solu ions.
CIFOR Occasional Pape Numbe 47. Cen e o
In e na ional Fo es y Resea ch, Bogo , Indonesia.
A ailable online a :
h p://www.ci o .cgia .o g/Pub
lica ions/Pape s/.
Swedish In e na ional De elopmen Coope a ion
Agency. 2007. Na u al esou ce enu e: a posi ion
pape o SIDA. Swedish In e na ional De elopmen
Coope a ion Agency, S ockholm, Sweden. A ailable
online a :
h p://www.sida.se/sida/jsp/sida.jsp?d=11
8&a=32805&language=en_US.
Takasaki, Y., B. L. Ba ham, and O. T. Coomes.
2004. Risk coping s a egies in opical o es s:
loods, illnesses and esou ce ex ac ion. En i onmen
and De elopmen Economics 9(2):203-224.
Vedeld, P., A. Angelsen, E. Sjaas ad, and G.
Kobugabe Be g. 2004. Coun ing on he
en i onmen : o es incomes and he u al poo .
En i onmen al Economics Se ies Numbe 98.
Wo ld Bank, Washing on, D.C., USA. A ailable
online a :
h p://lnweb18.wo ldbank.o g/essd/en ex .
ns /44ByDocName/Coun ingon heEn i onmen Fo
es Incomesand he Ru alPoo 2004/$FILE/
Coun ingon heEn i onmen Fo es Incomesand heR
u alPoo .pd .
Whi e, A., and A. Ma in. 2002. Who owns he
wo ld’s o es s? Fo es enu e and public o es s in
ansi ion. Fo es T ends and Cen e o
In e na ional En i onmen al Law, Washing on, D.
C., USA. A ailable online a :
www. o es - ends.o g/
documen s/publica ions/ enu e epo _whoowns.pd
.
Wo ld Bank. 2006. India: unlocking oppo uni ies
o o es -dependen people in India. Repo
Numbe 34481-IN. Wo ld Bank, Washing on, D.C.,
USA. A ailable online a :
h p://si e esou ces.wo l
dbank.o g/INDIAEXTN/Resou ces/Repo s-
Publica ions/366387-1143196617295/
Fo es y_Repo _ olume_I.pd .
Zhou, L., and G. Veeck. 1999. Fo es esou ce use
and u al po e y in China. Fo es y Economics 4
(1):80-92.