Wazza, Melkamu Tadesse; Ayele, Sei e; Shano, Be hanu Kuma
A icle
Place-based app oach o u al de elopmen : E hiopia in
con ex
Economies
P o ided in Coope a ion wi h:
MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel
Sugges ed Ci a ion: Wazza, Melkamu Tadesse; Ayele, Sei e; Shano, Be hanu Kuma (2025) : Place-
based app oach o u al de elopmen : E hiopia in con ex , Economies, ISSN 2227-7099, MDPI,
Basel, Vol. 13, Iss. 3, pp. 1-29,
h ps://doi.o g/10.3390/economies13030061
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/329341
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/
Academic Edi o : Richa d J. Cebula
Recei ed: 12 Decembe 2024
Re ised: 3 Feb ua y 2025
Accep ed: 4 Feb ua y 2025
Published: 22 Feb ua y 2025
Ci a ion: Wazza, M. T., Ayele, S., &
Shano, B. K. (2025). Place-Based
App oach o Ru al De elopmen :
E hiopia in Con ex . Economies,13(3),
61. h ps://doi.o g/10.3390/
economies13030061
Copy igh : © 2025 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license
(h ps://c ea i ecommons.o g/
licenses/by/4.0/).
A icle
Place-Based App oach o Ru al De elopmen : E hiopia
in Con ex
Melkamu Tadesse Wazza 1,2,* , Sei e Ayele 3and Be hanu Kuma Shano 4
1Depa men o Ru al De elopmen , Wolai a Sodo Uni e si y, Sodo P.O. Box 138, E hiopia
2Depa men o P ojec Managemen , PBTA ica College, Addis Ababa P.O. Box 62045, E hiopia
3Ins i u e o De elopmen S udies, Uni e si y o Sussex, B igh on BN1 9RH, UK; [email p o ec ed]
4Depa men o Ag icul u al Economics, Wolai a Sodo Uni e si y, Sodo P.O. Box 138, E hiopia;
[email p o ec ed]
*Co espondence: [email p o ec ed] o [email p o ec ed]
Abs ac : Place-speci ic socioeconomic ea u es a e unique and, unlike i s -na u e geog-
aphy, a e shaped and eshaped by human and ins i u ional in e ac ions. In E hiopia,
howe e , policy hinking has no p og essed much beyond i s -na u e geog aphy, o e -
looking he mul idimensional socio-spa ial o ma ions o u al a eas. This s udy, based
on na ionally ep esen a i e socioeconomic panel da a om 2018/19 and 2021/22, used
a place-based amewo k o explo e he complex na u e o u al de elopmen and i s
ela ionship wi h mul idimensional, place-speci ic key de e minan s, namely u ali y and
en ep eneu ial ecosys ems. Indices o he key a iables we e de eloped by educing
hei dimensions using P incipal Componen Analysis o measu e mul idimensional a i-
ables, including u al de elopmen , and unde ake subsequen examina ions. The s udy
examines he e ec s o he key de e minan s on u al de elopmen using he Fixed E ec s
Ins umen al Va iables–Two-S age Leas Squa es eg ession model, owing o endogenei y
conce ns wi h he key de e minan s. The s udy shows signi ican e ec s o bo h u ali y
and en ep eneu ial ecosys ems on u al de elopmen . I o e s insigh s in o he complex
socio-spa ial o ma ions and explana o y powe o u al con ex s and con ibu es o he
unde s anding o a place-based app oach o u al de elopmen . The s udy also con ibu es
o na ional and sub-na ional s a egies o add ess u al challenges in E hiopia and beyond.
Keywo ds: socio-spa ial; place-based; u ali y; en ep eneu ial ecosys em; FE-2SLS
1. In oduc ion
A ica is p ima ily u al (Ch is iaensen & Mae ens,2022), ye i s uggles wi h c i ical
policy challenges in u al de elopmen (RD). Pa icula ly, Sub-Saha an coun ies lack a clea
ocus on RD, and o en, he e is a na ow emphasis on ag icul u al p oduc i i y. They also
inadequa ely adap o changing socioeconomic condi ions, such as high u al e ili y, u al-
o-u ban mig a ion and egional inequali ies (OECD,2016;OECD/PSI,2020). Mo eo e ,
hey o e look g ow h po en ial h ough a place-based app oach (OECD,2016, p. 181).
Fo ins ance, Cô e d’I oi e and Tanzania p esen a compelling case. Cô e d’I oi e ocused
on ag icul u al p oduc i i y and has achie ed imp essi e esul s in a ming. Likewise,
Tanzania pu sued an explici RD s a egy and achie ed no able economic g ow h since
2000. Howe e , Cô e d’I oi e s ill con on s ex eme po e y and low human de elopmen ,
and Tanzania has s uggled o ansla e g ow h in o imp o ed u al socioeconomic wel a e.
While he labou o ce is g owing apidly, A ican coun ies including E hiopia a e no
o e ing adequa e job oppo uni ies (OECD,2016;OECD/PSI,2020).
Economies 2025,13, 61 h ps://doi.o g/10.3390/economies13030061
Economies 2025,13, 61 2 o 29
Wi h 130 million inhabi an s, and being 77% u al, E hiopia is he second mos pop-
ulous and he i h mos u al coun y in A ica (OECD/PSI,2020;T ading Economics,
2024). I implemen ed he wo ld’s mos p ominen Chilalo Ag icul u al De elopmen
Uni (CADU) and Wolai a Ag icul u al De elopmen Uni (WADU) u al p og ammes i e
decades back (Cohen,1987). Howe e , exis ing u al de elopmen s a egies in E hiopia
o en o e look he di e se socioeconomic con ex s o u al a eas (OECD/PSI,2020), elying
on he de ec i e concep ualiza ion o u al a eas ha classi y a eas as non-u ban spaces,
mainly using popula ion h eshold (FAO,2018). While u al a eas a e shaped and eshaped
by complex in e ac ions be ween human and ins i u ional ac o s a ec ing hei de el-
opmen ajec o ies (Gadde o s & Ande son,2019;Laza o e al.,2019;Zah a e al.,2014),
policy hinking in he coun y has no p og essed much beyond i s -na u e geog aphy,
o e looking he mul idimensional socio-spa ial o ma ions o u al a eas, making i a key
s udy coun y o his opic.
Agains his backg ound, his s udy aims o unde s and he complex socio-spa ial
na u e o u al de elopmen and i s ela ionship wi h key mul idimensional, place-speci ic
de e minan s, namely u ali y and he en ep eneu ial ecosys em, o in o m e ec i e
policymaking. D awing om he concep s, heo ies, and empi ical e idence and using
he Li ing S anda ds Measu emen S udy (LSMS) da a o E hiopia, i es s a cen al
hypo hesis (H) ha mul idimensional, place-speci ic ea u es a e key de e minan s o u al
de elopmen (RD), expounded wi h wo speci ic hypo heses: H
1
: u ali y (RUR) o a place
has a s a is ically signi ican e ec on RD, and H
2
: he en ep eneu ial ecosys em (EE) o a
place has a s a is ically signi ican e ec on RD. In associa ing hese de e minan s di ec ly
wi h RD and quan i ying hei e ec s, o he bes o he au ho s’ knowledge, he e a e no
subs an i e s udies, pa icula ly in E hiopia.
The exis ing knowledge base in unde s anding such socio-spa ial dimensions in
E hiopia is limi ed. The ailu e o imp o e his may esul in di icul y in add essing
he pe sis en socio-spa ial mic o-le el challenges such as inequali ies and mig a ion. Mos
s udies on economic g ow h in E hiopia a e no di ec ly ela ed o he u al con ex ; ins ead,
hey di e om he p esen s udy in a ious common aspec s. They o en ocus on he
mac o-economy, ely on ( eal) GDP as a singula indica o o economic pe o mance, and
ha e no subs an i e conce n in dis inguishing u ban and u al spa ial se ings (Alemu e al.,
2019;Nege a,2021;Tes aye & Bekana,2020;Rao & Bedada,2017). S udies om A ica also
p ima ily emphasize mac oeconomic de e minan s ha unde play he de imen al ole
played by local communi ies and mul idimensional de elopmen aspec s (Nguyen,2023;
Sendi e al.,2022;Thaddeus e al.,2024). Though GDP is a popula economic indica o ,
i alls sho o assessing RD as i emphasizes mone a y alues (Kaˇca e al.,2016). In
u al economies like E hiopia, impo an economic ac i i ies such as subsis ence a ming,
communi y se ices and en i onmen al conce ns a e i al o li elihoods bu a e o en
igno ed in GDP measu emen s. Thus GDP o e looks he essen ial and mul idimensional
aspec s o RD. This s udy add esses he d awbacks o GDP by di ec ly measu ing and
analyzing RD. Mo eo e , GDP da a a e una ailable a he g ass oo s (mic o) le el o many
Sub-Saha an coun ies, including E hiopia, whe e one canno ind hem e en a a egional
le el, usually due o a lack o s a is ical capaci y (Ga es,2013). Thus, di ec measu emen
has an addi ional bene i in ackling da a una ailabili y.
The mic o-le el e idence a ailable in E hiopia ends o emphasize indi idual ac o s
o was me ely collec ed a he sub-na ional le el, o wi hou conce n o endogenei y
p oblems (see K useman e al.,2006;Ogunleye,2010;Panahi,2015;Yilmaz e al.,2010).
The empi ical ep esen a ion o place-speci ic ac o s is commonly limi ed o simple and
single-dimensional a iables such as ma ke access (Hausmann e al.,2021). Unlike some
expe iences in objec i ely analyzing u ali y and en ep eneu ial ecosys ems (Leende se
Economies 2025,13, 61 3 o 29
e al.,2021;Li e al.,2015;Sánchez-Ma eos,2018;Yilmaz e al.,2010), he p esen s udy
iden i ies hese indica o s as mul idimensional place-speci ic eg esso s and quali ies hem
as key explana o y powe s. The au ho s a e unawa e o a s udy ha applied mul idimen-
sional eg esso s as ac o s di ec ly a ec ing RD, o i so, a leas empi ically es ablished
o wha ex en such en i ies explain RD, anywhe e o a leas in E hiopia. This s udy
le e ages quan i a i e panel da a and e ec i ely con on s he challenges o place he e o-
genei ies and endogenei ies, o e ing conside able echnical ad an ages (see G eene,2020;
Woold idge,2020).
In line wi h he shi in he concep s o RD o a e i o ial, “bo om-up” app oach
ha acknowledges u al places (Bee e al.,2020;Ca aneo e al.,2022), his s udy applies
place-based de elopmen modelling and o e s insigh in o he u al con ex ’s complex
socio-spa ial o ma ions. I u he unde sco es he explana o y na u e o he mul idi-
mensional socio-spa ial ea u es in d i ing socioeconomic changes. The s udy in o ms
policymake s and p ac i ione s o “wha ” and “how” o a ge he undamen al socio-
spa ial conce ns easibly. The analy ical amewo k enables he e ec i e iden i ica ion and
measu emen o place-speci ic ea u es and highligh s he spa ial in e play ha can help he
g owing concep ualiza ion o a place-based app oach. Adop ing a place-based app oach
and in eg a ing i in o b oade egional and na ional s a egies will help add ess he unique
u al challenges and enhance policy e ec i eness in E hiopia and beyond.
The es o he pape includes i e pa s. Fi s , a sho accoun o he place-based
app oach is p esen ed. Following his, he de e minan s o RD, da a and me hods, key
indings, discussions, and conclusions a e elabo a ed on.
2. Place-Based App oach and De e minan s o Ru al De elopmen
2.1. Place-Based App oach o Ru al De elopmen : Spa ial Concep s and Empi ical E idence
Ru al de elopmen (RD) is a b oad concep ha inco po a es ag icul u al and non-
ag icul u al ac i i ies, wi h due emphasis on imp o ing he li ing s anda ds o u al socie y
(Singh & Shishodia,2023). The app oaches in add essing RD ha e shi ed om hei sim-
plis ic app oaches o ackle he ‘ alling’ u al a eas wi h u baniza ion, mode niza ion, and
indus ializa ion in o he b oade concep o a e i o ial model and mo e owa d commu-
ni y and place-based app oaches (Bee e al.,2020;Ca aneo e al.,2022) ha emphasize
concep s o space, place, and e i o y. Space is he basic laye emb acing e i o y and
place. I is geog aphically concep ualized wi h loca ion, e i o y, and place (Dua e,2017).
Bo h e i o y and place a e slices o space wi h which people’s (indi idual, g oup, o
social) meaning ul in e ac ions and alues a e asc ibed. The place is a speci ic geog aphic
loca ion whe e people li e, and i exhibi s a ec i e and subjec i e alues. Te i o y on
he o he hand has en o ced alues ha end o go e n he occupan s (pp. 4, 79). The
place-based app oach o u al de elopmen (PBRD) is an app oach o de elopmen ha
seeks o achie e desi ed socioeconomic changes in a gi en geog aphic place such as local
u al a eas o egions (Bee e al.,2020). Though conno ed as place-based, i is ul ima ely
designed o imp o e he well-being o u al people, making people–place duali y unna u al
(By on,2010).
The unde lying p inciple o he place-based app oach is equi y and e iciency. While
he equi y conce n is due o commonly seen une en spa ial de elopmen , he e iciency
conce n is due o ma ke ailu e (Du an on & Venables,2021). In his ega d, he PBRD ap-
p oach ecognizes wo essen ial elemen s: (i) he unique iden i y and sense o place inhe en
in communi ies, and (ii) he ailo ing o local s a egies wi h egional and na ional policies.
The i s elemen ocuses on he hema ic conce ns o a place. This enables le e aging
angible and in angible unique endowmen s and bo om-up planning. The emphasis hus
is on he needs and p essing conce ns o he local ( egional) de elopmen (Manioudis &
Economies 2025,13, 61 4 o 29
Angelakis,2023). In a he e ogeneous con ex o a place, he app oach goes beyond he
adi ional sec o s such as manu ac u ing o inance o u he p omo e he g owing sec o
o al e na i e economies such as c ea i e economies such as c a , gaming, and media.
These sec o s di e om adi ional capi alis models ha in ensi y inequali ies and en-
i onmen al challenges. This economy elies less on na u al esou ces and no ably o e s
job oppo uni ies o young ci izens and women compa ed o adi ional sec o s and helps
e icien ly d i e de elopmen in local a eas (Healy,2020;Manioudis & Angelakis,2023).
The second elemen , i.e., ailo ing, emphasizes he impo ance o he in e ac ion and
in e dependence be ween local and ex a-local ac o s, and ac i e communi y engagemen
(Manioudis & Angelakis,2023). Consequen ly, he place-based app oach encou ages a
socio-spa ial unde s anding o bo h local and b oade ac o s. Be ween he local and he
b oade na ional and global ac o s a e egional ac o s se ing as a pa o a la ge na ional
s a egy aimed a add essing socioeconomic challenges, including egional inequali ies,
economic s agna ion, and unemploymen . Place-o ien ed egional planning hus ecognizes
he signi icance o place-speci ic ea u es such as u ali y and local en ep eneu ship and
inno a ion (Bee e al.,2020). En ep eneu ship is usually a egional phenomenon and
when connec ing and in e ac ing wi h o he componen s o o m an EE, i has a s ong
and e ec i e capaci y o mobilize he capi al, labou , and esou ces o places (O’Conno
e al.,2018). Such mechanisms u ge ha nessing he un apped local po en ial and add essing
social exclusion in a speci ic ( egional) con ex . Thus, he PBRD app oach h ough i s
equi y and e iciency conce ns p omo es sus ainable ( egional) de elopmen (Almusaed &
Almssad,2023).
The p ominen p oponen s o place-based policymaking include he EU, OECD, he
Aus alian go e nmen , and A DB (A ican De elopmen Bank) (A DB/OECD,2015;
Gea ches e al.,2023;OECD/PSI,2020). Wi h he place-based app oach, while he EU
add esses con inen al challenges such as popula ion decline, ageing, and limi a ions in
social se ices (Ba ca e al.,2012;Eu opean Union,2022), he Aus alian go e nmen deals
wi h a ea ma ginaliza ion and egional inequali ies (Gea ches e al.,2023). The applica ion
o he app oach is g owing in A ican coun ies such as Nige ia and Sou h A ica (Abagna
e al.,2024;A DB/OECD,2015;Pugalis & G ay,2016). In Nige ia, o ins ance, while
Lawal and Osayomi (2021) applied c oss-sec ional place-based modelling o COVID-19
social ulne abili y analysis, KTN (2021) conduc ed a place-based inno a ion audi in a
u al s a e. Using a h ee-decade (1990–2020) household da ase om 10 A ican coun ies,
Abagna e al. (2024) also showed he e ec i eness o a place-based policy (Special Economic
Zones—SEZs) on household weal h
1
. An OECD s udy on E hiopia a gued agains he
u al–u ban di ide and ecommended shi ing away om a sec o al-based policy o a
PBRD (OECD/PSI,2020, pp. 22, 102) ha s eng hens u al–u ban linkage and add esses
g owing egional inequali ies (p. 162).
2.2. De e minan s o Ru al De elopmen
De e minan s o economic de elopmen a e complex and a y ac oss and wi hin
coun ies (Nguyen,2023). They a e gene ally g ouped in o economic (e.g., esou ces and
capi al) and noneconomic (e.g., ins i u ions) ca ego ies (Jhingan,2016). Wi hin his b oad
amewo k, se e al ac o s a ec he le el o RD such as esou ce endowmen s, human
esou ces, a ailable capi al, cu ing-edge echnology, and he e ec i eness o ins i u ions
and o ganiza ions (Aga wal e al.,2009;Banaka & Pa il,2018;Singh & Shishodia,2023;
Tae-Hwa & Seung-Ryong,2016). Unde s anding he in e ac ions be ween each o hese
a ious ac o s in spa ial and social se ings is essen ial o a success ul policy (Gadde o s
& Ande son,2019;Laza o e al.,2019).
Economies 2025,13, 61 5 o 29
Recen empi ical esea ch on ac o s a ec ing RD was pe o med in di e en pa s
o he wo ld such as Tu key (Yilmaz e al.,2010), I an (Panahi,2015), and Nige ia (Ogunl-
eye,2010), and e iew wo k in Czech (S aka & Tuzo á,2016), among o he s. E idence
om Tu key’s u al p o ince indica ed ha RD is a ec ed by na u al, land-use, demo-
g aphic, in as uc u e, and socioeconomic s uc u es (Yilmaz e al.,2010). In de e mining
de elopmen in a u al con ex , i.e., RD, esou ce endowmen s such as land and li es ock;
echnology such as imp o ed seed, e ilize s, and machine y; and human capi al such as
gene al and echnical educa ion a e impo an ac o s (Singh & Shishodia,2023). Local
o ganiza ions and ins i u ions gene ally apply o he op imum u iliza ion o p oduc ion
ac o s (Jhingan,2016). In u al Poland, o ins ance, inc eased ins i u ional quali y was
associa ed wi h highe socioeconomic de elopmen (Ba kowiak-Bakun,2018). Mo eo e ,
RD is a ec ed no only by local ac o s bu also by ex a-local ac o s (Kim,2024) such as
u ban agglome a ion (F ick & Rod íguez-Pose,2017), con inen al oad ne wo k (Pa el &
Moldo an,2019), and globaliza ion (Bee e al.,2020).
In unde s anding he de e minan s o u al de elopmen , he socio-spa ial concep u-
aliza ion o u al a eas is also help ul as socio-spa ial ac o s a e among he majo de e -
minan s o RD (Aga wal e al.,2009). Beyond being a geog aphic en i y, space is whe e
socioeconomic phenomena a e pu oge he and shaped. Such socioeconomic in e ac ions
p oduce a place wi h dis inc ea u es ha ac as explana o y ac o s (Bee e al.,2020;
Coe e al.,2013). The dis inc i eness o a place (wi hin and be ween e i o ies) eme ges
om i s physical na u e (e.g., esou ce endowmen , ou is ic and se lemen landscape) as
well as om human ac i i y (e.g., o ms o go e nmen , cul u e, eligion, weal h s a us,
buil en i onmen , social ela ionships, e c.), and he in e ac ion be ween each o hese. A
place is o med o his o ical and exis ing changes wi hin an a ea and elsewhe e (Coe e al.,
2013). Mo eo e , a place is a secu e po ion o space, and one expe iences space h ough a
sense o place—a ac o in geog aphic iden i y ha is acknowledged as in angible capi al
wi hin a geog aphic space (a u al place) ha can a ec labou a ailabili y, en ep eneu -
ship, and commodi y supply (Bol on,1992). Thus, a place wi h socio-spa ial ea u es and
in e ac ions a ec s RD (Bee e al.,2020;Coe e al.,2013;Gadde o s & Ande son,2019;
Zah a e al.,2014).
The majo socio-spa ial ea u es in luencing u al de elopmen include u ali y (RUR)
and he u al en ep eneu ial ecosys em (EE). Ru ali y embodies a unique socioeconomic
iden i y shaped by a s ong sense o place, socioeconomic in e ac ions, and his o ical
con ex (Woods,2011). The u al en ep eneu ial ecosys em ep esen s he in e ac ions
among en ep eneu ial playe s—en ep eneu s, businesses, go e nmen o ganiza ions,
and esea ch ins i u ions— ha can ei he suppo o hinde en ep eneu ial pe o mances
(S am,2015;Wo ld Economic Fo um,2013). While en ep eneu ship is o en d i en by
indi idual ini ia i e, i is meaning ully a ec ed by unique u al con ex s, which, wi h i s
ma ked socio-spa ial ea u e, shapes u al en ep eneu ship di e en ly han gene al/u ban
en ep eneu ship (Muñoz & Kimmi ,2019). Ru al a eas a e cha ac e ized by s ong social
in e ac ion and a e sensi i e o ma ke demands, bu hey ace challenges such as limi ed
access o inance, skills, echnology, in as uc u e, and dis ibu ion channels. Unlike u ban
en ep eneu ship, which ocuses on seizing oppo uni ies, u al en ep eneu ship p ima ily
elies on a ailable esou ces (Asmi e al.,2024;Pa o & Teixei a,2014).
2.3. Ru al De elopmen in E hiopia
Since he mid-1990s, “ag icul u e and u al de elopmen (ARD)” has been he ocus
o E hiopia’s de elopmen s a egy, guided by he ADLI’s (Ag icul u al De elopmen -Led
Indus ializa ion) amewo k. The ADLI’s main objec i e is o inc ease he ag icul u al
p oduc i i y and he p oduc i e capaci y o smallholde a me s. ADLI helped o b ing
Economies 2025,13, 61 6 o 29
o h se e al subsequen de elopmen plans including he Sus ainable De elopmen and
Po e y Reduc ion P og am (SDPRP) and he G ow h and T ans o ma ion Plans (GTP I
and II) (OECD/PSI,2020). Wi hin he ADLI amewo k, he mos no able RD s a egy is he
“Ru al De elopmen Policy and S a egies”, endo sed in 2003; hough i conside s ag oeco-
logical a ia ions (MoFED,2003, p. 16), i ails o acknowledge he o e all socioeconomic
he e ogenei ies o u al a eas. Following he design and implemen a ion o hose plans and
s a egies, albei wi h une en ou comes, he coun y has expe ienced signi ican economic
g ow h, pa icula ly since 2004. The cu en egime, in place since 2019, has made no majo
change o he app oach. Recen ly, he coun y implemen ed a en-yea pe spec i e plan o
he coun y (2021–2030). Howe e , he plan con inues o p io i ize smallholde ag icul u e
and ans o ma ion in o comme cial a ming, al hough p i a e sec o in ol emen emains
limi ed. The plan does no comp ehensi e u al de elopmen model. Ra he , he e is
a high emphasis on u ban de elopmen (see PDC,2020). The o e all policy di ec ions,
hough wi h egis e ed GDP g ow h, ailed o achie e economic ans o ma ion and he
much- equi ed quali y g ow h in E hiopia. A ecen OECD/PSI s udy iden i ied h ee main
challenges o E hiopia: demog aphic, economic, and spa ial. The ongoing popula ion
g ow h, pa icula ly in u al a eas, poses signi ican labou ma ke challenges. E hiopia is
expe iencing a decline in ag icul u al GDP and employmen , ye ag icul u e s ill employs a
ema kable pe cen age o he wo k o ce (73%). Fo mos u al households, c op a ming is
essen ial, con ibu ing o e 70% o hei income, while he non- a m sec o plays a minimal
ole. The OECD/PSI s udy unde sco ed ha he ADLI amewo k s uggles o add ess he
cu en issues (OECD/PSI,2020).
3. The Analy ical F amewo k o Place-Based Ru al De elopmen
This s udy applies a place-based analy ical amewo k which s uc u es he di e en
socio-spa ial ac o s a ec ing u al de elopmen . I ames he ac o s in a speci ic place
aking accoun o he o ma ion o mul idimensional, place-speci ic de e minan s wi h local
and ex a-local in e play. In his con ex , endogenous (local/mic o-le el) and exogenous
(ex a-local/mac o) o ces in e ac wi hin and be ween u al a eas. Such in e ac ions
usually gene a e mul idimensional ac o s unique o a place. As illus a ed in Figu e 1,
hese o ces in luence u al de elopmen (RD), empowe ing g ass oo s communi ies. In his
s udy, he unique ea u es o u al places—speci ically u ali y (RUR) and en ep eneu ial
ecosys ems (EEs)—a e assumed o be key socio-spa ial de e minan s. Thus, u al a eas a e
no me ely i s -na u e geog aphy o non-u ban homogenous spaces. They a e complex
socio-spa ial con ex s in eg al o de elopmen p ocesses. Because economic de elopmen
has a spa ial dimension (Komo ,2020) and places, wi h hei a ying socio-spa ial ea u es
and in e ac ions, a ec RD (Bee e al.,2020;Coe e al.,2013;Gadde o s & Ande son,2019;
Zah a e al.,2014), a u al place wi h a spa ial elemen ac s as an impo an ac o a ec ing
RD (Komo ,2020).
He e, i is impo an o unde sco e he basic no ion unde lying RUR and RD. RUR
ep esen s he unique socio-spa ial cha ac e is ics o a u al place, and RD, on he con a y,
e lec s he le el o de elopmen achie ed. I is he ou come o he de elopmen p ocess
shaped by a ious place-speci ic ac o s, including RUR. While RUR can be quan i ied
h ough aspec s such as ag icul u al engagemen , popula ion size, o co e age o communal
spaces, measu ing RD equi es a di e en se o indica o s. These include e alua ing
p oduc i i y le els (o labou o land) and consump ion le els (bo h ood and non- ood).
Ul ima ely, he concep o RD seeks o cap u e he esul s o a de elopmen p ocess, while
RUR aims o de ine he a ibu es ha cha ac e ize a speci ic a ea.
Economies 2025,13, 61 7 o 29
Economies 2025, 13, x FOR PEER REVIEW 7 o 31
Figu e 1. A place-based analy ical amewo k o he p oposed RD p ocess (au ho s’ p esen a ion).
He e, i is impo an o unde sco e he basic no ion unde lying RUR and RD. RUR
ep esen s he unique socio-spa ial cha ac e is ics o a u al place, and RD, on he con-
a y, eflec s he le el o de elopmen achie ed. I is he ou come o he de elopmen
p ocess shaped by a ious place-specific ac o s, including RUR. While RUR can be quan-
ified h ough aspec s such as ag icul u al engagemen , popula ion size, o co e age o
communal spaces, measu ing RD equi es a diffe en se o indica o s. These include e al-
ua ing p oduc i i y le els (o labou o land) and consump ion le els (bo h ood and non-
ood). Ul ima ely, he concep o RD seeks o cap u e he esul s o a de elopmen p ocess,
while RUR aims o define he a ibu es ha cha ac e ize a specific a ea.
4. Da a and Me hods
4.1. The S udy A ea and Da a
This s udy was conduc ed in E hiopia, a ede al coun y in Eas A ica. Wi h o e 130
million people (in 2024), i is A ica’s second mos populous na ion, and wi h a 76.84%
u al popula ion, i is he fi h mos u al coun y (T ading Economics, 2024). Known o
i s di e se geog aphy, cul u e, and li elihood (Lie & Mesfin, 2018; OECD/PSI, 2020), E hi-
opia’s economy elies hea ily on ag icul u e, wi h smallholde a me s accoun ing o
90% o p oduc ion. Howe e , he coun y aces significan challenges, pa icula ly land
agmen a ion; 40% o landholde s cul i a e less han 0.5 hec a es each (OECD/PSI, 2020).
This s udy employs na ionally ep esen a i e panel da ase s om he E hiopian So-
cioeconomic Panel Su ey (Panel II: ESPS 4 o 2018/19 and ESPS 5 o 2021/22), which is
p oduced by he E hiopian S a is ical Se ice (ESS, o me ly he Cen al S a is ical
Agency), in pa ne ship wi h he Li ing S anda ds Measu emen S udy (LSMS) o he
Wo ld Bank. The sample uni s a e he same wi hin he panel. Da a we e collec ed om
p edefined u al and u ban cen es, including small, medium, and la ge owns, co e ing
communi y, household, and ag icul u al ac i i ies, wi h specific aen ion o c ops and
Figu e 1. A place-based analy ical amewo k o he p oposed RD p ocess (au ho s’ p esen a ion).
4. Da a and Me hods
4.1. The S udy A ea and Da a
This s udy was conduc ed in E hiopia, a ede al coun y in Eas A ica. Wi h o e
130 million people (in 2024), i is A ica’s second mos populous na ion, and wi h a 76.84%
u al popula ion, i is he i h mos u al coun y (T ading Economics,2024). Known
o i s di e se geog aphy, cul u e, and li elihood (Lie & Mes in,2018;OECD/PSI,2020),
E hiopia’s economy elies hea ily on ag icul u e, wi h smallholde a me s accoun ing o
90% o p oduc ion. Howe e , he coun y aces signi ican challenges, pa icula ly land
agmen a ion; 40% o landholde s cul i a e less han 0.5 hec a es each (OECD/PSI,2020).
This s udy employs na ionally ep esen a i e panel da ase s om he E hiopian So-
cioeconomic Panel Su ey (Panel II: ESPS 4 o 2018/19 and ESPS 5 o 2021/22), which is
p oduced by he E hiopian S a is ical Se ice (ESS, o me ly he Cen al S a is ical Agency),
in pa ne ship wi h he Li ing S anda ds Measu emen S udy (LSMS) o he Wo ld Bank.
The sample uni s a e he same wi hin he panel. Da a we e collec ed om p ede ined u al
and u ban cen es, including small, medium, and la ge owns, co e ing communi y, house-
hold, and ag icul u al ac i i ies, wi h speci ic a en ion o c ops and li es ock p oduc ion.
Enume a ion A eas (EAs), he smalles local-le el sampling uni s, se ed as he s udy’s uni
o analysis (ESS and LSMS,2024). In his s udy, he EA is a speci ically de ined geog aphic
sec ion o a communi y ep esen ing a u al place.
EAs e ec i ely ep esen u al communi ies, allowing o de ailed mic o-le el analysis
and he conside a ion o b oade sub-na ional e ec s. A wo-s age p obabili y sampling
echnique was used o da a collec ion, wi h EAs selec ed in p opo ion o hei size wi hin
egions, ollowed by household selec ion. The ocus on u al de elopmen (RD) es ic s
he analysis o EAs in u al a eas and small owns, excluding medium and la ge u ban
cen es, ensu ing he indings a e ele an o u al con ex s and enhancing his s udy’s
applicabili y (see Table 1).
Economies 2025,13, 61 8 o 29
Table 1. Su ey da ase s wi h hei basic s a is ics.
Desc ip ion Planned Obse a ion Obse a ions in he S udy a
Yea ESPS-4 ESPS-5 ESPS-4 ESPS-5
Numbe o Enume a ion
A eas:
Ru al 297 223 262 b223
U ban 244 215 — —
To al 541 438 262 223
Numbe o Households:
Ru al 3239 2325 2841 2325
U ban 3655 2674 — —
To al 6894 4999 2841 2325
No e:
a
Some EAs in he coun y and he whole Tig ay egion we e no included in he ESPS-5 due o secu i y
easons,
b
Because ollow-up (ESPS-5) da a om Tig ay egion we e no a ailable, Tig ay EAs om ESPS-4 we e
also excluded. Sou ce: au ho s’ p esen a ion.
4.2. Va iables and Measu emen Using P incipal Componen Analysis
The majo a iables employed a e as ollows.
(1)
Dependen a iable: u al de elopmen (RD)
This s udy cons uc ed he dependen a iable, RD, as a composi e a iable by adap -
ing he dimensions and a iables p esen ed by Singh and Shishodia (2023). The s a us o
RD was measu ed by including po en ial ou come a iables, making i a “ esul index” (see
Table 2). Acco ding o Tae-Hwa and Seung-Ryong (2016), a esul index is a componen
alue made o di e en indica o domains o in es iga e he le el o de elopmen . De el-
opmen in E hiopia is also explained by imp o emen in ag icul u al p oduc i i y and he
le el o in es men in public in as uc u e (Dube e al.,2019).
Table 2. Domains and indica o s employed in he index cons uc ion o u al de elopmen .
No. Domains and Indica o s Desc ip ion and Uni o Measu emen
A Ag icul u al p oduc i e e iciency
1 Land (ag icul u al) p oduc i i y C op ha es ed in quin als pe a m size (c op yield)
2 Labou (ag icul u al) p oduc i i y a
To al c op ha es ed in kilog am pe o al man-hou s (adul
equi alen adjus ed) (labou ac i i y pos -plan ing and
pos -ha es ing season)
B Wo k o ce di e si ica ion
3 Pe cen non-ag icul u al wo k o ce To al o - a m employmen pe wo king-age popula ion) ×
100
C Ru al educa ional and heal h in as uc u e
4 P ima y and seconda y schools Numbe
5 P ima y heal h cen es Numbe
6Elec i ied acili ies: p ima y and seconda y
schools Numbe
D Ru al ameni ies
7 Ru al HHs wi h d inking wa e (HH membe s using clean d inking wa e in a
place/popula ion size) ×100 (%)
8 Ru al HHs wi h elec ici y connec ion (HH membe s wi h elec ici y u ili ies in a place/popula ion
size) ×100 (%)
9 Ru al HHs wi h oile s
(HH membe s wi h oile acili ies in a place/popula ion size)
×100 (%)
Economies 2025,13, 61 15 o 29
socioeconomic ac o s like educa ion, employmen , sa e wa e , and sani a ion, Kuzna
discusses he u al–u ban gap. Howe e , his s udy unco e s ema kable dispa i ies wi hin
u al a eas. Gi en egions in E hiopia a e s uc u ed by e hnic c i e ia, inequali ies in line
wi h he e hnici y o Kuzna ’s indings also ha e implica ions o egional dispa i ies.
Economies 2025, 13, x FOR PEER REVIEW 15 o 31
This s udy showed ma ked egional inequali ies in he mean sco e alues o RD,
RUR, and EE ac oss egions (Figu e 4). Ru al O omia, Gambela, and Ha a i ank highes
in de elopmen , and Somali, Amha a, and A a sco e he lowes in descending o de .
While O omia is leading in RD, A a is a he boom. Addi ionally, he Somali, A a , and
Benishangul-Gumuz egions exhibi ed he highes u ali y, wi h Somali ha ing he mos .
In e ms o EE, O omia anks highes , ollowed by Amha a and he o me SNNP (Sou h-
e n Na ions, Na ionali ies, and Peoples) egion. These findings e eal subs an ial pe o -
mance diffe ences among u al a eas11. The egional inequali ies ound ag ee wi h ea lie
findings in E hiopia by A gaw (2017) and Kuzna (2019), who a ibu ed dispa i ies o
diffe ences in oppo uni ies, socioeconomic s a us, and e hnic di e si ies. While A gaw
ocuses on socioeconomic ac o s like educa ion, employmen , sa e wa e , and sani a ion,
Kuzna discusses he u al–u ban gap. Howe e , his s udy unco e s ema kable dispa -
i ies wi hin u al a eas. Gi en egions in E hiopia a e s uc u ed by e hnic c i e ia, ine-
quali ies in line wi h he e hnici y o Kuzna ’s findings also ha e implica ions o egional
dispa i ies.
Figu e 3. Pae ns o RD, RUR, and EE o e ime. Au ho s’ ep esen a ion.
Figu e 3. Pa e ns o RD, RUR, and EE o e ime. Au ho s’ ep esen a ion.
Economies 2025, 13, x FOR PEER REVIEW 16 o 31
Figu e 4. The ex en o RD, RUR, and EE in he egional s a es o E hiopia. Au ho s’ ep esen a ion.
Such dispa i ies in de elopmen le els call o u gen , place-based in e en ions
(Mo ei, 2022). Remo e u al a eas a e s uggling wi h popula ion decline in Eu ope
(Cas illo e al., 2023) and economic s agna ion in China (Li e al., 2015). The egional di -
e ences make p i a e and public se ice deli e y difficul , wo sening egional inequali-
ies (Cas illo e al., 2023). Place-based s a egies a e gene ally mo e ele an o specific
challenges and ma ginaliza ion, as hey u ilize local social dynamics and ne wo ks (Win-
e on e al., 2014). Ini ia i es o PBRD a e gaining ecogni ion in Eu ope (Sánchez-
Zamo a & Galla do-Cobos, 2020), Aus alia (Gea ches e al., 2023), he USA (Pa ke e al.,
2022) and A ica (A DB/OECD, 2015), wi h E hiopia ale ed o adop simila solu ions
(OECD/PSI, 2020, pp. 22, 104). The PBRD app oach aligns wi h E hiopia’s cons i u ional
commi men o ede a ion and decen aliza ion ha aims o inclusi e g ow h. Also, by
add essing local needs, he app oach can enhance he low en ep eneu ship and
s eng hen he weak u al–u ban linkages in he coun y.
5.3. Findings om he P incipal Componen Analysis
A e a imax o a ion, ou componen s (Comp1, Comp2, Comp3, and Comp4) we e
ex ac ed in he PCA o RD wi h p opo ions o 0.2188, 0.1490, 0.1406, and 0.1371 in he
fi s su ey yea 2018/19; and 0.1939, 0.1860, 0.1557 and 0.1124 in he second su ey yea
2021/22 (Table 4). The fi s and he o e all componen s ha e join ly explained a high pe -
cen age o he o al a ia ion in he o iginal da ase o bo h su ey yea s. While he fi s
componen alue was 21.9% and 19.4%, o ESPS-4 and ESPS-5, espec i ely, he o e all
a ia ion in he da a explained by he ou componen s was 64.55% and 64.8%, espec-
i ely, which implies ha he a ia ion explained is adequa e o de e mine he s uc u e
o RD (Finch, 2013), and he model, hus, is accep able (Hai e al., 2019).
Table 4. Ini ial componen ex ac ion o u al de elopmen , S a is ics o Va imax Ro a ed Compo-
nen s.
Figu e 4. The ex en o RD, RUR, and EE in he egional s a es o E hiopia. Au ho s’ ep esen a ion.
Such dispa i ies in de elopmen le els call o u gen , place-based in e en ions
(Mo e i,2022). Remo e u al a eas a e s uggling wi h popula ion decline in Eu ope
(Cas illo e al.,2023) and economic s agna ion in China (Li e al.,2015). The egional di e -
Economies 2025,13, 61 16 o 29
ences make p i a e and public se ice deli e y di icul , wo sening egional inequali ies
(Cas illo e al.,2023). Place-based s a egies a e gene ally mo e ele an o speci ic chal-
lenges and ma ginaliza ion, as hey u ilize local social dynamics and ne wo ks (Win e on
e al.,2014). Ini ia i es o PBRD a e gaining ecogni ion in Eu ope (Sánchez-Zamo a &
Galla do-Cobos,2020), Aus alia (Gea ches e al.,2023), he USA (Pa ke e al.,2022) and
A ica (A DB/OECD,2015), wi h E hiopia ale ed o adop simila solu ions (OECD/PSI,
2020, pp. 22, 104). The PBRD app oach aligns wi h E hiopia’s cons i u ional commi men
o ede a ion and decen aliza ion ha aims o inclusi e g ow h. Also, by add essing
local needs, he app oach can enhance he low en ep eneu ship and s eng hen he weak
u al–u ban linkages in he coun y.
5.3. Findings om he P incipal Componen Analysis
A e a imax o a ion, ou componen s (Comp1, Comp2, Comp3, and Comp4) we e
ex ac ed in he PCA o RD wi h p opo ions o 0.2188, 0.1490, 0.1406, and 0.1371 in he
i s su ey yea 2018/19; and 0.1939, 0.1860, 0.1557 and 0.1124 in he second su ey yea
2021/22 (Table 4). The i s and he o e all componen s ha e join ly explained a high
pe cen age o he o al a ia ion in he o iginal da ase o bo h su ey yea s. While he i s
componen alue was 21.9% and 19.4%, o ESPS-4 and ESPS-5, espec i ely, he o e all
a ia ion in he da a explained by he ou componen s was 64.55% and 64.8%, espec i ely,
which implies ha he a ia ion explained is adequa e o de e mine he s uc u e o RD
(Finch,2013), and he model, hus, is accep able (Hai e al.,2019).
Table 4. Ini ial componen ex ac ion o u al de elopmen , S a is ics o Va imax Ro-
a ed Componen s.
Su ey Da a: ESPS-4 ESPS-5
Componen
Va iance P opo ion
Cumula i e Componen
Va iance P opo ion
Cumula i e
Comp1 1.9695 0.2188 0.2188 Comp1 2.32653 0.1939 0.1939
Comp2 1.3409 0.1490 0.3678 Comp2 2.23201 0.1860 0.3799
Comp3 1.2654 0.1406 0.5084 Comp3 1.8688 0.1557 0.5356
Comp4 1.2337 0.1371 0.6455 Comp4 1.3483 0.1124 0.6480
Comp5 0.89720 0.0997 0.7452 Comp5 0.90972 0.0758 0.7238
Comp6 0.69803 0.0776 0.8228 Comp6 0.70951 0.0591 0.7829
Comp7 0.63911 0.0710 0.8938 Comp7 0.65126 0.0543 0.8372
Comp8 0.54929 0.0610 0.9548 Comp8 0.57741 0.0481 0.8853
Comp9 0.40681 0.0452 1.0000 Comp9 0.48477 0.0404 0.9257
Comp10 0.44418 0.0370 0.9627
Comp11 0.31252 0.0260 0.9888
Comp12 0.13495 0.0112 1.0000
ESPS-4 ESPS-5
Numbe o obse a ions 255 218
Numbe o componen s 9 12
Kaise –Meye –Olkin measu e o sampling adequacy (KMO) (o e all): 0.6278 0.6454
De e minan o he co ela ion ma ix 0.338 0.026
Ba le es o sphe ici y a
Chi-squa e = 271.204 775.493
p- alue = 0.000 0.000
No e: aH0: Va iables a e no in e co ela ed.
The inal bes model is ob ained om se e al ials and e o s, which ended in 9 and
12 a iables in he ESPS-4 and ESPS-5, espec i ely. See Table 5 o componen loadings
(eigen ec o s o he eigen alues) which in o m he s eng h o associa ion be ween he
loading a iables and he ex ac ed p incipal componen s (PCs). I is usually ecommended
Economies 2025,13, 61 17 o 29
o include componen loading alues g ea e han 0.3 (WinC oss,2022). The componen s
loading on he iden i ied a iables we e signi ican and hus, hey explain RD (Finch,2013).
This loading allows us o ob ain an RD sco e o each communi y in ol ed in he s udy.
Table 5. Componen loadings o u al de elopmen ( o bo h su ey yea s).
1Va iable ESPS-4 ESPS-5
Comp1 Comp2 Comp3 Comp4 Comp1 Comp2 Comp3 Comp4
2Ru al HH wi h d inking
wa e (%) 0.5406 0.446
3 HH wi h oile acili ies (%) 0.5998 0.591
4 HH wi h elec ic u ili ies (%) 0.5861
5E ec i e li e acy a e
( ead/w i e: 0.636 0.696
6 School en ollmen (mean) 0.642 0.611
7P ima y and seconda y
schools (numbe ) 0.366 0.564 0.463
8Non-ag icul u al wo k o ce
(%) 0.737 0.522
9Expendi u e on non- ood
i ems 0.692 0.558
10 Annual expendi u e on
educa ion 0.720 0.547
11 Elec i ied acili ies, schools;
numbe 0.557
12 Pe cen u al HHs wi h
mode n household ameni ies
0.611
13 Annual expendi u e on
u ili ies 0.515
14 Accessibili y o comme cial
banks −0.406
Componen in e p e a ion: ESPS-4 ESPS-5
Comp1:
En i onmen al hygiene and sani a ion En i onmen al hygiene and sani a ion
Comp2: Li e acy Non- ood consump ion
Comp3: Non- ood consump ion In as uc u e
Comp4: (Public) In as uc u e Li e acy
Signi ican loading: >0.3; emp y cells: loadings less han 0.3. Sou ce: au ho s’ compu a ion.
Based on he esul s in Table 5, he e was signi ican posi i e loading on Comp 1 o
a iables such as pe cen u al households wi h d inking wa e , oile acili ies, and elec ic
u ili ies. F om he na u e o he a iables, i is e iden ha he componen is “En i onmen al
Hygiene and Sani a ion”. Componen 2 has a signi ican loading on e ec i e li e acy a e
(household membe s aged o e o equal o 5 and who can ead o w i e), mean school
en ollmen , and he o al numbe o p ima y and seconda y schools in he communi y.
This componen is hus in e p e ed as “Li e acy”. Comp3 has a signi ican ly high loading
on annual expendi u e on non- ood i ems and educa ion. This componen is labelled
as “Non- ood consump ion”. In he ou h componen o he i s su ey, he impo an
loadings we e on he numbe o p ima y and seconda y schools and he pe cen age o
non-ag icul u al wo k o ce in he communi ies. This is hus “non- a m employmen ”. In
all ou componen s, he highe he sco e o he componen loadings, he s onge he RD
would be and ice e sa, i.e., a di ec ela ionship.
In ESPS-5 (2021/22), Comp1 has a high and signi ican loading on he pe cen o u al
households wi h d inking wa e , oile acili ies, and o al mode n household ameni ies.
This componen is simila o he indings in he p e ious su ey. I is in e p e ed as
Economies 2025,13, 61 18 o 29
“En i onmen al Hygiene and Sani a ion”. Componen 2 has a high posi i e loading on
annual expendi u e on non- ood i ems, educa ion, and u ili ies. Hence, i is “Non- ood
consump ion”. Componen 3 has high posi i e loading on he o al numbe o p ima y
and seconda y schools in he communi y, he pe cen age o non-ag icul u al wo k o ce, he
o al numbe o elec i ied acili ies in he communi y (p ima y and seconda y schools), and
nega i e loading on accessibili y o comme cial banks (dis ance in km). This is in e p e ed as
public in as uc u e. Componen 4 has a high and posi i e loading on e ec i e li e acy a e
and mean school en ollmen . All he componen s in he ESPS-5 ha e a di ec ela ionship
wi h RD.
F om he abo e componen analysis, RD in E hiopia as an ou come can be explained
mainly in e ms o en i onmen al sani a ion and ameni y le els, li e acy a e and school
en ollmen , non- ood consump ion, and public expendi u e. Ou pu s o he wo key
eg esso s can be ound in he Supplemen a y Ma e ials Tables S1 and S2 o RUR and
Tables S3 and S4 o EE.
5.4. Key Findings Using FE IV-2SLS Reg ession
This s udy cons uc ed indices o he key mul idimensional dependen a iable o
u al de elopmen (RD) and i s key endogenous eg esso s: u ali y (RUR) and he u al
en ep eneu ial ecosys em (EE). D awing on heo e ical insigh s and empi ical esea ch,
he s udy es ed he hypo hesis ha hese mul idimensional, place-speci ic ea u es a e
i al de e minan s o RD in E hiopia. The panel ixed e ec s me hod enhances ou analysis,
while he 2SLS me hod signi ican ly educes po en ial endogenei y issues. Residual plo s
indica e no se ious conce ns ega ding non-linea i y, ou lie s, o he e oskedas ici y, wi h
a iance a ound ze o appea ing uni o mly dis ibu ed.
The ou pu om he model wi hou and wi h con ol a iables is p esen ed in
Table 6
.
The inal esul s e e o he p edic ion wi h he FE IV-2SLS model including all he con ol
a iables wi h clus e obus s anda d e o
12
(see columns C and D). Gi en he la ge
clus e s in he s udy, he clus e s’ obus s anda d e o s emain consis en , e en wi h he -
e oskedas ici y o au oco ela ion (Came on & Mille ,2015)
13
. The es ima ion is alida ed
by weak ins umen and o e -iden i ica ion es s, con i ming he ele ance and alidi y o
he ins umen s (see he ou pu a he bo om o Table 6). The i s -s age eg ession demon-
s a es a s a is ically signi ican ela ionship be ween he ins umen s and he endogenous
eg esso s (see Table 6, bo om). Ins umen alida ion is u he suppo ed by obus
empi ical e idence (see Sec ion 5.7). I was ound ha bo h he RUR and EE s a is ically
signi ican ly a ec RD (see de ails in he sec ions below). Mo eo e , key a iables ela ed
o echnology and capi al, such as he pe cen age o households using imp o ed c op seeds
(p< 0.1) and he numbe o a m ools pe hec a e (p< 0.01), a e also s a is ically signi ican
explaining he ole played by ag icul u e in RD (Table 6).
Table 6. FE IV-2SLS es ima ion ou pu .
AaB C Db
Dependen Va iable:
Ru al de elopmen (RD)
FE IV-2SLS Coe
(wi hou con ol
a s and
non- obus s d. e .)
(-x i eg-
command)
FE IV-2SLS Coe
(con ol a iable
and non- obus s d.
e .)
(-x i eg-
command)
FE IV-2SLS Coe
(con ol a s and
clus e - obus sd
e ) c
(-x i eg-
command)
FE IV-2SLS Coe
(wi h con ol
a iables and
clus e - obus sd
e ( inal model)
(-x i eg2-
command)
Economies 2025,13, 61 19 o 29
Table 6. Con .
AaB C Db
Endogenous explana o y (ins umen ed) a iables
Ru ali y (RUR) −0.261 (0.117) ** −0.270 (0.1266) ** −0.270 (0.1121) ** −0.270 (0.1106) **
Ru al en ep eneu ial
ecosys em (EE) 0.294 (0.165) * 0.456 (0.1724) *** 0.456 (0.1748) ** 0.456 (0.1726) ***
Exogenous con ol a iables (included ins umen s)
Pe cen HH u ilizing
imp o ed c op seed
(ag icul u al echnology)
— 0.208 (0.0806) ** 0.208 (0.1225) * 0.208 (0.1210) *
Cos o e ilize (global
p ice) — 0.000035 (0.00001) −0.00035 (0.00082) −0.0003 (0.00008)
TLU o al ( esou ces) — 0.0007 (0.0008) 0.0007 (0.0006) 0.0007 (0.0006)
Fa m ools pe a m
holding ( a m capi al) —−0.001 (0.004) ** −0.001 (0.003) *** −0.001 (0.003) ***
Land ce i ica e ( u al
ins i u ion) — 0.096 (0.0638) 0.096 (0.0701) 0.096 (0.0692)
Cons an 0.227 *** 0.123 (0.0549) ** 0.123 (0.0630) * —
No. o Obs 390 337 337 252
P ob > F 0.080 0.0029 0.0000 0.0085
Tes o o e iden i ying es ic ions (using he -x o e id- command) (Ou igh ) d
Sa gan–Hansen s a is ic 4.406 2.274 2.462 2.462 (J) e
Chi-sq(2) p- alue = 0.1104 0.3208 0.2921 0.2921
Fi s -s age eg ession esul s
Ins umen al Va iables
(Excluded Ins umen s)
Dependen
Va iable:
Ru ali y (RUR)
Dependen Va iable: En ep eneu ial Ecosys em (EE)
Coe icien ( obus s anda d e o s)
Pe cen bushland 0.015 (0.0045) *** −0.03 (0.0047)
Pe cen o es land 0.07 (0.0014) *** 0.026 (0.0016)
Licence paymen
(non- a m) −0.01 (0.003) *** 0.015 (0.0029) ***
U ban dis ance 0.01 (0.002) *** −0.017 (0.029) ***
Tes s o he inal model epo ed, FE IV-2SLS.
Unde iden i ica ion es (Kleibe gen–Paap k LM s a is ic) :9.766
Chi-sq(3) p- al = 0.0207
Weak iden i ica ion es (Kleibe gen–Paap k Wald F s a is ic): g9.904
S ock–Yogo weak ID es c i ical alues: 5% maximal IV ela i e bias 11.04
10% maximal IV ela i e bias 7.56
20% maximal IV ela i e bias 5.57
30% maximal IV ela i e bias 4.73
10% maximal IV size 16.87
15% maximal IV size 9.93
20% maximal IV size 7.54
25% maximal IV size 6.28
Economies 2025,13, 61 20 o 29
Table 6. Con .
AaB C Db
NB: C i ical alues a e o C agg–Donald F s a is ic and i.i.d. e o s.
Endogenei y es o endogenous eg esso s: 6.091
Chi-sq(2) p- al h=0.046
No e:
a
A model wi hou he con ol a iables is good o examining he es ima es wi h espec o he subsequen
ull model. Howe e , one canno s op he e as i su e s om omi ed a iables bias ha may mislead in e p e a ion,
b
Final coe icien es ima es in column D a e epo ed using he use ’s command o -x i eg2- (Scha e ,2010).
Column C is he ou pu using S a a’s o icial command (-x i eg-). No di e ence in he coe icien es ima es
excep o he absence o a cons an in he -x i eg2-. The -x i eg2- command has he ad an age o he ou igh
p esen a ion o he a ious pos -es ima ion checks,
c
Clus e - obus s anda d e o s: obus o he e oscedas ici y
and co ela ion o e o s wi hin he indica ed panel iden i ica ion (id) clus e s and se ial co ela ion,
d
The -x i eg2-
use command p esen s he es esul ou igh , i.e., wi hou he -x o e id- command,
e
Hansen J s a is ics; Ho:
he ins umen s a e alid (i.e., unco ela ed wi h he e o e m),
Ho: The ma ix o educed o m coe icien s
has ank = K1-1 (unde iden i ied),
g
Ho: The equa ion is weakly iden i ied,
h
Ho: The speci ied endogenous
eg esso s a e exogenous (using endog op ion in -x i eg2- command). *** p< 0.01; ** p< 0.05; * p< 0.1. Sou ce:
au ho s’ compu a ion.
5.5. E ec o Ru ali y on Ru al De elopmen
This sec ion p esen s he es esul o he hypo hesis ha u ali y (RUR) has a s a-
is ically signi ican e ec on RD. I was ound ha a 1-uni dec ease in RUR esul s in a
s a is ically signi ican inc ease o 0.27 uni s in RD (p< 0.05), holding all o he a iables
cons an (Table 6). The esul con i med he in e se ela ionship exhibi ed be ween RD
and RUR in Sec ion 5.2. The esul ag ees wi h indings om Eu ope (Cas illo e al.,2023)
and China (Li e al.,2015) whe e a eas wi h highe u ali y—such as emo e u al a eas—
expe ience lowe le els o economic de elopmen . By iewing RUR as a mul idimensional,
place-speci ic ac o a ec ing de elopmen , his s udy ag ees wi h Gol zsche (2022) and
Chigbu (2013). Al hough hese s udies app oached he issue di e en ly, hey highligh he
need o ecognize u ali y (o u al ypology) in de elopmen discussions. The e idence
om China u he con i ms he link be ween u ali y and RD (Li e al.,2015). Addi ionally,
he indings go along wi h s udies by K useman e al. (2006) in No he n E hiopia and
Hausmann e al. (2021) in Mexico.
Dis inguishing he di e ences and simila i ies wi h he ea lie s udies helps unde -
s and place-speci ic ac o s. K useman e al. (2006) iden i ied adi ional ac o s like ma ke
access, ag icul u al po en ial, and popula ion densi y o a ec p oduc ion sys ems and
de elopmen in No he n E hiopia. They emphasize ha ealizing adi ional place-speci ic
ac o s is impo an in unde s anding he e ogenei y and policy a ge ing. Howe e , apply-
ing se e al adi ional indica o s may con ine policy a ge ing. The p esen s udy, ins ead,
p esen s an inno a i e app oach by in eg a ing he a ious indi idual place-speci ic indica-
o s in o a single index which ag ees wi h he indings by Hausmann e al. (2021). The la e
is a ecen s udy by Hausmann e al. (2021) in Chiapas, Mexico’s poo es egion. I unde -
sco es he impo ance o mul idimensional, place-speci ic ea u es in de elopmen . Thei
indings show ha he Economic Complexi y Index
14
, he mul idimensional place-speci ic
ac o hey examined, signi ican ly a ec ed wage and employmen g ow h a he municipal
le el. The inclusion in o he model o his a iable inc eased he explana o y powe o
he model by a g ea e han ha o indi idual ac o s like educa ion and adi ional
place-speci ic ac o s such as oad in as uc u e and c edi ma ke s.
The use o such mul idimensional ac o s is mo e manageable han ying o add ess
he many indi idual place-speci ic ac o s. Thei e ec s a e no ably s onge in u al a eas
as hese economies ely hea ily on na u al esou ces and clima e (Gol zsche,2022). Places
a e speci ic geog aphic loca ions. The inding on RUR would imply he e ec o geog aphy.
Geog aphy shapes economic de elopmen h ough ac o s like ag icul u al p oduc i i y,
esou ce a ailabili y, and clima e (Boldeanu & Cons an inescu,2015). Howe e , his s udy
Economies 2025,13, 61 21 o 29
goes beyond adi ional unde s anding o p o ide compelling e idence ha a single,
mul i ace ed cha ac e is ic—such as u ali y—holds signi ican explana o y powe . This
e eals he deep connec ions be ween spa ial en i ies, human dynamics, and ins i u ions.
Ru ali y en ails mo e han a simple geog aphic en i y. Beyond his, i is ma ked by unique
ma e ial and social capi al shaped h ough in e ac ions be ween people and a place o ming
dynamic, place-speci ic ea u es. Such unde s anding compels acknowledging u al a eas
as ib an cen es o economic ac i i y and social ad ancemen in a way ha enhances bo h
u al de elopmen and in si u u baniza ion.
5.6. E ec o En ep eneu ial Ecosys em on Ru al De elopmen
This sec ion p esen s he es esul s o he hypo hesis ha he en ep eneu ial ecosys-
em (EE) has a s a is ically signi ican e ec on RD. I is impo an o no e ha he EE
analysis is limi ed by he a ailabili y o da a o local-le el indica o s o ce ain en e p ise
a iables. Fo ins ance, in o ma ion is usually una ailable on high- ech i ms ha help
gain deepe insigh s in o EE. Da a inadequacy in measu ing and analyzing en ep eneu ial
ecosys ems is no uncommon (see Li e al.,2023;Spigel e al.,2020). While na ional-le el
in o ma ion be e exis s, da a o local and egional analyses a e o en inadequa e (S am,
2017). The issue is pa icula ly e iden in Sub-Saha an A ican coun ies like E hiopia
whe e mos i ms ace echnological cons ain s (Ci e a e al.,2023). The s udy, howe e ,
e ec i ely u ilized da a om ex a-local sou ces on la ge-scale en e p ises. App oaching
such local da a limi a ions is also no uncommon (see Mainle el Consul ing,2024). This is
because en ep eneu ial ecosys ems a e usually o med wi h he in e ac ion be ween local
and ex a-local s akeholde s’ ac ions whe e he e a e chances o la ge and high- ech i ms
o in e ac (Li e al.,2023). Thus, his s udy ca e ully cons uc ed p oxy indica o s o such
missing a iables (see Table S2). Acco dingly, he esul s p esen ed in Figu e 3and Table 6
show a di ec co ela ion be ween RD and EE, and Figu e 4illus a es egional inequali ies
in EE. A uni inc ease in EE was associa ed wi h a s a is ically signi ican 0.46-uni inc ease
in he RD (p< 0.01), holding all o he a iables cons an . En ep eneu ship is o en seen
as an indi idual endea ou , bu i is deeply a ec ed by he su ounding spa ial con ex .
The connec ions be ween en ep eneu s, en e p ises, and s akeholde s—including he
ma ke and enabling en i onmen —a e c ucial (Ko sgaa d & Tan ig,2015). A u al a ea is
beyond a physical loca ion shaped by dynamic socio-cul u al in e ac ions be ween people
and place. This unde s anding acknowledges he iew o “ u al” as an essen ial “spa ial
dimension” in he de elopmen p ocess in gene al and en ep eneu ship in pa icula .
En ep eneu ship is a key d i e o inno a ion and economic g ow h. Deepe han e e , RD
has a s ong and s a egic associa ion wi h en ep eneu ship (Kulka ni & Na khede,2016).
Wi h he cu en socioeconomic landscape ea u ing high u al e ili y and signi ican
u al- o-u ban mig a ion in he coun y, u ban a eas ace high unemploymen a es, la gely
due o his mig a ion and limi ed a mland in u al egions (OECD/PSI,2020). The une en
en ep eneu ial le els ac oss he coun y (see Figu e 4) exace ba e hese challenges. In
such cases, a place-based ib an EE has emendous po en ial o job c ea ion (Kulka ni &
Na khede,2016), po e y alle ia ion (Naminse e al.,2019), and ans o ma i e solu ions
o employmen and esou ce managemen (P iya & Mohanasunda i,2024).
5.7. Pos -Es ima ion Checks
5.7.1. Weak Ins umen s and O e iden i ica ion Tes s
Two ins umen s o each endogenous eg esso we e examined o ensu e obus
indings. Pe cen ages o land co e ed by bush and o es in a communi y a e pi o al o
assessing u ali y (RUR). Addi ionally, annual licencing ees paid by non- a m en e p ises,
which se e as a p oxy o he egula o y amewo k, and he dis ance o he nea es majo
Economies 2025,13, 61 22 o 29
u ban cen e in o m he en ep eneu ial ecosys em (EE). Ru ali y can be de ined by p ima y
economic ac i i ies like a ming and o es y (FAO,2018), making land co e an essen ial
ea u e. The p opo ions o bush and o es land co e ages no only cha ac e ize he u al
a eas bu a e s ongly co ela ed wi h u ali y. Likewise, he de elopmen o u al egions
is in luenced u he by local ins i u ional ac o s (WBG,2016) and ex a-local ac o s such
as nea by u ban agglome a ions (Ca aneo e al.,2022) o dis ance om he nea es u ban
cen es (La esson,2018). Non- a m licencing ees and dis ance o u ban cen es p omo e he
en ep eneu ial ecosys em by impac ing business s a -up oppo uni ies (WBG,2016). The
i s -s age eg ession ou pu (2SLS) con i ms he ele ance o hese ins umen s, showing a
signi ican s a is ical ela ionship wi h hei espec i e endogenous eg esso s (
p
< 0.0000),
as p esen ed in Table 615.
5.7.2. Robus ness Check
In alida ing he es ima es, he endogenei y es equi es mo e han s a is ical checks
o obus ness (Cla ke & Ma a,2018). The examina ions ac oss a ious scena ios showed
consis en and obus coe icien es ima es, suppo ing hei eliabili y. Fi s , hey we e
checked using a di e en sample, emo ing he pas o al ep esen a i es. The pas o al
sys em, co e ing 61% o he coun y’s e i o y, e lec s dis inc li elihoods and cul u es
(Gebeye,2016). F om pas o al samples ( egions and zones), 127 Enume a ion A eas (EAs)
we e excluded, ocusing solely on he highland sample. The esul s showed consis en
es ima es o he key coe icien s (see Table S7). Also, he con ol a iables we e exam-
ined in a ious ways o assess hei e ec on RD. By excluding all con ol a iables, and
ein oducing hem one a a ime—an app oach suppo ed in he li e a u e (see Ka ka,
2024)— he key coe icien s emained consis en in sign and magni ude (see Table 6, column
A). Mo e impo an ly, he plausible exogenei y es was assessed based on Conley e al.
(2012) o challenge he exclusion es ic ion o he ins umen al a iable assump ion. Such
a p ocedu e is no uncommon p ac ice (Nguyen-Phung & Le,2024;Zheng e al.,2023).
The esul s indica e ha when ins umen s a e no s ic ly exogenous, he socio-spa ial
de e minan s s ill signi ican ly a ec RD (Figu e 5).
Economies 2025, 13, x FOR PEER REVIEW 23 o 31
Figu e 5. The 95% confidence in e als o he effec o u ali y on u al de elopmen using he local-
o-ze o (LTZ) me hod o he plausible exogenei y assump ion. Sou ce: au ho s’ ep esen a ion.
6. Conclusions and Implica ions
This s udy aims o unde s and he complex socio-spa ial na u e o RD and i s ela-
ionship wi h key mul idimensional, place-specific de e minan s, namely u ali y and he
en ep eneu ial ecosys em. Using a place-based amewo k, i s uc u ed he socio-spa ial
ac o s o u al locali ies o unde sco e he explana o y powe o socio-spa ial de e mi-
nan s. I applied ins umen al a iables es ima ion o deal wi h he endogenei y p oblem
associa ed wi h hese place-specific de e minan s. The es ima ion esul s using he fixed
effec s model wi h ins umen al a iables showed s a is ically significan effec s o bo h
u ali y (nega i e) and he en ep eneu ial ecosys em (posi i e) on RD. Ru ali y is a dis-
inc socioeconomic cha ac e is ic o a place. The p esen finding depic s ha u al a eas
a e beyond fi s -na u e geog aphy and hey ha e unequal de elopmen le els. Thus, hei
embedded unique ea u es a e composed o se e al local ac o s such as geog aphy, ins i-
u ions, and go e nance, which in e ac p ima ily wi h local and ex a-local ea u es, o
shape hei unique de elopmen , and in u n, d i e each in a diffe en de elopmen a-
jec o y. Likewise, a unique en ep eneu ial ecosys em eme ges wi hin he u al con ex .
While en ep eneu ial ac i i y en ails indi idual o g oup effo s, i is shaped by he u al
con ex ha leads o a unique en ep eneu ial ecosys em in a u al locali y o egion. This
ecosys em nu u es en e p ise oppo uni ies and add esses u al needs in a way ha sig-
nifican ly impac s RD.
This s udy also ound a s a is ically significan effec o a m capi al and echnology
on RD. This implies he impo an ole ag icul u al de elopmen plays in success ul RD.
Howe e , he s udy goes beyond sec o al ou look o socio-spa ial conce ns. I unde pins
a conside able significance o he inhe en and blended o ma ion o na u al and human
ac o s in a u al place. The finding hus emphasizes he undamen al influence o he
second-na u e geog aphy (human and ins i u ional effec ) on he fi s na u e, suppo ing
a need o add ess he mul idimensional place-specific ac o s as a unique explana o y
powe . E hiopia is a p edominan ly u al and he e ogeneous coun y acing p essing chal-
0
Es ima ed β
Figu e 5. The 95% con idence in e als o he e ec o u ali y on u al de elopmen using he
local- o-ze o (LTZ) me hod o he plausible exogenei y assump ion. Sou ce: au ho s’ ep esen a ion.
Economies 2025,13, 61 23 o 29
6. Conclusions and Implica ions
This s udy aims o unde s and he complex socio-spa ial na u e o RD and i s ela-
ionship wi h key mul idimensional, place-speci ic de e minan s, namely u ali y and he
en ep eneu ial ecosys em. Using a place-based amewo k, i s uc u ed he socio-spa ial
ac o s o u al locali ies o unde sco e he explana o y powe o socio-spa ial de e mi-
nan s. I applied ins umen al a iables es ima ion o deal wi h he endogenei y p oblem
associa ed wi h hese place-speci ic de e minan s. The es ima ion esul s using he ixed
e ec s model wi h ins umen al a iables showed s a is ically signi ican e ec s o bo h
u ali y (nega i e) and he en ep eneu ial ecosys em (posi i e) on RD. Ru ali y is a dis inc
socioeconomic cha ac e is ic o a place. The p esen inding depic s ha u al a eas a e
beyond i s -na u e geog aphy and hey ha e unequal de elopmen le els. Thus, hei
embedded unique ea u es a e composed o se e al local ac o s such as geog aphy, in-
s i u ions, and go e nance, which in e ac p ima ily wi h local and ex a-local ea u es,
o shape hei unique de elopmen , and in u n, d i e each in a di e en de elopmen
ajec o y. Likewise, a unique en ep eneu ial ecosys em eme ges wi hin he u al con ex .
While en ep eneu ial ac i i y en ails indi idual o g oup e o s, i is shaped by he u-
al con ex ha leads o a unique en ep eneu ial ecosys em in a u al locali y o egion.
This ecosys em nu u es en e p ise oppo uni ies and add esses u al needs in a way ha
signi ican ly impac s RD.
This s udy also ound a s a is ically signi ican e ec o a m capi al and echnology
on RD. This implies he impo an ole ag icul u al de elopmen plays in success ul RD.
Howe e , he s udy goes beyond sec o al ou look o socio-spa ial conce ns. I unde pins
a conside able signi icance o he inhe en and blended o ma ion o na u al and human
ac o s in a u al place. The inding hus emphasizes he undamen al in luence o he
second-na u e geog aphy (human and ins i u ional e ec ) on he i s na u e, suppo ing
a need o add ess he mul idimensional place-speci ic ac o s as a unique explana o y
powe . E hiopia is a p edominan ly u al and he e ogeneous coun y acing p essing
challenges such as egional inequali ies. Ye , i has implemen ed a policy biased owa ds
homogenei y and u ban way, i.e., designing RD in a way ha p omo es an “u banely”
app oach. This u ban dominance isks high u al- o-u ban mig a ion ha ou paces job
c ea ion in u ban cen es, esul ing in u baniza ion wi hou g ow h. Thus, wi h hese socio-
spa ially in e wo en u al challenges, he indings in his s udy boil down o adop ing
a place-based u al app oach o de elopmen (PBRD) ha necessi a es a ge ing a place,
and he unique place ea u es he eo . I also helps u he le e age ex a-local gains o
imp o e he wel a e o he people. A PBRD p omo es an app oach ha acknowledges
he u al li es yle, maximizing he bene i o a sense o place. This consequen ly enhances
de elopmen , educing he cos o mig a ion and allowing a simul aneous nu u ing o
a m and non- a m ac i i ies. Such an app oach aligns wi h he Sus ainable De elopmen
Goals, which p omo e RD wi hou subsuming i wi hin u ban de elopmen , leading o
inclusi e and sus ainable g ow h. Neglec ing place-speci ic ac o s in policymaking could
hus lead o ailu e in sus ainable de elopmen .
In line wi h he PBRD, his s udy con ibu es o he heo y and p ac ices o RD in wo
essen ial ways. Fi s , i p o ides a i id depic ion o he complex socio-spa ial con ex o
a eas o policy a ge ing and implemen a ion. The mul idimensional ea u es p esen ed
a e con ex -applicable so ha RD policymake s and p ac i ione s wo king in go e nmen al
and non-go e nmen al sec o s can ca ch hold o and cou ageously u ilize hem in designing
and implemen ing a a ge ed PBRD. Second, he analy ical amewo k enables he e ec i e
iden i ica ion and measu emen o mul idimensional place-speci ic ea u es in a complex
socio-spa ial in e play. This g ea ly helps he g owing concep ualiza ion o a place-based
app oach in e ining he ole o place in local and egional de elopmen . Adop ing a
Economies 2025,13, 61 24 o 29
PBRD and in eg a ing i in o he b oade na ional and sub-na ional s a egies hus will
help add ess he unique u al challenges and enhance policy e ec i eness in E hiopia and
beyond. The s udy is cons ained by ob aining local-le el indica o s o some a iables,
pa icula ly o en e p ise analysis such as da a on high- ech i ms. The la e is he na u e
o he i ms in he coun y, whe e he echnology is limi ed. Howe e , he s udy included
da a om he ex a-local le els on la ge-scale en e p ises by cons uc ing possible p oxies,
hus enabling analysis o he complex suppo se ices and business in e ac ions.
Supplemen a y Ma e ials: The ollowing suppo ing in o ma ion can be downloaded a h ps://
www.mdpi.com/a icle/10.3390/economies13030061/s1. Table S1. Domains and indica o s em-
ployed in he index cons uc ion o u ali y; Table S2. Domains and indica o s employed in con-
s uc ing he EE index (en ep eneu ial ecosys em); Table S3. Ini ial componen ex ac ion o u-
ali y;
Table S4
. Componen loadings o u ali y; Table S5. Ini ial componen ex ac ion o he
en ep eneu ial ecosys em; Table S6. Componen loadings o he en ep eneu ial ecosys em; Table S7.
Sample size and he es ima ion ou pu o an al e na i e sample.
Au ho Con ibu ions: Concep ualiza ion, M.T.W., S.A. and B.K.S.; Me hodology, M.T.W.; Da a Cu a-
ion, M.T.W.; Fo mal analysis, M.T.W.; So wa e, M.T.W.; W i ing—o iginal d a , M.T.W.; W i ing—
e iew & edi ing, M.T.W., S.A. and B.K.S.; Visualiza ion, M.T.W.; Supe ision, S.A. and B.K.S. All
au ho s ha e ead and ag eed o he published e sion o he manusc ip .
Funding: This esea ch ecei ed no ex e nal unding.
In o med Consen S a emen : No applicable.
Da a A ailabili y S a emen : Da a a ailable in a publicly accessible eposi o y.
Acknowledgmen s: We would like o exp ess ou g a i ude o he Wo ld Bank and CSA o E hiopia
(now known as ESS) o pe mi ing us o use he LSMS da a. We also app ecia e aluable commen s
and sugges ions om academic edi o , h ee anonymous e iewe s and ou academic colleagues
du ing he d a ing p ocess. The esea ch pape e lec s he au ho s’ p o essional judgmen s.
Con lic s o In e es : The au ho s decla e no con lic o in e es .
No es
1Egyp , E hiopia, Ghana, Kenya, Mali, Mozambique, Nige ia, Tanzania, Uganda and Zambia.
2TLU: opical li es ock uni .
3Allows us o keep hose componen s wi h eigen alues g ea e han uni y (>1).
4Tables S1–S7 a e Supplemen a y Ma e ials.
5H0: No he e ogenei y: all Ui = 0.
6H0: The andom e ec s model is p e e ed; H1: he ixed e ec s model is p e e ed.
7Ho: The speci ied endogenous eg esso s a e exogenous.
8The panel da a used in his s udy a e unbalanced.
9ESPS (E hiopian Socio-Economic Panel Su ey, 2018/19 and 2021/2022) is he main da a used in he p esen s udy.
10 Each measu ed on a con inuous scale o a minimum o 0 and maximum o 1.
11
Due o p e ailing con lic in he coun y, da a o he Tig ay egion we e una ailable in he la es ollow-up su ey. Also, he
ecen ly o med ou egions we e conside ed oge he wi h hei o me pa en egion (SNNP) pe he s udy da a.
12
The inal model speci ica ion (column D, Table 6) excludes a ime dummy (yea - ixed e ec s) due o he limi ed na u e o he
panel da a, which co e only wo pe iods, and he use o se e al local, egional, and global a iables. The nume ous local and
ex a-local a iables esul ed in collinea i y wi h he ime dummy, making he co e a iables di icul o iden i y. Howe e ,
inco po a ing global and egional indica o s o e ed mo e ad an ages. As he ime dummy assumes uni o m e ec s ac oss
he e ogeneous u al communi ies, i is less sui ed o he mic o-le el analysis. In e nal con lic s in E hiopia, especially since 3
No embe 2020, ha e in ensi ied, pa icula ly in he Tig ay egion, which is excluded om he su ey da a. The Amha a egion
is also expe iencing ac i e con lic bu emained la gely s able du ing da a collec ion, aside om a b ie inciden . Figu e 3also
clea ly shows ha he dependen a iable (RD) has consis en ly inc eased since 2016 and emained s able du ing he su ey
pe iod (2018/19–2021/22). S a is ical es s o yea dummies, using he - es pa m- command, indica e hei join insigni icance
(p> 0.5) in he model, u he suppo ing he decisions made in ou speci ica ion.
