Rod íguez, O lando; V achioli, Ma ia; Wüppe , Da id; Saue , Johannes
A icle — Published Ve sion
Wo ld He i age S a us and Fa me s’ Income: E idence
F om a Reg ession Discon inui y Design in Colombia
Ag icul u al Economics
P o ided in Coope a ion wi h:
John Wiley & Sons
Sugges ed Ci a ion: Rod íguez, O lando; V achioli, Ma ia; Wüppe , Da id; Saue , Johannes (2025) :
Wo ld He i age S a us and Fa me s’ Income: E idence F om a Reg ession Discon inui y Design in
Colombia, Ag icul u al Economics, ISSN 1574-0862, Wiley, Hoboken, NJ, Vol. 56, Iss. 5, pp. 728-748,
h ps://doi.o g/10.1111/agec.70019
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/330161
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/4.0/
Ag icul u al Economics
ORIGINAL ARTICLE
Wo ld He i age S a us and Fa me s’ Income: E idence F om
a Reg ession Discon inui y Design in Colombia
O lando Rod íguez1Ma ia V achioli1Da id Wüppe 2Johannes Saue 1
1Technical Uni e si y o Munich, Munich, Ge many 2Uni e si y o Bonn, Bonn, Ge many
Co espondence: O lando Rod íguez (o lando[email p o ec ed])
Recei ed: 20 Ap il 2024 Re ised: 4Feb ua y2025 Accep ed: 13 Feb ua y 2025
Funding: This s udy was suppo ed by he Ge man Resea ch Founda ion (DFG), unde Ge many’s Excellence S a egy EXC 2070, G an No. 390732324 —
PhenoRob.
Keywo ds: co ee p oduc ion | eg ession discon inui y | wo ld he i age
ABSTRACT
The co e o Colombia’s co ee g owing egion was designa ed as a Wo ld He i age (WH) si e in 2011, making a dis inc ion be ween
“ he co e” o he egion and “ he pe iphe y.” Howe e , he co ee cul u al he i age does no ab up ly s op a he WH bounda y bu
i exis s inside and ou side he bounda y. This allows us o use a eg ession discon inui y design (RDD) o iden i y he income e ec
o a me s loca ed jus inside he WH si e. We ind ha he WH designa ion o he egion inc eases co ee a me s’ income by up
o $757 pe mon h. The mechanism includes mo e ou ism ac i i ies inside he co ee a ms by 39%, highe adop ion o sus ainable
a ming p ac ices, inc easing o ganic co ee p oduc ion by 8.38%, and inc easing paymen s o en i onmen al se ices (PES) by
2.35%.
JEL Classi ica ion: O13, Q12, Q57
1 In oduc ion
Co ee p ices a e ola ile, leading o high unce ain y o co ee
a me s, and his is exace ba ed by inc easing clima ic and pes
p oblems. Ye , he e a e s ill hal a million amilies in Colombia
ha depend on co ee p oduc ion o hei li elihood. Co ee is
one o he leading expo p oduc s in Colombia, and a me s
ha e de eloped a cul u al iden i y a ound co ee p oduc ion. In
2011, he “Co ee Cul u al Landscape o Colombia” (CCL) was
insc ibed on o he Wo ld He i age (WH) Lis wi h he main
aim o p o ec Colombia’s unique co ee p oduc ion cul u e and
adi ions and wi h a seconda y goal o imp o ing he li ing
s anda d o Colombian co ee p oduce s.
This pape explo es whe he he WH s a us, in ac , led o
di e ences in income be ween co ee a me s ac oss he WH
bo de . We exploi he ac ha he WH si e was c ea ed wi h
a sha p bounda y bu aims o p o ec he cul u al he i age o
an a ea ha does no o he wise ha e a sha p bounda y. The
Co ee Cul u al Landscape o Colombia co e s a la ge a ea wi h
a long and amous co ee adi ion. The bounda y o he WH si e
was placed inside his as a ea such ha close o he bounda y,
bo h sides ha e simila co ee g owing condi ions—cul u ally,
en i onmen ally—and he only s a k di e ence is ha since 2011,
one side is inside he WH si e and one side is ou side. This
is an op imal na u al expe imen o use a spa ial eg ession
discon inui y design (RDD) o iden i y he causal income e ec
o he WH si e.
UNESCO (1972) de ines WH as one si e o ou s anding uni e sal
alue om a his o ical, e hnological, aes he ic, o an h opological
poin o iew. The Colombian co ee WH a ea was delinea ed
acco ding o he combina ion o 16 cul u al and en i onmen al
a ibu es o p o ec “ he co e” o he adi ional co ee p oduc ion
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly
ci ed.
© 2025 The Au ho (s). Ag icul u al Economics published by Wiley Pe iodicals LLC on behal o In e na ional Associa ion o Ag icul u al Economis s.
728 Ag icul u al Economics, 2025; 56:728–748
h ps://doi.o g/10.1111/agec.70019
FIGURE 1 Loca ion o a ms a ound he co ee WH bo de .
a ea wi h all i s cul u al signi icance. By a , no he en i e egion
was included in he WH si e, and adi ional co ee-g owing
communi ies a e equally ound on bo h sides o he WH bo de .
Recognizing his, he Colombian go e nmen al eady signed Ac
2245 o July 2022 (Cong ess o Colombia 2022), p omo ing he
ecogni ion o mo e e i o ies o he Colombian co ee egion in
he WH s a us. The go e nmen claims ha he adi ional co ee
egion is bigge han he ame d awn by UNESCO. C i icism
om he Go e nmen claims ha he Co ee Cul u al he i age
and he adi ional co ee p oduc ion is no limi ed by he WH
bo de p oposed by UNESCO. Co ee p oduc ion in Colombia
was de eloped du ing he 20 h and 21s cen u ies in he same
condi ions inside and ou side o he WH a ea. The loca ions jus
inside and ou side he WH bounda y a e simila in adi ion and
cul u e, co ee g owing condi ions, and o he con ex ual a iables
de e mined be o e he es ablishmen o he WH si e. The e is
also g ea con inui y in who owns and manages he a ms in
he egion, which is domina ed by small amily es a es ha a e
handed down h ough he gene a ions.
Figu e 1shows he loca ion and ex en o he WH si e. One
cha ac e izing ea u e o i is ha i is no a con iguous a ea,
bu i is a he agmen ed in o six subzones. The eason o his
agmen a ion was o lea e ou mode n co ee a ms wi h mod-
i ied in as uc u e and se lemen s wi h mode n u baniza ion.1
The WH has 207,000 hec a es, dis ibu ed in a e ical shape,
embedded in he adi ional co ee egion. The a e age dis ance
be ween one WH bo de o he opposi e bo de is 4.8 km. The
highes dis ance be ween wo segmen s o he bo de is 23
km, bu 70% o he WH a ea has less han 6 km o dis ance
b/w bo de s. We included in ou pooled sample he closes
co ee a ms o he WH bo de , expec ing ha we could collec
da a om co ee a ms wi h simila condi ions and he mos
simila cases. To decide which pa s o he co ee g owing egion
should be decla ed as WH, a lis o objec i e c i e ia was used:
na u al he i age, wa e sou ce a ailabili y, a chi ec u al he i age,
a cheological he i age, concen a ed se lemen and agmen ed
p ope y s uc u e, in luence o mode niza ion, u ban he i age,
co ee p oduc ion his o ical adi ion, moun ain g own co ee,
co ee ins i u ions and ela ed ne wo ks, co ee p edominance,
slope o co ee g owing, age, smallholding as landowne ship sys-
em, mul iple c ops, and sus ainable p oduc ion and echnologies
in he co ee p oduc i e chain. These c i e ia a e known as he 16
a ibu es o he WH si e (UNESCO 2011). I was no easible o
include all communi ies ha sco ed high acco ding o he c i e ia,
bu i was ied o include a “ ep esen a i e a ea.” Impo an ly, he
a eas loca ed jus wi hin he WH si e ne e had a sha p bounda y
wi h he a eas loca ed jus ou side, and whe e o exac ly se he
bounda y o he WH si e was app oxima ely a bi a y. In he end,
he p agma ic ou come was o include a ep esen a i e selec ion
o “ e edas”2wi hin 51 municipali ies among he ou mainly
ele an co ee p oducing depa men s (12 om Risa alda, 18 om
Caldas, 11 om Quindio, and 10 om Valle de Cauca) and d aw
a bo de be ween hese and hei neighbo ing municipali ies.
Figu es 2and 3show how he WH bo de was d awn ac oss he
s ee s o he municipali ies, passing h ough buildings, houses,
and blocks. In some a eas, hal o he same in as uc u e was
included in he WH s a us, and he o he hal was le ou . The
WH si e has unique condi ions highligh ed by he UNESCO
c i e ia; howe e , bo h he a ea jus wi hin and jus ou side he
WH si e exempli y hese condi ions. Thus, he WH si e ep esen s
he main ace s o he Colombian co ee adi ion, bu he neigh-
bo ing municipali ies do no ha e less co ee adi ion o di e in
o he ele an ways om he included ones. Renewal a chi ec u e
a oided some a eas o be included in he WH a ea. Howe e , he
economic ac i i y and he co ee adi ion a e he same.
Ag icul u al WH si es ha e no ye ecei ed much academic
a en ion o da e, al hough in many egions o he wo ld, a me s
ha e low and uns able incomes, and suppo ing hem is egula ly
729
FIGURE 2 Wo ld He i age bo de ac oss he municipali y o Pe ei a (Risa alda).
FIGURE 3 Wo ld He i age bo de ac oss he municipali y o Ci casia (Quindio).
730 Ag icul u al Economics,2025
FIGURE 4 Ho els and es au an s in he co ee egion.
FIGURE 5 In es men s in oad in as uc u e in he co ee egion
in housands o US dolla s a e 2011.
an explici ly s a ed policy goal when u al WH si es a e p oposed.
Pa icula ly in low-income and middle-income coun ies, he e
is o en a goal con lic be ween p o ec ing he en i onmen
and aising local li elihoods. Po en ially, WH si es can suppo
bo h, bu he li e a u e shows a he mixed esul s ega ding his
syne gy (Fu e al. 2020;Sue al.2020).
Ou main con ibu ion is he iden i ica ion o he causal
e ec o he WH bo de . The impac o he WH is pa icula ly
anspa en ly and obus ly es ima ed he e, which is enabled
by how he bo de o he WH si e was d awn, no coinciding
wi h any p e-exis ing discon inui ies, (i.e., in geog aphy, cul u e,
poli ics), and he e is no endogenous so ing because he a ms
in he egion a e owned and managed by he same amilies o e
gene a ions. This is an op imal con ex o a RDD (Wueppe
and Finge 2023). We collec ed su ey da a om 337 co ee
a me s si ua ed bo h inside and ou side he WH si e and we
use a spa ial RDD o iden i y he causal e ec o he WH lis ing.
Ou main inding is ha WH s a us inc eased he income o
he co ee a me s by $756.97 pe mon h, which was mos ly
d i en by inc eased ac i i ies o co ee g owe s’ in ol emen in
ou ism ne wo ks (39%). Consis en wi h his, Figu e 4shows
ha he numbe o ho els and es au an s inc eased in he co ee
egion since he es ablishmen o he WH s a us om 819 in 2005
o 4641 in 2019 (Ci u 2021). In he same line, Figu e 5shows
highe in es men s in oad in as uc u e inside he WH a ea and
Figu e 6desc ibes mo e kilome e s buil in he WH a ea a e 2011,
allowing access o co ee a ms and ou is ic ameni ies h ough
he adi ional co ee egion (Minis y o T anspo 2024). Second,
highe income among co ee g owe s inside he WH a ea was also
in pa due o inc eased mon hly paymen s o en i onmen al
se ices (PES) o $17.77. Finally, WH s a us changed a ming
p ac ices, namely he e was a shi o mo e o ganic co ee
p oduc ion, which inc eased 8.38%. The main esul s p o ide
inpu s o policy-making p ocess, ela ed o he suppo o
esilience among ulne able adi ional g owe ’s amilies.
The emainde o he pape is o ganized as ollows. The nex sec-
ion p esen s he heo e ical e iew, he hi d sec ion in oduces
da a and me hods, he ou h sec ion p o ides he esul s, and he
inal sec ion o e s ou conclusions.
2Li e a u e Re iew
Ways o ensu ing a me s’ income du ing economic c ises ha e
been an ac i e ield o esea ch o a long ime. Fa me s’ income
can be in luenced by luc ua ions in p ices paid o ecei ed,
shi s in g oss domes ic p oduc , echnological change, a ia ion
in in e es and exchange a es, and land p ices (Beckman and
Schimmelp enning 2015). Thus, g owe s ace boom-bus cycles
leading o economic unce ain y. G ea e a m income a iabili y
leads o mo e o - a m ac i i ies (Mish a and Goodwin 1997).
The e o e, u he s a egies a e needed o complemen he
income o p oduce s. Pa icula ly, some li e a u e has ocused
on he e ec s o WH s a us on g owe s’ income. Local ne wo ks
ake ad an age o highe isibili y wi hin he WH lis , looking o
o he economic ac i i ies ha inc ease he income o domes ic
p oduce s (Fauconbe g e al. 2017). WH s a us is used by local
go e nmen s like ci y ma ke ing labels o p omo e social, eco-
nomic, and en i onmen al goals in a gi en egion among hei
s akeholde s like ou is s, in es o s, and dono s.
Fa me s’ income in WH a eas could inc ease h ough mecha-
nisms like ou ism and PES, which compensa e o down-swings
du ing c op p ice c ises. Some e idence shows ha he in eg a-
ion o mul iple he i age alues wi h ou ism is highly impo an
o sus aining ag icul u al he i age sys ems and u al commu-
ni ies (Su e al. 2020). Howe e , some au ho s conside ha i
s ill needs o be explo ed on how o alle ia e he con adic ion
be ween en i onmen al p o ec ion and ou ism de elopmen
aced by WH si es and how o achie e ade-o s and syne gies
o ecosys em se ices and compensa ions (Fu e al. 2020).
The e is ex ensi e li e a u e on he economic bene i s o WH
lis ing, bu he indings a e mixed, mainly highligh ing he
bene i s o WH s a us on p omo ing ou ism. On he one hand,
Wueppe (2017) and Su and Lin (2014) ound ha he economy
in and a ound WH si es can bene i om inc eased ou ism
e enues. On he o he hand, se e al s udies ound ha ou ism
did no inc ease due o he WH lis ing, including Gao and Su
(2019) and Huang e al. (2012). Pa icula ly s iking, Yang and Lin
(2014) concluded ha cul u al WH si es can bene i economically
om hei lis ing, whe eas Wueppe and Pa y (2016) concluded
ha less amous, mo e emo e, and u al des ina ions bene i he
mos om WH s a us.
The WH s a us linked o en i onmen al a ibu es is also man-
aged by local go e nmen s as an oppo uni y o launching PES
731
FIGURE 6 Kilome e s o oad in as uc u e buil in he co ee egion a e 2011.
schemes o p o ec na u al esou ces and inc ease he income
o local g owe s. PES schemes o e economic incen i es o
p oduce s o p o ec he ecosys ems (Wunde 2005). The e o e,
PES complemen s a me s’ income by ul illing en i onmen al
policies. Hube man (2009) concludes ha he WH lis s eng h-
ens PES schemes, because he global known WH s a us p omo es
he links be ween o eign dono s and local needs. The ele ance
o PES inc eases a ound a WH si e o ensu e he con inui y
o speci ic na u al o cul u al a ibu es wi hin he c i e ia o
he WH con en ion. Howe e , Nuppenau (2002) poin ed ou
ha policies wi h PES schemes a e di icul o e alua e because
PES ag eemen s ep esen a p incipal-agen issue. On he one
hand, he a me s’ ( he agen s’) in e es is o maximize p o i s,
acco ding o hei beha io as homo economicus. On he o he
hand, he public ( he p incipal) wan s o p o ec en i onmen al
se ices. Fu he mo e, K onenbe g and Hubacek (2013)wa n
ha PES schemes could lead o a esou ce cu se because PES
p og ams could cen alize e enues on a ew ac i i ies, leading
o dependency on PES in he long e m. The e o e, PES schemes
should be complemen ed wi h u he mechanisms o a oid he
lack o di e si ica ion among egional p oduce s.
WH s a us c ea es isibili y in delimi ed ag icul u al a eas.
Tou is s a e a ac ed o isi he WH si es, eco e ing domes ic
economic ac i i ies. Fu he , unde s o PES schemes pe cei e
he needs ela ed o en i onmen al and social p o ec ion wi hin
he WH zone, launching ag eemen s wi h local a me s. O e all,
i is expec ed ha WH s a us p omo es no only ou ism and
PES p og ams bu can also enhance he income o he local
g owe s.
3Da a and Me hods
3.1 Sou ces o Da a
The WH bo de was delimi ed acco ding o geo e e enced coo -
dina es, de ined by UNESCO (2011) du ing he insc ip ion o he
Colombian Co ee Cul u al Landscape in he WH lis in 2011.
The coo dina es o he main zone and he bu e zone a he
WH a ea o his s udy we e aken om he Colombian Minis y
o Cul u e (2017), which is he na ional au ho i y in cha ge o
he p o ec ion o he WH a ea. The bu e zone is a p o ec ion
ame o he main a ea o he WH. Bo h he bu e zone and he
main zone cons i u e he WH a ea. “The CCLC con ains e y ew
con empo a y incong uous addi ions o i s adi ional a chi ec u al
and landscape pa e ns, and no subs an ial modi ica ions o he
small owns loca ed in he p ope y as well as in he bu e zone.
Aspec s such as adi ions, language, and o he o ms o in angible
he i age, ha e been p ese ed, mos ly by owne s and he local
communi y, who ha e a high sense o social app op ia ion o hei
cul u al he i age” (UNESCO 2011).
Da a ela ed o he income o a me s, hei pa icipa ion in
ou is ne wo ks, he amoun o PES, and he numbe o PES
ecipien s we e collec ed by conduc ing on-si e su eys wi h 337
co ee a me s in he Wes e n and Cen al Moun ain anges o he
Andes in 2017. The sample co e ed he depa men s o Risa alda,
Quindio, Caldas, and Valle del Cauca, which ep esen he mos
adi ional co ee p oduc ion egions in Colombia. The da ase
con ains de ailed in o ma ion on co ee p oduc ion, household
cha ac e is ics, socioeconomic in o ma ion, and en i onmen al
a iables (Table 1).
The sample selec ion was ocused on a ms loca ed close o he
WH bo de in a ame o 5 km o bo h sides o he WH on ie ,
seeking ha he obse a ions we e consis en in he a iable
“Dis ance.” The sampling ame is ep esen a i e o he a ea and
consis en wi h a simple andom sampling o co ee a ms along
he WH bo de . Figu e 1shows he spa ial dis ibu ion o he main
zone and he bu e zone o he WH a ea a ound he co ee egion.
The popula ion o co ee a ms in he adi ional co ee egion
was 93,500 (Dane 2016). Following he classic Coch an o mula
(1977)inEqua ion(1), he sample size should be highe han
336.8 obse a ions, whe e 𝑒is a ma gin o e o o 5.34%, 𝑝is he
es ima ed popula ion ha con i ms he a ibu e (50%), 𝑞is he
732 Ag icul u al Economics,2025
TABLE 1 De ini ion o a iables.
Va iable Uni De ini ion
To al income US $ Mon hly o al income o co ee g owe s
In ol emen in ou ism ne wo ks Dummy 1 i co ee g owe is in ol ed in ou ism ne wo k, o he wise 0
Mon hly paymen o ES US $ Mon hly mone a y compensa ion ecei ed pe co ee g owe
om PES
Pa icipa ion in PES schemes Dummy 1 i co ee a me is in ol ed in PES schemes; o he wise 0
O ganic co ee p oduc ion Pe cen age Sha e o o ganic co ee ou o o al co ee p oduc ion
Age Yea s The ime someone has been ali e
Educa ion Le el Educa ion le el is classi ied acco ding o he ollowing
ca ego ies: 1 =None; 2 =Some p ima y; 3 =P ima y comple ed;
4=Some seconda y; 5 =Seconda y comple ed; 6 =Technical
aining a e seconda y; 7 =Uni e si y
Na u al ee co e po en ial Sco e P edic ed pe cen age o ee co e wi hou human impac , based
on 10 summa ized a iables: annual mean empe a u e, annual
p ecipi a ion, p ecipi a ion seasonali y, mean empe a u e o he
we es qua e , p ecipi a ion o he d ies qua e , o ganic
ca bon s ock om ce o o 15 cm, The da a on he na u al ee
co e po en ial a e om Bas in e al. (2019) dep h o bed ock,
sand con en om ce o o 15 cm, ele a ion, and hill-shade
Al i ude Me e s Loca ion o a m abo e he sea le el
Tempe a u e Celsius Deg ees o ho ness o coldness
Loca ion Coo dina es Geog aphic coo dina es o each a m acco ding o longi ude and
la i ude
Dis ance o he bo de Kilome e s The leng h o space be ween he WH bo de and he a m
Bounda y ixed e ec s Yes/No Se o bounda y segmen o know which bo de is he closes o
he a m
Membe ship Yea s Time a ilia ed o he Na ional Fede a ion o co ee g owe s
Fa m owne ship Dummy 1 i co ee a me is owne o he a m, o 0 i a me is no he
owne
Al e na i e cu -o Km Dis ance o new placebo bo de o he WH bo de
Poli ical bo de s Depa men s F on ie be ween wo depa men s
Land pe a m Hec a es To al a ea wi hin he a m sys em
Co ee land Hec a es A ea o ien a ed o co ee p oduc ion wi hin he a m
Co ee c ops in 2007 Hec a es A ea pe municipali y in he co ee egion
Co ee c ops be o e 2011 Hec a es A ea o co ee ields
Volume o co ee be o e 2011 Tons Weigh o co ee p oduc ion
es ima ed popula ion ejec ing he a ibu e (50%), and 𝑧 alue is
1.96, which is calcula ed acco ding o a con idence le el o 95%.
Sample =𝑧2∗𝑝∗𝑞
𝑒2(1)
We ma ched sa elli e images wi h he geo e e enced loca ion o
each a m in ou da abase h ough GIS analysis. The Minis y o
Cul u e ha o e sees p o ec ing he a ibu es o he Colombian
Co ee Cul u al Landscape wi hin he WH p o ided us wi h he
o icial geo e e enced coo dina es o he WH a ea. Figu e 1shows
he app oxima e dis ibu ion o a ms wi hin he main zone and
he bu e zone o he WH and ou side o he WH, p o ec ing he
iden i y and pe sonal in o ma ion o a me s. GPS coo dina es
allowed us o classi y he a ms in he con ol g oup ou side
he WH and he ea men g oup inside he WH a ea (Table 2).
Fu he mo e, we calcula ed he dis ance in kilome e s om each
a m o he closes bo de o he WH a ea.
We augmen ed he da ase wi h en i onmen al a iables, such as
he p edic ed na u al ee co e om Bas in e al. (2019), which
p o ides a con enien summa y index o na u al condi ions
ac oss Colombia’s co ee-g owing egion. The au ho s p o ide
a high- esolu ion, global map o po en ial ee co e wi hou
human impac . Fo his, Bas in e al. (2019)useda andom
o es machine-lea ning algo i hm, ained on p o ec ed a eas all
a ound he wo ld, o p edic —based on o e 200 en i onmen al
and na u al a iables— he na u al ee co e consis en ly glob-
733
TABLE 2 Summa y s a is ics.
Mean S d. De Min Max
Va iable Pooled Inside Ou side Pooled Inside Ou side Pooled Inside Ou side Pooled Inside Ou side
A. Co ee a ms a e WH s a us
To al income 902.80 980.37 783.81 989.03 1.111.834 751.83 125 125 125 8000 8000 5100
In ol emen in
ou ism
ne wo ks
0.26 0.26 0.25 0.44 0.44 0.44 0 0 0 1 1 1
Mon hly
paymen o PES
19.05 21.57 16.44 36.77 40.76 33.47 0 0 0 250 250 201
Pa icipa ion in
PES schemes
0.5 0.52 0.46 0.50 0.50 0.50 0 0 0 1 1 1
O ganic co ee
p oduc ion
5.34 5.88 4.51 22.51 23.58 20.83 0 0 0 100 100 100
Age 54.41 55.97 52.05 12.25 11.50 13.02 20 20 20 86 86 79
Educa ion 3.45 3.46 3.44 1.73 1.73 1.73 1 1 1 7 7 7
Na u al ee
co e po en ial
82.37 82.50 82.18 11.678 11.53 11.91 38.05 52.55 38.05 100 100 100
Al i ude 1493.41 1475 1521 270.49 249.90 298.14 811 811 989 2391 2391 2262
Tempe a u e 19.83 19.6 20.19 1.94 1.60 2.33 17 17 17 28 27 28
Membe ship 4.05 4.01 4.02 2.64 2.59 2.74 0 1 0 10 10 10
Fa m owne ship 0.96 0.99 0.92 0.19 0.09 0.28 0 0 0 1 1 1
Dis ance o
bo de
0.23 1.01 2.14 1.82 0.96 1.02 0 0 0 4.15 4.15 4.13
Loca ion
(la i ude)
4.98 4.98 4.99 0.19 0.17 0.21 4.50 4.51 4.50 5.31 5.29 5.31
Loca ion
(longi ude)
−75.87 −75.85 −75.88 0.12 0.12 0.12 −76.13 −76.13 −76.10 −75.59 −75.59 −75.62
Bounda y ixed
e ec s
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Land a ea pe
a m
9.89 10.17 9.46 12.39 12.96 11.50 1 1 1 90 90 80
Co ee land 5.62 5.47 5.85 6.65 5.93 7.65 0.07 0.07 0.07 56 33.75 56
N337 204 133
B. Co ee p oduc ion be o e WH
Co ee c ops in
2007 (pe
municipali y)
61 2264 2778 9268
N90
Co ee c ops
be o e 2011
1.9 1.8 2.0 4.90 4.06 5.38 0.01 0.01 0.01 261 100 261
Volume o co ee
be o e 2011
2.1 2.0 2.2 5.46 4.67 5.91 0.01 0.01 0.01 281 126 281
N75077 29552 45525
No e: A. Da a based on su ey o 337 co ee a me s in he Wes e n and Cen al moun ain anges o The Andes in Colombia. B. Da a om he Na ional Ag icul u al
E alua ion (Minis y o Ag icul u e 2022) and he Thi d Na ional Ag icul u al Census (Dane 2016).
734 Ag icul u al Economics,2025
ally wi hou human impac (wi h a spa ial esolu ion o 30 a c
sec). The p edic i e a iables we e hen summa ized in ac o
a iables local, cap u ing annual mean empe a u e, annual
p ecipi a ion, p ecipi a ion seasonali y, mean empe a u e o
he we es qua e , p ecipi a ion o he d ies qua e , o ganic
ca bon s ock om 0 o 15 cm, dep h o bed ock, sand con en
om 0 o 15 cm, ele a ion, and hill-shade. The andom o es
machine-lea ning algo i hm linked hese 10 a iables o p edic
he na u al ee co e po en ial in pe cen age. We ma ched his
da a o ou a m-le el da ase o es whe he we would ind any
discon inui y in he na u al ege a ion ac oss he WH bo de .
We added o he global map o ee co e po en ial a laye in
GIS so wa e wi h he GPS loca ion o each a m. The machine-
lea ning algo i hm o Bas ian e al. (2019) p o ided he sco e o
ee co e po en ial o he geog aphical coo dina es o each a m.
The smoo h ansi ion in his p e-exis ing ou come con i ms he
en i onmen al homogenei y in he WH bo de egion.
Da a ela ed o co ee c ops be o e he WH s a us is based on wo
sou ces. Fi s , we used he da abase om he Minis y o Ag icul-
u e (2022) ha includes he annual epo o na ional ag icul u al
p oduc ion be ween 2007 and 2018 o each municipali y in
Colombia. We ma ched he da a o co ee p oduc ion in 2007
wi h he GPS loca ion o co ee a ms om ou sample and he
coo dina es o he WH bo de h ough GIS so wa e. The analysis
allows us o illus a e he homogenei y o co ee p oduc ion
a ound he adi ional co ee egion be o e he WH s a us. Second,
he Thi d Na ional Ag icul u al Census p o ided da a on bo h
hec a es o co ee c ops and co ee p oduc ion in ons (Dane 2016).
Fu he mo e, om he same Census, we ex ac ed da a ela ed
o he yea o he co ee plan a ion and he illage whe e each
co ee plan a ion was loca ed. Dane (2016) epo ed 93,500 co ee
ields in he adi ional co ee egion among he depa men s o
Caldas, Quindio, Risa alda, and Valle del Cauca. Acco ding o
Dane (2016), 80% o co ee c ops in he adi ional co ee egion
we e plan ed be o e he WH s a us. We used his sample o 75,000
co ee c ops o es homogenei y in co ee p oduc ion inside and
ou side o he WH a ea be o e he WH s a us.
Finally, we impo ed he sa elli e image and da a o nigh -
ime ligh s in 2009, p ocessed by he Na ional Oceanic and
A mosphe ic Adminis a ion om he Depa men o Comme ce
o USA Go e nmen , and collec ed by US Ai Fo ce Wea he
Agency (NOOA 2014). We designed a GIS map o illus a e he
dis ibu ion o mode n u ban se lemen s, which helps o explain
he agmen a ion among he six discon inuous WH subzones
wi hin he bigge a ea o he adi ional co ee egion.3
3.2 Empi ical S a egy
As he WH s a us es ablishes a spa ial bounda y, we es ima e i s
causal e ec on income using a RDD4; o de ails, see Wueppe
and Finge (2023). RDD allows us o assess he ea men e ec
o WH s a us on co ee g owe s’ income in he adi ional
Colombian co ee egion. The cu o is he WH bo de be ween
simila co ee a ms. We examine whe he he WH bo de leads
o discon inui y in he income o co ee g owe s, acco ding o
hei loca ion and he known co a ia es, la i ude, and longi ude,
om he geo e e ence o each co ee a m. We included a ms less
han 5 km om he WH bo de , e alua ing he ea men e ec
on a homogenous g oup close o he WH bounda ies. The con ol
g oup includes a ms up o 5 km o he bo de ou side o he
WH, while he ea men g oup included a ms up o 5 km o he
bo de inside he WH, loca ed in he co ee egion in he Andes.
RDD educes he bias by choosing he sample wi h obse a ions
close o he bo de and dismissing obse a ions loca ed a
away om he cu o . The sample o obse a ions close o he
bo de allows us o analyze a ms in simila geog aphical and
en i onmen al condi ions.
A cen al assump ion o he RDD is ha he e is no ea men
con ounding. The only di e ence is he h eshold o he WH
bo de . We explici ly es ed he RDD assump ion, e alua ing dis-
con inui y in di e en con ounding a iables, ollowing he bes
p ac ices sugges ed by Alix-Ga cia e al. (2018) and Wueppe and
Finge (2023). We es ed en i onmen al a iables like he na u al
ee co e po en ial, al i ude, and empe a u e. In addi ion, we
es ed discon inui y in social a iables like age, educa ion, and
a m owne ship. We es ed homogenei y in co ee c ops (hec a es)
and olume in co ee p oduc ion ( ons) be o e he WH s a us.
Fu he , we e alua ed discon inui y acco ding o poli ical bo de s
be ween depa men s. Finally, we es ed he RDD assump ion by
es ima ing “placebo” bo de s.
The WH bo de o ms a de e minis ic and discon inuous unc ion
o co a ia es (longi ude and la i ude) (Ca aneo, Id obo, and
Ti iunik 2019). The wo-dimensional spa ial iden i ica ion o
longi ude and la i ude allowed us o pe o m a mo e obus
in e p e a ion ac oss he bounda y (Keele and Ti iunik 2015). The
h eshold is he bo de d awn by he geog aphical coo dina es o
he WH s a us, which di ides wo simila g oups o co ee a ms.
E e y co ee a m ecei ed a sco e acco ding o i s geog aphic
loca ion (Dell 2010). The dis ance sco e a he cu o is ze o.
Dis ance o he igh o ze o e e s o a ms inside he WH a ea,
and dis ance o he le o ze o e e s o a ms ou side o he WH.
The sco e o he dis ance o a ms inside he WH a ea o he WH
bo de ecei es a posi i e sign, and he dis ance o a ms loca ed
ou side o he WH a ea a e ma ked wi h a sco e wi h a nega i e
sign o he ep esen a ion in he igu es. RDD analyzes discon-
inui y be ween obse a ions wi h he minimum sco e eligible
by he WH s a us and obse a ions wi h de icien sco es o be
included in he WH a ea. The e o e, RDD es ima es causal ea -
men e ec s o a ms nea he bo de . Finally, “Dis ance” should
be con inuously dis ibu ed nea he cu o , ensu ing he p esence
o obse a ions close o he bo de (Calonico e al. 2014).
The RDD iden i ies causal ea men e ec s close o he h eshold
ha de e mines ea men assignmen (gi en he alidi y o all
iden i ying assump ions, such as ha he bo de is exogenous,
e c.). RDD is needed in ou analysis as a anspa en me hod
o e lec he causal ea men e ec o he WH s a us, a he
han using he simple compa ison o mean incomes inside and
ou side he WH bo de . Compa ing means inside and ou side he
WH bo de would gi e a highly biased es ima e o he causal
income e ec . The es ima ion o RDD does no in end o es ima e
di e ences wi hin each subsample, bu be ween bo h g oups.
The jump a he h eshold be ween he con ol g oup and he
ea men g oup is a anspa en alida ion o RDD.
Fi s , we used he ha e sine o mula (Aldie i and Cince a 2007)in
Equa ion (2) o calcula e he dis ance om he geog aphic coo -
735
TABLE 5 Tes o s abili y— ea men e ec de i a i e (TED) o
o al income.
Polynomial o de Coe icien S.E p alue
Linea −52.82 377.92 0.889
Quad a ic 656.84 1547.87 0.671
Cubic −2932.39 3241.25 0.366
No e: This able epo s es o s abili y TED o o al income. The signi icance
o his es is assessed wi h he p alue o he es ima ed coe icien . Ke nel:
iangula . Op imal bandwid h choice: minimum squa e e o o RD (mse d).
None o he esul s a e signi ican , acco ding o p alue.
4.4 Ke nel-Weigh ed Local Polynomial
We es ed discon inui y by es ima ion o ke nel-weigh ed local
polynomial a he WH bo de in Figu e 15. The non-pa ame ic
es ima o has pa ame e s ha con ol he local weigh s and
he size o he neighbo hood used o es ima ion. The me hod
combines mul iple eg ession models in k-nea es -neighbo -
based model. We con i med a jump in o al income ac oss he WH
bo de .
4.5 Endogenous So ing
We es ed endogenous so ing by a donu es in Figu e 16 (Ba eca
e al. 2016), whe e we igno ed he obse a ions loca ed wi hin 1
km o he bo de , whe he inside o ou side o he WH bo de . The
donu es sol es conce ns abou he sys ema ic so ing o co ee
a ms a ound he WH a ea, which could dis o he es ima ions
o he RD e ec o he WH bo de .
We complemen ed he analysis o po en ial endogenous so ing,
inqui ing abou he mig a ion o co ee g owe s owa d he WH
a ea and ising income in co ee a ms. We examined a m sales
eco ds as empi ical e idence o changes in he owne s o he
land. Howe e , we did no ind adjus men s in he pa e ns o
land sales. The WH bo de does no ollow a speci ic e i o ial
di ision be ween na i e popula ion g oups, and he whole a ea
is known as he co ee egion. Co ee g owe s do no mig a e
hei co ee p oduc ion o ano he place, e en i he economic
condi ions change o e ime. Unlike co ee g owe s, ha es e s5
used o mig a e o he co ee egion du ing he ha es ing pe iod.
Tempo al mig a ion o ha es e s sa is ies he need o seasonal
labo o collec he beans om he co ee ees (Ca iazo e al.
2002). Ra he , amilies o co ee g owe s do no mo e co ee
p oduc ion om one loca ion o ano he . The pe manence o
co ee g owe amilies in he zone helped o build he co ee
cul u al landscape, which is embodied by he oo s o s ong
amily adi ions. Figu es 17 and 18 show us ha p ope y sales
inside he WH a ea ollowed he business cycle as usual wi hou
inc easing ex ensi ely. The e was no a wa e o mig a ion and
in es men s in o he zone a e he WH s a us; a he , local
co ee g owe s u ned hei own a ms in o ou is ic des ina ions.
Figu e 18 shows a m sales wi hou he main ci ies in he egion:
Pe ei a, A menia, and Manizales, conside ing ha adi ional
co ee a ms a e loca ed mainly in u al a eas. The la a m sales
igu es show ha co ee a ms ha e been owned by he same am-
ilies be o e and a e he WH s a us wi hou endogenous so ing
a ound 2011. The WH bo de made he di e ence be ween bo h
sides.
The e was no mig a ion o co ee a me s om ou side o he
WH a ea o inside a e he UNESCO designa ion. Co ee g owe s
emained in hei o iginal a ms inside o ou side o he WH
a ea wi hin he co ee egion, which is bigge han he WH zone.
The mig a ion o co ee g owe s be ween u al a eas is unusual,
acco ding o he FNC (2021).
The adi ion o s aying in he same place o p oduce co ee is
one o he cha ac e is ics highligh ed by UNESCO o include he
adi ional co ee egion in he WH lis . Two socio-economic cha -
ac e is ics helped o build hei cul u al landscape. Fi s , co ee
a me s inhe i ed a long co ee adi ion wi hin he a ms o hei
amilies. The s udy om he Economics Resea ch Depa men
o FNC applied a su ey o he di ec o s o local commi ees
o co ee g owe s, epo ing ha 76% o su ey esponden s
conside ha co ee g owe s emain in he co ee ac i i y because
co ee p oduc ion is a amilia adi ion (FNC 2021). The cul u al
he i age o co ee g owe s a ec s hei willingness o change,e en
du ing he downswing o co ee p ices. The s ong oo s o co ee
amilies called he a en ion o UNESCO, which ema ked he
Au hen ici y o he Colombian Co ee Cul u al Landscape o i s
designa ion like WH: “Aspec s such as adi ions, language and
o he o ms o in angible he i age, ha e been p ese ed, mos ly by
owne s and he local communi y, who ha e a high sense o social
app op ia ion o hei cul u al he i age” (UNESCO 2011). The cu -
en co ee a m owne s in he whole egion a e p edominan ly
amilies who ha e owned hese a ms o gene a ions. The second
cha ac e is ic is ha co ee amilies a e small p oduce s. The
a e age o co ee c ops inside he WH is 2.6 hec a es (Minis y o
Cul u e 2017). Pe manen labo is ep esen ed by he co ee amily
i sel , and he e o e, he WH in Colombia is a li ed landscape.
Co ee amilies do no ecei e la ge-scale ou side in es men s.
Ra he , one o he main c i e ia o PES p og ams is o o e
bene i s o small co ee g owe s, leading o a highe impac
om he economic incen i e o he en i onmen al se ice
p o ision.
4.6 Mechanisms Inc easing Income o Fa me s
In his pa , we explain he ela ionship be ween he WH s a us
and he highe income o co ee g owe s. The eg ession discon i-
nui y e ec o WH s a us on in ol emen in ou ism ne wo ks is
highly signi ican , as p esen ed in Table 6.
The ou ism plan in he documen o he nomina ion o he
CCL o he WH lis pu sued o a oid mass ou ism by p omo ing
sus ainable ou ism wi hin he p oduc i e oca ion o he co ee
land (Minis y o Cul u e 2010). Co ee a ms inside he WH
zone al eady had he a chi ec u al he i age o he co ee cul u al
expe ience. Howe e , ou is s used o go o jus 16 ou o he
47 municipali ies wi h “ e edas” loca ed inside he WH a ea
be o e he WH s a us (Minis y o Cul u e 2009). By con as ,
a e he insc ip ion o he Colombian Co ee Cul u al Landscape
in he WH Lis in 2011, 84% o ou is s who a i ed by ligh o
he co ee egion s ayed a he municipali ies wi h “ e edas”
included in he WH s a us (SITUR 2018). Co ee g owe s ha e
adap ed hei a ms a e he WH s a us in o es au an s wi h
742 Ag icul u al Economics,2025
FIGURE 18 Fa m Sales in he co ee egion 2000–2016 wi hou u ban cen e s.
TABLE 6 RD e ec o WH s a us on mechanisms inc easing income.
Va iable RD es ima o Bandwid h Le sample Righ sample
In ol emen in ou ism ne wo ks 0.39*** 0.925 133 204
(0.075)
Mon hly paymen o ES 17.77*1.085 133 204
(9.245)
Pa icipa ion in PES schemes 0.39*1.405 133 204
(0.223)
O ganic co ee p oduc ion 0.08** 1.485 133 204
(0.035)
Co ee Re enue 340.67** 1.138 133 204
(115.64)
No e: This able epo s RD e ec o WH s a us on mechanisms inc easing income: In ol emen in ou ism ne wo ks, Mon hly paymen o ES, Pa icipa ion
in PES schemes and o ganic co ee p oduc ion. Ke nel: iangula . Op imal bandwid h choice: minimum squa e e o o RD (mse d). Coe icien s signi ican
di e en om ze o a e deno ed wi h he ollowing sys em: *10%, **5%, ***1%.
adi ional gas onomy, spa cen e s, ho els, o handic a shops,
whe e ou is s can eel imme sed in adi ional co ee cul u e.
Fa me s inside he WH zone include ou ism supply and pa ic-
ipa ion in ou ism commi ees in hei ag icul u al ac i i ies a
a highe le el han g owe s ou side o he WH. The WH s a us
allowed he di e si ica ion o adi ional co ee p oduce s’ income
h ough ou ism ne wo ks a ound he Colombian co ee cul u e
(Figu e 19). Tou ism a ound co ee cul u e ensu es he con inui y
o co ee p oduc ion. Tou ism ne wo ks we e complemen ed by
easie access o he WH a eas by inc easing in es men s in oad
in as uc u e a e he WH s a us in 2011. Figu e 20 epo s
he RD e ec o he dis ance om each co ee a m o he
closes ame o he highway ne wo k buil o adjus ed a e
2011. The dis ance ep esen s he kilome e s om co ee a ms o FIGURE 19 RD e ec o WH on in ol emen in ou ism ne wo ks.
743
FIGURE 20 RD e ec o oad ne wo k on o al income in co ee
a ms.
FIGURE 21 RD e ec o WH on pa icipa ion o a me s in PES.
he highway ne wo k. The sign ep esen s he loca ion o co ee
a ms, being ou side (nega i e) o inside (posi i e) o he WH
a ea. Resul s a e consis en wi h Su and Lin (2014), which indica e
a posi i e ela ionship be ween ha ing WH si es and ou ism
ac i i y. The economic oppo uni ies ound by co ee a me s
wi hin he WH a ea con i m he conclusion o Wueppe (2017)
abou mo e likely bene i om WH b anding in u al se lemen s
h ough conse a ion and economic incen i es.
One o he a ac ions o ou is s is he co ee landscape. The
biodi e si y in he Wes e n and Cen al Moun ain anges o he
Andes a ound co ee p oduc ion o e s en i onmen al se ices
o isi o s. Co ee a me s ecei e economic incen i es om
he go e nmen , NGOs, and p i a e i ms i hey ac i ely p o-
ec o es s, wa e sheds, biodi e si y, o he cul u al landscape.
The PESs con ibu e as a mechanism o ensu e he income o
co ee amilies h ough o mal ag eemen s. PES ep esen s bo h
addi ional income o co ee g owe s and he p o ec ion o he
en i onmen . The eg ession discon inui y e ec o WH s a us
on he pa icipa ion o co ee g owe s in PES is signi ican a
10% in Table 6. The e is a jump ac oss he bo de o he WH
a ea acco ding o pa icipa ion in PES schemes (Figu e 21). PES
can o e an addi ional sou ce o income o co ee g owe s
acing unce ain y o p ice ola ili y in he co ee ma ke . The
eg ession discon inui y e ec o he WH bo de on mon hly PES
is signi ican a 10% in Table 6. The RDD es ima o o he mon hly
PES amoun is 17.7 USD, which ep esen s 2.35% o he sha e in
mon hly income inc ease o 756.97 USD ac oss he WH bo de .
FIGURE 22 RD e ec o WH on he mon hly paymen s o ES
(linea polynomial o de ).
FIGURE 23 RD e ec o WH on he mon hly paymen s o ES
(quad a ic polynomial o de ).
PES has become a mechanism o complemen he income o
co ee g owe s inside he WH a ea (Figu es 22 and 23).
Finally, PES schemes p omo ed cleane p ac ices a ound co ee
c ops inside he WH, leading o he speci ic a ibu es o o ganic
co ee. Highe quali y is ep esen ed by highe p ices o cus-
ome s. Discon inui y o o ganic co ee p oduc ion is signi ican
inside he WH a ea a 5% (Table 6). Highe p oduc ion o quali y
co ee (8.38%) is an addi ional mechanism o inc ease he income
o a me s inside o he WH (Figu e 24). The WH s a us p omo ed
sus ainable p oduc ion and en i onmen al p o ec ion o he WH
a ea, inc easing o ganic co ee p oduc ion in esponse o co ee
p oduce s o he WH s a us. We es ed he RD e ec o he WH
bo de on co ee p oduc ion i sel . Table 6 epo s discon inui y a
he bo de in co ee e enue. Discon inui y shows ha WH s a us
no jus inc eases ou ism and en i onmen al p o ec ion, bu WH
also helps o p ese e co ee p oduc ion.
The economic incen i es o p omo e he balance be ween co ee
p oduc ion, en i onmen al p o ec ion, and social s abili y a e
a icula ed o he sugges ions om UNESCO. The Managemen
Plan o he Co ee Cul u al Landscape o Colombia by UNESCO
(2009) men ions he legal ools o de end he ecogni ion o
“Ca é de Colombia” as P o ec ed Geog aphic Indica ion (PGI).
These guidelines call o he in e en ion o go e nmen s o
de end he unique social and en i onmen al condi ions o he
744 Ag icul u al Economics,2025
FIGURE 24 RD e ec o WH on o ganic co ee p oduc ion.
WH. PES answe o his ini ia i e, d awing g een co ido s
wi hin he WH bo de s and nex o he i e basins. Mos PES
p og ams a ge co ee a ms inside he WH a ea. PES p og ams
include communica ion pla o ms ha connec co ee a me s
wi h dono s, like Banco2 (2024)h ps://banco2.com/. Dono s
could choose he loca ion o he co ee a m, and hey could ead
a b ie e iew o he co ee amilies.
Colombia is well known o he associa i e model among co ee
p oduce s by he Na ional Fede a ion o Co ee G owe s (NFC),
which has a long adi ion since 1927. NFC is one o he main
ac o s ha p omo e he WH. PES p og ams like Banco2 and
Re o es a ion o Magdalena Ri e used o p o ide he economic
incen i e h ough he NFC such as in e media y, because o NFC
has he go e nance s uc u e o ensu e he channels o commu-
nica ion and inancial alloca ion wi h local co ee g owe s.
WH s a us called he a en ion o go e nmen , p i a e sec o ,
ou is s and NGOs o e he WH a ea. Go e nmen and p i a e
i ms designed ins umen s like PES schemes o en i onmen-
al p o ec ion o co ee cul u al landscape inc easing mon hly
income in 2.35%. Cleane p ac ices a ound co ee c ops inside
he WH inc eased o ganic co ee p oduc ion (8.38%). Tou ism
ac i i ies in co ee a ms inc eased 39%. WH s a us led o PES
schemes and ou is ic ne wo ks a ound co ee a ms, which
became mechanisms o inc ease income o co ee g owe s.
5 Policy Implica ions
Ou indings ha e implica ions o policy making. I is much
deba ed whe he WH s a us is ac ually impac ing local li eli-
hoods. Fo u al communi ies, he e is usually high hope ha
WH insc ip ion o a si e will b ing economic bene i s, bu he
empi ical e idence is mixed so a (Wueppe 2017; Gao and Su
2019). Pa icula ly o Colombia’s co ee a me s, incomes a e
ola ile, bo h because o ma ke luc ua ions and because o
en i onmen al shocks ha a e exace ba ed by clima e change.
Ou inding ha WH s a us makes hem inancially be e is
encou aging. Resul s can p o ide inpu o policy make s o ake
ad an age o WH s a us, leading o esilience o he adi ional
ag icul u al sec o in de eloping coun ies. The main indings
p omo e esea ch oppo uni ies abou he alloca ion o PES and
hei cha ac e is ics.
The case o he Colombian Co ee Cul u al Landscape can
encou age policymake s in o he de eloping coun ies o ake
ad an age o adi ional ag icul u al p oduc ion. Co ee p oduc-
ion ep esen s a cul u al iden i y in Colombia. Co ee a me s
con inue p oducing co ee because hey a e en enched in he
iden i y o being co ee p oduce s as an idiosync a ic ea u e.
The adi ional Colombian co ee egion con i ms ha policy-
make s should conside de eloping and celeb a ing local cul u al
cosmogonies o handle economic c ises. In addi ion, WH s a us
and PES schemes a e examples o how local co ee g owe s
could bene i om global ini ia i es om in e na ional o gani-
za ions like UNESCO o o eign dono s. This is an oppo uni y
o s eng hen adi ional ag icul u al p ac ices. Co ee g owe s
need addi ional policies o p o ec adi ional co ee p oduc ion
om land use change. En i onmen al se ices a e h ea ened by
al e na i e economic ac i i ies, he eby a ec ing he ecosys em
equilib ium. The WH s a us p omo ed policies o en i onmen al
p o ec ion and cul u al ecogni ion, which became sou ces o
complemen a y economic ac i i ies and educed he nega i e
e ec o he p ice c isis on co ee g owe income. Fu he mo e,
policymake s should be awa e o he isk o p o iding un ai
dis ibu ion o PES. O he wise, en i onmen al p o ec ion o he
WH a ea could lead o nega i e economic consequences o
a me s ou side o he WH s a us because ou ism and o he
economic sou ces ocus on a pa icula a ea.
The balance be ween en i onmen al p o ec ion and highe
income is no pe ec in he Colombian Co ee Cul u al Land-
scape. The dis up ion o he WH bo de wi hin a municipali y,
lea ing ou some co ee a ms, makes i mo e di icul o con ol
he land use by policymake s. The e o e, allowed economic
ac i i ies in he zone o he municipali y ou side o he WH
pe mea e he in isible WH bo de s, h ea ening he p o ec ed
a eas by he WH s a us and PES schemes.6
6 Conclusion
WH s a us p omo es al e na i e sou ces o income o adi ional
co ee p oduce s. Co ee a ms inside he WH bo de ecei ed
highe income han a ms loca ed ou side he WH bo de . The
inc eased income o co ee g owe s is suppo ed h ough h ee
main mechanisms. Fi s , co ee a me s in he WH a ea became
highly in ol ed in ou ism ne wo ks because he WH s a us
p omo ed u al ou ism in he egion. The in e es o ou is s
in eme ging local co ee p oduc ion adi ions enabled co ee
a me s o complemen hei income. The go e nmen and he
uni e si ies launched s a egies o ake ad an age o adi ional
co ee p oduc ion as a landma k. Local ne wo ks o co ee g ow-
e s adap ed hei a ms o ecei e ou is s who a e a ac ed by he
co ee cul u e; he e o e, such ou ism ope a es wi hin he co ee
p oduc ion ac i i ies. The daily ou ine o g owe s includes exo ic
p ac ices, which a ac he in e es o he ou is communi y.
None heless, go e nmen s and p i a e i ms s ill ha e a long way
o go o ully ake ad an age o his label.
Second, he co ee egion has become a ecipien o PES p og ams
o p o ec he equilib ium o ecosys ems om land use change
745
a e he co ee p ice c isis. PES supplemen ed he lack o income
a e he co ee c isis and p o ided incen i es o p o ec he Co ee
Cul u al Landscape highligh ed by UNESCO h ough WH s a us.
Finally, p omo ing u al ou ism and en i onmen al p o ec ion
p opi ia ed a con ex o o ganic co ee p oduc ion inside he
WH a ea. The di e en speci ica ions o RDD me hods show a
jump in o al income a he WH bo de . The jump coincides
wi h discon inui y a he bo de in he in ol emen o co ee
p oduce s in ou ism ne wo ks and hei pa icipa ion in PES and
o ganic co ee p oduc ion. Thus, WH s a us could be conside ed
an e ec i e incen i e o inc ease o al income in u al a eas.
Mo eo e , he e is no e idence o endogenous so ing in he co ee
egion, as mos co ee a ms a e handed down he gene a ions
and no sold and bough on a ma ke . Figu es 17 and 18 show
he lack o changes in he owne s o he land like empi ical
e idence o he pa e ns o land sales. The discon inui ies we
es ima e e lec ha WH s a us made a di e ence in he income
o co ee a me s inside o he WH a ea, which is qui e sizable
and impo an o he a me s.
Because his s udy elies on a c oss-sec ional da ase , a ew
limi a ions a e implied and shall be discussed. Fi s , he main
mechanism o how WH designa ion inc eases he a me s’
income is inc eased ou ism e enues. This aises he conce n
o spillo e s, ei he posi i e ones (like he WH a ac ing ou is s
and his could also gene a e income oppo uni ies nea by, bu
ou side he WH) o nega i e ones (like he WH a ac ing ou is s
who would o he wise spend hei money in he places nea by, bu
ou side he WH). These spillo e s a e unlikely o be subs an ial,
as he inc ease in ou ism income is likely a mix o new isi o s
who a e Only awa e o / in e es ed in he egion because o he
“WH p omo ion” (Wueppe and Pa y 2016) and because hese
isi o s a e likely spending mo e money because hey hemsel es
ha e highe incomes. This sugges s li le in e ac ion wi h he
egions ou side he WH si e, ei he posi i e o nega i e. Ano he
limi a ion is ex e nal alidi y. Ou esea ch design, he spa ial
RDD, a guably achie es high in e nal alidi y bu in a e y local
ashion ( igh a he bounda y). I is unclea how much his can
be ex apola ed spa ially, o example, all he way o he cen e o
he WH si e. Howe e , we also ha e no e idence sugges ing ha
he e ec a he bounda y is a special case o ha he con ex is
as ly di e en om o he pa s o he egion.
Ou indings lead o u he esea ch ques ions. Fo ins ance, is
he highe income o co ee p oduce s in addi ion o he egula
alloca ion o esou ces, o is his a ze o-sum game, whe e o he
p oduce s a e losing economic bene i s? A e he e up-and-down
swings in income esilience in he long e m a e he WH s a us?
A e he e o he cases o income esilience among ag icul u al
p oduce s wi hin WH si es? Does he adap a ion o a ms o
ou ism a ec co ee p oduc i i y? O e all, he case o he Colom-
bian Co ee Cul u al Landscape shows us ha he in e na ional
ele ance o he WH s a us, linked o o he mechanisms, such
as ou ism ne wo ks and PES, is e ec i e o p omo ing highe
income among co ee g owe s a e he co ee p ice c isis.
The en i onmen al condi ions close o he WH bounda ies a e
qui e simila . Na u al ee co e po en ial pe cen age desc ibes
he same le els bo h inside and ou side o he WH bo de ; he e-
o e, he e a e no no able di e ences in he na u al en i onmen
ac oss he bo de . This ea u e con i ms ha he a ms in he con-
ol g oup and he ea men g oup ha e he same en i onmen al
oppo uni ies in co ee p oduc ion. The only majo di e ence
be ween hem is hei loca ion inside o ou side o he WH bo de .
Acknowledgmen s
Da id Wueppe acknowledges suppo by he Ge man Resea ch Founda-
ion (DFG), unde Ge many’s Excellence S a egy EXC 2070, G an No.
390732324— PhenoRob.
Open access unding enabled and o ganized by P ojek DEAL.
Con lic s o In e es
The au ho s decla e no con lic s o in e es .
Da a A ailabili y S a emen
The da a ha suppo he indings o his s udy a e a ailable on eques
om he co esponding au ho . The da a a e no publicly a ailable due o
p i acy o e hical es ic ions.
Endno es
1The nigh ligh s map in Figu e A3 in he Suppo ing Appendix illus a es
he dis up ion o some mode n in as uc u e, leading o discon inui y
in he WH a ea. Howe e , co ee a ms a e loca ed inside and ou side o
he WH a ea.
2The “ e edas” a e he minimum e i o ial s uc u e in he u al a ea
in Colombia, and hey a e compa ed o he neighbo hoods o u ban
se lemen s. The municipali y is in eg a ed by a g oup o e edas.
3The esul ed nigh ligh s map in he co ee egion is a ailable in he
Suppo ing Appendix in Figu e A3.
4The spa ial eg ession discon inui y design is based on c oss-sec ional
da a and he assump ion ha he e is a con inuously dis ibu ed a iable
wi h a sha p h eshold alue ha is exogenous o he ou comes o in e es
and ha de e mines ea men assignmen (e.g. dis ance o he WH
bo de and he bo de is he h eshold). RDD is no simila o o he
me hods like di e ence in di e ences (di -in-di ), which is based on
he a ailabili y o panel da a and he assump ion ha he e is a ea men
and a con ol g oup ha we e on pa allel pa hs be o e ea men , such
ha he di -in-di be ween ea men and con ol g oup be o e and a e
ea men iden i ies he causal ea men e ec . Fo an o e iew o such
esea ch designs, see Abadie and Ca aneo (2018).
5Labo in co ee p oduc ion is di ided in wo ca ego ies: pe manen
labo and seasonal labo . Ha es e s ep esen seasonal labo du ing
he ha es ing pe iod o collec he beans om he co ee ees. Co ee
p oduc ion has wo ha es ing pe iods: he small one is called ‘mi aca’
be ween Ap il and May, which ep esen s 30% o annual co ee p oduc-
ion. The main ha es ing pe iod is be ween Oc obe and No embe
wi h 70% o annual co ee p oduc ion. Co ee g owe s a e a me s,
who manage he whole p ocess o co ee p oduc ion since he plan ing
p ocess o co ee ees un il he pos -ha es ing p ocess. Pe manen
labo use o be ep esen ed by he amilies o co ee g owe s. Co ee
cul u al landscape is mainly embodied by co ee g owe s. Ha es e s
come o co ee egion om di e en egions and cul u es, e en om
o he coun ies and s ay in he zone jus o he ha es ing pe iod.
6O icial da a om ou ism ac i i y be o e he WH s a us was epo ed
a he le el o municipali y a he han e eda le el. The e a e some
municipali ies wi h e edas inside he WH a ea, and ano he e edas
ou side o he WH zone. The e o e, da a om ou ism au ho i ies is no
clea abou which sha e o his municipali y da a belongs o he WH
a ea.
746 Ag icul u al Economics,2025
Re e ences
Abadie, A., and M. Ca aneo. 2018. “Econome ic Me hods o P og am
E alua ion.” Annual Re iew o Economics 10: 465–503.
Aldie i, L., and M. Cince a. 2007. “Geog aphic and Technological R&D
Spillo e s Wi hin he T iad: Mic o E idence F om US Pa en s.” Jou nal
o Technology T ans e 34: 196–211.
Alix-Ga cia, J. M., K. R. E. Sims, V. H. O ozco-Ol e a, L. E. Cos ica,
J. D. Fe nández Medina, and S. Romo Mon oy. 2018. “Paymen s o
En i onmen al Se ices Suppo ed Social Capi al While Inc easing Land
Managemen .” P oceedings o he Na ional Academy o Sciences 115, no. 27:
7016–7021. h ps://doi.o g/10.1073/pnas.1720873115.
Banco2. 2024. Banco2. h ps://banco2.com/.
Ba eca, A., J. Lindo, and G. Waddel. 2016. “Heaping- Induced Bias in
Reg ession Discon inui y Designs.” Economic Inqui y 54, no. 1: 268–293.
h ps://doi.o g/10.1111/ecin.12225.
Bas in, J.-F., Y. Finegold, C. Ga cia, e al. 2019. “The Global T ee
Res au a ion Po en ial.” Science 365, no. 6448: 76–79.
Beckman, J., and D. E. Schimmelp ennig. 2015. “De e minan s o Fa m
Income.” Ag icul u al Finance Re iew 75, no. 3: 385–402. h ps://doi.o g/
10.1108/AFR-06-2014-0019.
Calonico, S., M. Ca aneo, M. Fa el, and R. Ti iunik. 2019. “Reg es-
sion Discon inui y Designs Using Co a ia es.” Re iew o Economics and
S a is ics 101, no. 3: 442–451.
Calonico, S., M. Ca aneo, and M. Fa ell. 2020. “Op imal Bandwid h
Choice o Robus Bias-Co ec ed In e ence in Reg ession Discon inui y
Designs.” Econome ics Jou nal 23, no. 2: 192–210. h ps://doi.o g/10.1093/
ec j/u z022.
Calonico, S., M. Ca aneo, and R. Ti iunik. 2014. “Robus Nonpa ame ic
Con idence In e als o Reg ession Discon inui y Designs.” Econome -
ica 82: 2295–2326. h ps://doi.o g/10.3982/ECTA11757.
Calonico, S., M. Ca aneo, and R. Ti iunik. 2015. “Op imal Da a-D i en
Reg ession Discon inui y Plo s.” Jou nal o he Ame ican S a is ical
Associa ion 110, no. 512: 1753–1769. h p://www-pe sonal.umich.edu/
~ca aneo/pape s/Calonico-Ca aneo-Ti iunik_2015_JASA.pd .
Ca iazo, F., D. L. Hue h, and E. U. Bo e o. 2002. Un Modelo de Au o-
Selección Pa a la Mig ación de T abajado es Hacia la Región Cen al
Ca e e a de Colombia. CEDE.
Ca aneo, M., N. Id obo, and R. Ti iunik. 2019. A P ac ical In oduc ion
o Reg ession Discon inui y Designs Founda ions. Elemen s in Quan i a i e
and Compu a ional Me hods o he Social Sciences. Camb idge Uni e si y
P ess.
Ce ulli, G., Y. Dong, A. Lewbel, and A. Poulsen. 2017. “Tes ing S abili y
o Reg ession Discon inui y Models.” Reg ession Discon inui y Designs:
Ad ances in Econome ics 38: 317–339.
Ci u . 2021. Regis o de P es ado es de Se icios Tu ís icos [Da abase
o Supplie s o Tou is ic Se ices]. Minis y o Comme ce, Indus y
and Tou ism o Colombia. Re ie ed om h p://www.ci u .go .co/
es adis icas/d _p es ado es_his o ico/depa men /41/1/6? =1#gsc. ab=
0.
Coch an, W. G. 1977. Sampling Techniques.3 ded.JohnWiley&Sons.
Cong ess o Colombia. 2022. Ley 2245 de 11 de julio de 2022.
Dane. 2016. Thi d Na ional Ag icul u al Census. h ps://mic oda os.dane.
go .co/index.php/ca alog/513.
Dane. 2020. Visualizado de Da os Sec o Tu ismo. S a is ics Nacional
Depa men . Re ie ed om h ps://si ios.dane.go .co/ u ismo.
Dell, M. 2010. “The Pe sis en E ec o Pe u’s Mining Mi a.” Econome ica
78, no. 6: 1863–1903.
Dong, Y., and A. Lewbel. 2015. “Iden i ying he E ec o Changing
he Policy Th eshold in Reg ession Discon inui y Models.” Re iew o
Economics and S a is ics 97, no. 5: 1081–1092.
Fauconbe g, A., P. Be hon, and J. Be hon. 2017. “Re hinking he
Ma ke ing o Wo ld He i age Si es: Gi ing he Pas a Sus ainable Fu u e.”
Jou nal o Public A ai s 18, no. 2: a No. e1655. h ps://doi.o g/10.1002/
pa.1655.
FNC – Na ional Fede a ion o Co ee G owe s. 2021. “Demog a ía,
Inclusión Social y Pob eza en los Hoga es Ca e e os (1993-2020).” Ensayos
Sob e Economía Ca e e a 35: 7–34.
Fu, Y., K. Xiong, and Z. Zhang. 2020. “Ecosys em Se ices and Eco-
logical Compensa ion o Wo ld He i age: Li e a u e Re iew.” Jou nal
o Na u e Conse a ion 60: 125968. h ps://doi.o g/10.1016/j.jnc.2021.1259
68.
Gao, Y., and W. Su. 2019. “Is he Wo ld He i age Jus a Ti le o Tou ism?”
Annals o Tou ism Resea ch 78: 102748.
Gelman, A., and G. Imbens. 2019. “Why High-O de Polynomials Should
No be Used in Reg ession Discon inui y Designs.” Jou nal o Business &
Economics S a is ics 37, no. 3: 447–456.
Huang, C.-H., J.-R. Tsau , and C.-H. Yang. 2012. “Does Wo ld He i age
Lis Really Induce Mo e Tou is s? E idence F om Macau.” Tou ism
Managemen 33, no. 6: 1450–1457.
Hube man, D. 2009. “Wo ld He i age Si es and he In eg a ion o Bio-
di e si y In o an In e na ional PES F amewo k.” Jou nal o Sus ainable
Fo es y 28, no. 3-5: 456–480.
Keele, L., and R. Ti iunik. 2015. Geog aphic Bounda ies as Reg ession
Discon inui ies. Poli ical Analysis. Camb idge Uni e si y P ess, 23, no. 1:
127–155.
K onenbe g, J., and K. Hubaceck. 2013. “Could Paymen s o Ecosys em
Se ices C ea e an ‘Ecosys em Se ice Cou se’?” Ecology and Socie y 18,
no. 1. h p://www.js o .o g/s able/26269284.
McC a y, J. 2008. “Manipula ion o he unning a iable in he eg ession
discon inui y design: a densi y es .” Jou nal o Econome ics 142, no. 2:
698–714. h ps://doi.o g/10.1016/j.jeconom.2007.05.005.
Minis y o Cul u e. 2009. Managemen Plan o Colombian Co ee Cul u al
Landscape. Minis y o Cul u e.
Minis y o Cul u e. 2010. “Co ee Cul u al Landscape:AnExcep-
ional Fusion o Na u e, Collec i e Human E o and Cul u e.”
Nomina ion Documen . Re ie ed om h ps://whc.unesco.o g/uploads/
nomina ions/1121.pd .
Minis y o Cul u e. 2017. Paisaje Cul u al Ca e e o. Re ie ed om h p://
paisajecul u alca e e o.o g.co/.
Minis y o T anspo . 2024. Repo No. 20243031389302 o he House o
Rep esen a i es. h ps://www.cama a.go .co/si es/de aul / iles/2024-09/
082024R asMinT anspo e.pd .
Minsi y o Ag icul u e. 2022. Ag icul u al E alua ions in
Municipali ies. h ps://www.da os.go .co/Ag icul u a-y-Desa ollo-
Ru al/E aluaciones-Ag opecua ias-Municipales-EVA/2pnw-mmge.
Mish a, A., and B. Goodwin. 1997. “Fa m Income Va iabili y and he
Supply o O -Fa m Labo .” Ame ican Jou nal o Ag icul u al Economics
79, no. 3: 880–887.
NOOA. 2014. Nigh ime Ligh s Time Se ies. Na ional Cen e o
En i onmen al In o ma ion. h ps://www.ngdc.noaa.go /eog/dmsp/
downloadV4composi es.h ml.
Nuppenau, E. -A. 2002. P ese a ion o Bio-Di e si y, He e ogenei y o
Fa m P ac ices and Paymen s o Cul u al Landscapes Unde Inhomo-
geneous Na u al Condi ions. X EAAE Cong ess ‘Explo ing Di e si y in
he Eu opean Ag i-Food Sys em’, Spain, Za agoza. In Wo ld Wide Web.
Re ie ed om h p://ageconsea ch.umn.edu/handle/24937.
Sa mien o, J. 2012. “Paisaje Cul u al Ca e e o Del Cen o Occiden e de
Colombia: Un Pa imonio Vi o.” Labo & Engenho, Campinas [B asil] 6,
no. 2: 08–15.
SITUR – Sys em o Tou is ic In o ma ion. 2018. IV In o me Téc-
nico de Es adís icas SITUR PCC. Cáma a de Come cio de A me-
747
nia. Re ie ed om h ps://www. lipsnack.com/ ede aciondeca e e os/
i -in o me- -cnico-de-es ad-s icas-si u -pcc.h ml.
Su, M., Y. Dong, G. Wall, and Y. Sun. 2020. “A Value-Based Analysis o
he Tou ism Use o Ag icul u al He i age Sys ems: Duo ian Ag osys em,
Jiangsu P o ince, China.” Jou nal o Sus ainable Tou ism 28, no. 12: 2136–
2155.
Su, Y. U- W., and H. -L. Lin. 2014. “Analysis o In e na ional Tou is
A i als Wo ldwide: The Role o Wo ld He i age Si es.” Tou ism Manage-
men 40: 46–58.
UNESCO. 1972. Con en ion Conce ning he P o ec ion o he Wo ld
Cul u al and Na u al He i age. UNESCO.
UNESCO. 2009. Managemen Plan Co ee Cul u al Landscape o Colom-
bia. h ps://whc.unesco.o g/en/documen s/115771.
UNESCO. 2011. Co ee Cul u al Landscape o Colombia. In Wo ld Wide
Web. Re ie ed om h p://whc.unesco.o g/en/lis /1121.
Wueppe , D. 2017. “Wha Is he Value o Wo ld He i age S a us o a
Ge man Na ional Pa k? A Choice Expe imen F om Jasmund, 1 Yea A e
Insc ip ion.” Tou ism Economics 23, no. 5: 1114–1123.
Wueppe , D., and R. Finge . 2023. “Reg ession Discon inui y Design
in Ag icul u al and En i onmen al Economics.” Eu opean Re iew o
Ag icul u al Economics 50, no. 1: 1–28. doi.o g/10.1093/e ae/jbac023.
Wueppe , D., and M. Pa y. 2016. “The Wo ld He i age Lis : Which Si es
P omo e he B and? A Big Da a Spa ial Econome ics App oach.” Jou nal
o Cul u al Economics 41, no. 1: 1–21.
Wunde , S. 2005. Paymen s o En i onmen al Se ices: Some Nu s
and Bol s. Ci o . Occasional Pape No. 42. In Wo ld Wide Web:
h p:// he edddesk.o g/si es/de aul / iles/ esou ces/pd /2012/paymen -
ecosys em-se ices-02.pd .
Yang, C. -H., and H. -Y Lin. 2014. “Re isi ing he Rela ionship Be ween
Wo ld He i age Si es and Tou ism.” Tou ism Economics 20, no. 1: 73–86.
Suppo ing In o ma ion
Addi ional suppo ing in o ma ion can be ound online in he Suppo ing
In o ma ion sec ion.
748 Ag icul u al Economics,2025
