Feue bache , A nd ; Luckmann, Jonas
A icle — Published Ve sion
Labou ‐sa ing echnologies in smallholde ag icul u e: An
economy‐wide model wi h ield ope a ions
Aus alian Jou nal o Ag icul u al and Resou ce Economics
P o ided in Coope a ion wi h:
John Wiley & Sons
Sugges ed Ci a ion: Feue bache , A nd ; Luckmann, Jonas (2023) : Labou ‐sa ing echnologies
in smallholde ag icul u e: An economy‐wide model wi h ield ope a ions, Aus alian Jou nal o
Ag icul u al and Resou ce Economics, ISSN 1467-8489, Wiley, Hoboken, NJ, Vol. 67, Iss. 1, pp. 56-82,
h ps://doi.o g/10.1111/1467-8489.12502
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/287818
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 p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/
56
|
Aus J Ag ic Resou Econ. 2023;67:56–82.
wileyonlinelib a y.com/jou nal/aja
Recei ed: 2 Decembe 2021
|
Accep ed: 10 Decembe 2022
DOI: 10.1111/1467-8489.12502
ORIGINAL ARTICLE
Labou - sa ing echnologies in smallholde ag icul u e:
An economy- wide model wi h ield ope a ions
A nd Feue bache 1 | JonasLuckmann2
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion-NonComme cial-NoDe i s License, which
pe mi s use and dis ibu ion in any medium, p o ided he o iginal wo k is p ope ly ci ed, he use is non-comme cial and no
modi ica ions o adap a ions a e made.
© 2023 The Au ho s. The Aus alian Jou nal o Ag icul u al and Resou ce Economics published by John Wiley & Sons Aus alia,
L d on behal o Aus alasian Ag icul u al and Resou ce Economics Socie y Inc.
1Ins i u e o Ag icul u al Policy and
Ma ke s, Ecological and Economic Policy
Modelling G oup, Uni e si y o Hohenheim,
S u ga , Ge many
2In e na ional Ag icul u al T ade and
De elopmen G oup, Humbold - Uni e si ä
zu Be lin, Be lin, Ge many
Co espondence
A nd Feue bache , Ins i u e o Ag icul u al
Policy and Ma ke s, Ecological and
Economic Policy Modelling G oup,
Schwe zs . 46, 70599 S u ga , Ge many.
Email: a. eue bache @uni-hohenheim.de
Funding in o ma ion
S i ung ia panis, Ulm, Ge many;
Humbold - Uni e si ä zu Be lin
Abs ac
Labou - sa ing echnologies a e ele an o ag icul u al
de elopmen . Ye , as his s udy shows, hey a e poo ly
in eg a ed in o ag icul u al p oduc ion unc ions o
economy- wide models. We epo a compu able gene al
equilib ium (CGE) model, which explici ly inco po a -
ing ield ope a ions (e.g. land p epa a ion, weeding o
ha es ing) in he con ex o smallholde ag icul u e. The
ield ope a ions app oach allows o model echnological
ade- o s in o ganic and con en ional p oduc ion sys ems
a a ious s ages o he ag icul u al p oduc ion p ocess.
Simula ing a s uc u al change scena io, we compa e he
pe o mance o he ield ope a ions app oach wi h pub-
lished benchma k p oduc ion s uc u es by assessing how
hey eplica e empi ically obse ed changes in land and
ag ochemical use. This benchma k analysis shows ha
inco po a ing ield ope a ions eplica es he obse ed em-
pi ical changes mos accu a ely and allows o mo e eal-
is ic modelling o labou - sa ing echnologies. We use he
ield ope a ions model o in es iga e h ee policy op ions
o mi iga e labou sho ages in he ag icul u al sec o o
Bhu an. Pe mi ing he employmen o Indian wo ke s
in ag icul u e has he highes sho - e m po en ial in his
espec . We ind ha subsidising ag icul u al machine y
hi ing se ices and emo ing impo a i s on ag ochemi-
cal inpu s a e ound o be less e ec i e. Fu he op ions
o model de elopmen s, such as combining ield ope a-
ions and labou ma ke seasonali y, a e highligh ed.
KEYWORDS
applied gene al equilib ium, baseline o ecas ing, cons an elas ici y
o subs i u ion, economic modelling, Leon ie echnology, model
alida ion
Feue bache and Luckmann ha e con ibu ed equally o his pape
|
57
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
1 | INTRODUCTION
His o ically, mos economies ha e pu sued de elopmen ajec o ies h ough which g ow h
in he seconda y and e ia y sec o s was accompanied by he p ima y sec o eleasing su -
plus labou (Ch is iaensen & Ma in,2018; Gollin,2014). Once su plus labou is no longe
a ailable, i.e. he Lewis u ning poin is eached, he ag icul u al sec o has o inc ease i s
p oduc i i y o elease labou o he es o he economy (Lewis,1954; McMillan e al.,2014).
The challenge o inc easing ag icul u al p oduc i i y is usually add essed by land- and labou -
sa ing echnologies, speci ically h ough he adop ion o , in e alia, high- yielding c op a ie -
ies, chemical e ilise s, a m machine y, and pes icides (Galla do & Saue ,2018). These ha e
been he main ing edien s o mode n ag icul u al de elopmen , pa icula ly du ing he e a o
he G een Re olu ion (Conway & Ba bie ,2013), and ha e led o a s ong educ ion in ag icul-
u al employmen (Collie & De con,2014). These ea u es o economic ans o ma ion a e o
high ele ance o policy analysis in he con ex o low- income coun ies, whe e a la ge sha e
o he popula ion's li elihood s ill elies on ag icul u e.
Ag icul u e and especially plan p oduc ion is subjec o biophysical p ocesses (wea he ,
soil e ili y, he occu ence o pes s, e c.) and a me s' decisions (on he p oduc ion sys em,
c op choice, ime o plan ing, quan i ies o e ilise applied, e c.). Plan p oduc ion ollows a
c op calenda , which de e mines he espec i e ield ope a ions ( om seeding o ha es ing)
and hei sequence (An le,1983; Jagnani e al.,2021). The complexi y o ag icul u al p oduc-
ion sys ems especially mani es s in low- income coun ies in he opics and sub opics, whe e
he ag icul u al sec o accoun s o a high sha e o GDP and employmen while being cha -
ac e ised by high labou in ensi y and small landholdings (F ija e al.,2020). The adop ion
o labou - sa ing echnologies in smallholde a ming is a con inuous p ocess and o en only
conce ns selec i e s ages o he ag icul u al p oduc ion p ocess. The use o powe ille s, o
ins ance, lessens he labou equi emen o land p epa a ion, bu o he ield ope a ions, such
as plan ing, weeding and ha es ing, may emain una ec ed. The in oduc ion o he bicides
o he adop ion o gene ically modi ied o ganisms may, o ins ance, solely educe he labou
needed o weeding.
The impac s o s uc u al (Bekke s e al.,2021; Mulanda Mulanda & Pun ,2021), policy
(Dixon & Rimme ,2022) o echnological change (Wi we & Bane jee,2015) on he ag icul-
u al sec o and he economy as a whole a e commonly assessed using economy- wide simu-
la ion models, such as compu able gene al equilib ium (CGE) models. Howe e , such models
ail o adequa ely depic he ea u es o smallholde sys ems, pa icula ly ega ding he ole
o labou (Dixon & Jo genson,2012). In e iewing he li e a u e, we show ha economy- wide
models o en inco po a e a a he simplis ic p oduc ion s uc u e, which does no allow us o
model he ealis ic po en ial o labou - sa ing echnologies. We epo an al e na i e and no el
p oduc ion s uc u e ha inco po a es ield ope a ions and hus pe mi s us o model echno-
logical ade- o s a a ious s ages o he ag icul u al p oduc ion p ocess. We demons a e
ha his app oach allows o a be e i wi h empi ically obse ed changes in he ag icul u al
sys em.
We use Bhu an as a case s udy, whe e he ag icul u al sec o employs app oxima ely 50%
o he labou o ce (Minis y o Labou and Human Resou ces,2019). C opping sys ems in
Bhu an a e cha ac e ised by small- scale p oduc ion, high labou in ensi y and a low use o
ag ochemicals. Thus, he p edominan p oduc ion sys em can be called ‘o ganic by de aul ’
JEL CLASSIFICATION
C68, J43, Q10
58
|
FEUERBACHER and LUCKMANN
( o b e i y, we e e o only ‘o ganic’ hence o h). This has led policymake s o sugges ha
he ag icul u al sec o should become 100% o ganic (Feue bache e al.,2018). Howe e , due
o inc easing le els o u banisa ion, u al labou sho ages a e becoming an u gen challenge
o a me s in Bhu an (MoAF,2013a, 2019a).
The con ibu ions o his s udy a e wo old. Fi s , we compa e he ield ope a ions model o
commonly used model app oaches based on a comp ehensi e li e a u e e iew. We simula e
a e e ence scena io o model he s uc u al change in Bhu an's economy and labou o ce
be ween 2012 and 2018. In his pe iod, he ag icul u al labou o ce dec eased by 5.9%. We
demons a e ha he ield ope a ions model ou pe o ms he benchma k app oaches in epli-
ca ing empi ically obse ed changes in land and ag ochemical use. Such igo ous compa isons
o p oduc ion s uc u es a e a he sca ce in he li e a u e bu highly ele an o assess he
alue- added and supe io i y o me hod de elopmen . The second con ibu ion comp ises an
analysis o h ee di e en policy esponses ha aim o mi iga e labou sho ages wi hin he ag-
icul u al sec o . These policies a e simula ed using he no el ield ope a ions model app oach
calib a ed o empi ical changes in land and ag ochemical use.
The emainde o his pape is s uc u ed as ollows: Sec ion2 gi es an o e iew o com-
monly used app oaches depic ing ag icul u al p oduc ion in CGE models. Sec ion3 in o-
duces he ield ope a ions model and da abase as well as he model benchma k app oach.
Sec ion4 in oduces he e e ence scena io ep esen ing he s uc u al change obse ed in he
Bhu anese economy in ecen yea s and h ee policy scena ios o mi iga e labou sho ages in
he ag icul u al sec o . In Sec ion5, he ou comes o he di e en model se - ups a e compa ed
and he esul s o he policy scena ios o Bhu an a e p esen ed. In sec ion6 we discuss he
modelling app oach's capabili y, u he model de elopmen op ions and policy implica ions.
Conclusions a e p esen ed in Sec ion7.
2 | AGRICULTURAL PRODUCTION STRUCTURES IN
ECONOMY- WIDE MODELS
In CGE models, he p oduc ion s uc u e o ac i i ies is desc ibed as a ‘ echnology ee’ con-
sis ing o a se ies o nes ed cons an elas ici y o subs i u ion (CES) p oduc ion unc ions o
ixed sha e agg ega es (i.e. ollowing he Leon ie assump ion). The unde lying assump ions
a e cons an e u ns o scale and sepa abili y, meaning ha he ma ginal a e o subs i u ion
be ween inpu s agg ega ed in one nes is independen o he quan i y o any o he inpu s used.
Howe e , he s uc u e o he ‘ echnology ee’ is a ely empi ically ounded and hus is o en
de e mined by esea che s' in ui ion (Simola,2015).
Mos s anda d models apply simple wo- o h ee- s age CES nes ing, whe eby he op- le el
agg ega e in e media e inpu s a e combined wi h o al alue- added (Figu e1). A he le el
below, on he one hand, alue- added is composed o p oduc ion ac o s and usually labou ,
capi al and land. On he o he hand, commodi ies a e agg ega ed o o m in e media e inpu .
The e is no consensus on when p oduc ion nes s should assume CES o Leon ie echnol-
ogy. Howe e , mos models assume (impe ec ) subs i u ion be ween alue- added and agg e-
ga ed in e media e inpu s as well as be ween single p oduc ion ac o s, o example, he GTAP
model (He el,1997). In e media e inpu s a e o en agg ega ed in ixed sha es, assuming no
adap a ion owa ds a ela i e p ice change in commodi ies, o example in he IFPRI s anda d
model (Lo g en e al.,2002) o in STAGE .2 (McDonald & Thie elde ,2015). This means,
o example, ha a a me always needs o pu chase a simila quan i y o seeds o p oduce a
ce ain quan i y o c ops, ha is seeds canno be subs i u ed by o he inpu s. This assump ion
is elaxed wi h he GTAP- AGR, which also allows o subs i u ion be ween pu chased ag icul-
u al inpu s (Keeney & He el,2005).
|
59
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
This simple p oduc ion s uc u e is widely applicable, as i equi es a limi ed amoun o
da a o pa ame e isa ion. Howe e , his app oach can be c i icised o i s o e simpli ica ions
and c ude assump ions, which migh lead o un ealis ic esul s, especially ega ding e ec s oc-
cu ing a he mic ole el scale. The e o e, in he li e a u e, he p oduc ion s uc u e has been
expanded by adding u he nes s, o example, o dis inguish labou ca ego ies o di e en
deg ees o subs i u abili y (McDonald & Thie elde ,2009).
Some esea ch ocusses especially on he p oduc ion s uc u e in he ag icul u al sec o
gi en i s ele ance o he use o land and wa e esou ces and i s con ibu ion o GDP, em-
ploymen and li elihoods in many coun ies. In addi ion, as poin ed ou abo e, he many ech-
nological ade- o s in ol ed make he ag icul u al p oduc ion sys em qui e complex. Some
s udies wi h a ocus on he ag icul u al sec o mo e in e media e inpu s such as e ilise ,
ag ochemicals, o eeds u o he alue- added side wi hin he p oduc ion nes o allow o
an adjus men o p oduc ion in ensi y (e.g. A güello & Valde ama- Gonzalez,2015; Jiménez
e al.,2021). The same has been done o d augh animal ploughing se ices as a p oduc cou-
pled wi h li es ock p oduc ion (Holden e al.,2005) and h ough he in eg a ion o i iga ion
wa e om di e en sou ces (e.g. Luckmann e al.,2014). The GTAP- AGR model links he
li es ock sec o mo e closely o he c opping sec o by in oducing a subnes unde he in e -
media e inpu composi e o di e en ia e eeds u s, which a e mo e easily subs i u able om
non eeds u inpu s (Keeney & He el,2005).
Osman e al.(2016) use a subannual ime dimension by in oducing seasonal c opping ac-
i i ies and season- speci ic wa e supply. Dixon and Rimme (2021) sol e hei model in qua -
e ly ime s eps o model seasonali y in he ag icul u al sec o , and Feue bache e al.(2020)
model seasonal labou ma ke s by in eg a ing he mon hly labou demand o ag icul u al ac-
i i ies wi hin he p oduc ion s uc u e. Kuipe (2005) de elops a illage- le el CGE model wi h
a de ailed ag icul u al p oduc ion s uc u e accoun ing o (impe ec ) subs i u ion be ween
chemical and o ganic e ilise (manu e), be ween animal and ac o ploughing and be ween
labou and chemical plan p o ec ion (PP). This app oach, howe e , s ill ea s labou as a
single p oduc ion inpu wi h only one di ec subs i u ion ela ionship wi hin he p oduc ion
s uc u e. Hence, he app oach does no conside he di e si y o subs i u ion ela ionships
and echnological choices ha allow o educe labou in ensi y. Chemical e ilise can eplace
manu e, and i equi es less labou o applica ion due o he highe nu ien densi y. This la-
bou sa ing po en ial is no e lec ed in he app oach used by Kuipe (2005). The same holds
o he ela ionship be ween ac o and labou - in ensi e animal ploughing.
Despi e he desc ibed de elopmen s in he p oduc ion s uc u e o he ag icul u al sec o ,
many ecen s udies wi h a ocus on ag icul u e employ s anda d p oduc ion s uc u es. Fo
example, he s anda d IFPRI p oduc ion s uc u e is used by Ben ica e al.(2019) o in es iga e
he implica ions o an ag icul u al in es men plan in Mozambique and by Mulanda Mulanda
and Pun (2021) o analyse changes in ansac ion cos s and capi al a ailabili y in he Zambian
ag icul u al sec o .
FIGURE 1 P oduc ion s uc u e in a s anda d CGE model (Sou ce: Adap ed om Lo g en e al.,2002 p. 9)

60
|
FEUERBACHER and LUCKMANN
3 | METHOD AND DATA
3.1 | Model
The CGE model adap ed o his s udy is a single- coun y, compa a i e- s a ic CGE model
mainly de eloped om STAGE .2 (McDonald & Thie elde ,2015) and STAGE- DEV models
(A agie e al.,2016). We modi y he model's p oduc ion sys em o inco po a e he ield op-
e a ions o c opping ac i i ies. In he ollowing, we e e o his model se - up as ‘ ieldops’. To
adequa ely assess he me i s o his model de elopmen , we compa e he ieldops se - up o wo
benchma k model se - ups ha e lec commonly applied p oduc ion s uc u es in economy-
wide modelling (see Sec ion3.2). Excep o di e ences in he p oduc ion sys em, simula ions
wi h he ieldops and benchma k se - ups o he wise ely on iden ical model pa ame e s (see
AppendixA).
The agen s in he model a e p oduc ion ac i i ies, households, inco po a ed en e -
p ises, he go e nmen and he capi al ma ke . We model households' demand beha iou
( he demand sys em) as a wo- le el LES- CES nes . The LES le el is he linea expendi-
u e sys em de i ed om S one– Gea y u ili y unc ions assuming u ili y- maximising be-
ha iou . A his le el, households de e mine he op imum consump ion le els o agg ega e
commodi ies. A he CES le el, households choose wel a e- maximising combina ions o
‘na u al’ commodi ies subjec o ela i e commodi y p ices and he cons ain s o p e e -
ences, income, a ailable labou esou ces and subsis ence equi emen s. This se - up allows
households o subs i u e simila goods and se ices, o example ice, maize and o he ce-
eals. The income elas ici ies o demand o commodi y g oups a he LES le el we e es i-
ma ed using c oss- sec ional household da a om he 2012 Bhu an Li ing S anda d Su ey
(Feue bache ,2019). The CES pa ame e s used o agg ega e he commodi y g oups a e doc-
umen ed in AppendixB. De ails o he p oduc ion sys em a e p o ided below. Following
he A ming on(1969) insigh , demand o domes ically p oduced commodi ies is di e en-
ia ed om impo s and speci ied by a CES unc ion. Domes ically p oduced commodi ies
a e supplied o he domes ic and wo ld ma ke s (i.e. expo s) using cons an elas ici y o
ans o ma ion (CET) unc ions.
3.2 | Model se - ups wi h p oduc ion sys em a ian s
The ieldops and wo benchma k model se - ups di e in hei p oduc ion sys em design as ex-
plained below. This only conce ns he p oduc ion sys em o c opping ac i i ies, while all o he
economic ac i i ies emain unchanged. Gene ally, all model se - ups a e disagg ega ed by con-
en ional and o ganic p oduc ion sys ems, ha is whe he he use o ag ochemicals is allowed
o banned. A Cobb– Douglas unc ion is used o agg ega e he na ional ou pu o con en ional
and o ganic ac i i ies.
The model se - ups a e de eloped om he o iginal h ee- le el nes ed CES p oduc ion
s uc u e o STAGE .2. In CES p oduc ion unc ions, he lexibili y in he agg ega ion o
inpu s acco ding o hei ela i e p ices is de e mined by a pa ame e , subs i u ion elas ici y
𝜎
(see Pauw(2003) o a comp ehensi e o e iew). This pa ame e can ake alues be ween
0 and in ini y. I
𝜎=0
, he p oduc ion unc ions a e iden ical o a Leon ie echnology
acco ding o which p oduc ion inpu s a e agg ega ed in ixed sha es. The Cobb– Douglas
unc ion ep esen s a special case o he CES p oduc ion unc ion wi h a uni a y subs i-
u ion elas ici y (
𝜎=1
) esul ing in a cons an alue sha e o inpu s. The poin es ima es
o he CES elas ici ies used in he h ee model a ian s a e p esen ed oge he wi h hei
sou ces in AppendixA.
|
61
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
3.2.1 | The benchma k se - ups
As a benchma k, we chose a CES nes ing based on s anda d CGE models ex ended by a
e ilise - land nes . This e lec s he app oach o A güello and Valde ama- Gonzalez(2015)
and allows o mo e lexibili y ega ding he p oduc ion in ensi y. Based on his model, we c e-
a e wo benchma k con igu a ions: benchma k_CES and benchma k_Leon ie . The only di -
e ence be ween hem is ha hey agg ega e alue- added a le el L2.2 (Figu e2a) using ei he
CES
(
𝜎
L
2.2 >0
)
o Leon ie
(
𝜎
L
2.2 =0
)
echnology. We use bo h se - ups, as empi ical s udies
show ha labou – land subs i u ion is impe ec , wi h elas ici y alues es ima ed a close o
ze o: o ins ance, Lopez(1980) epo s a mean land- labou subs i u ion elas ici y o 0.113 and
He el e al.(2016) epo a subs i u ion elas ici y alue o capi al– land– labou o ag icul-
u al ac i i ies o 0.24. We use he la e es ima e o he benchma k_CES se - up.
Apa om he alue- added nes subs i u ion speci ica ion a le el L2.2, bo h benchma ks
a e iden ical. In e media e inpu s and alue- added componen s a e agg ega ed acco ding o
Leon ie echnology (le el L1 in Figu e2a). In e media e inpu s a e also demanded in ixed
FIGURE 2 P oduc ion s uc u e o model se - ups (Sou ce: Au ho s' own elabo a ion). No e: All model se -
ups dis inguish be ween con en ional and o ganic c op p oduc ion. In o ganic p oduc ion, he e is no chemical
e ilise applica ion no chemical plan p o ec ion (PP). Hence, o hese ac i i ies, he nes s a L4.2 and L6.2 only
ha e he espec i e o ganic inpu .
62
|
FEUERBACHER and LUCKMANN
sha es (L2.1). In e media e inpu s include all commodi ies excep o chemical e ilise and
animal manu e, which a e in eg a ed on he alue- added side (see L4.2). Fe ilised land (L3.1)
and he capi al– labou composi e (L3.2) a e agg ega ed o o m o al alue- added a L2.2,
wi h ei he CES o Leon ie echnology. Land (a ea cul i a ed by c opping ac i i ies) and e -
ilise s a e agg ega ed using CES echnology. All o ms o capi al (including powe ille s) and
labou a e agg ega ed wi hin he espec i e nes s a L4.3 and L4.4. The nes a L4.1 agg ega es
he a ious land ypes (i iga ed and ain ed land), bu in his s udy, c ops a e linked o spe-
ci ic land ypes. No e: The e ilise agg ega e a L4.2 only comp ises e ilise commodi ies,
whe eas in he ieldops se - up, i will also include he co esponding labou needed in e ilise
applica ion.
3.2.2 | Field ope a ions se - up
The ieldops model se - up ex ends he p oduc ion s uc u e only o c opping ac i i ies o in-
eg a e he ield ope a ions (Figu e2b). The di e ences in he benchma k begin om L3.1 on-
wa ds, which go e ns he c opping ac i i ies' deg ee o in ensi ica ion. The e ilise agg ega e
a L4.2 he e consis s o he ield ope a ions o o ganic (manu e) and chemical e ilise applica-
ion. These wo ope a ions di e la gely in hei labou equi emen s (see Table1). Pe uni o
nu ien s, o ganic e ilisa ion equi es abou i e imes mo e labou compa ed wi h chemical
e ilise applica ion, as manu e mus be collec ed, s o ed and anspo ed o he ields, while
he nu ien densi y is much lowe as compa ed o chemical e ilise . These labou equi e-
men s o he applica ion o o ganic e ilise s a e no conside ed in he benchma k se - ups.
We modi y he nes a L4.1 o agg ega e land and all emaining ield ope a ions acco ding o
Leon ie echnology and e e o i as ‘a ea cul i a ed’. Assuming a ixed sha e be ween ield
ope a ions and land is easonable, inc easing he a ea cul i a ed would also lead o a highe
need o labou , which, o c opping ac i i ies, is mo ed om L4.4 o be included in he ield
ope a ions. In addi ion o land p epa a ion and PP, all ield ope a ions a e di ec ly agg ega ed
a L5.2 using Leon ie echnology. A L6.1 and L6.2, he h ee di e en land p epa a ion ech-
nologies and o ganic and con en ional PP p ac ices a e agg ega ed. No e ha only con en-
ional c opping ac i i ies can subs i u e be ween o ganic and con en ional PP.
The ield ope a ion ac i i ies hemsel es combine p oduc ion ac o s and in e media e in-
pu s as shown in Table1. As hey a e nonc opping ac i i ies, node L3.1 emains emp y in hei
p oduc ion s uc u e.
3.3 | Benchma k analysis
To assess he pe o mance o he ieldops model se - up, we simula e a Re e ence scena io e-
lec ing he s uc u al changes in Bhu an's labou ma ke and economy occu ing be ween
2012 ( he model's base yea ) and 2018 wi h he ieldops and wo benchma k model se - ups (see
Sec ion4.1). We compa e he esul s o each model se - up o empi ical es ima es o changes in
land and ag ochemical use. Be ween 2012 and 2018, he ag icul u al sys em in Bhu an gene -
ally in ensi ied wi h a 10.7% d op in he o e all c op a ea ha es ed. Paddy ice cul i a ion, he
mos labou - in ensi e c op in Bhu an wi h high ele ance o ood secu i y and sel - su iciency,
expe ienced a 10.6% decline. A he same ime, he use o chemical e ilise s and pes icides pe
a ea inc eased s ongly by 48.5% and 38.2% espec i ely.1 Based on hese ou indica o s, we
1The change in c opped a ea is based on MoAF(2013a, 2019a). Changes in ag ochemical use a e es ima ed ia a linea end
analysis o 5- yea mo ing a e ages. The annual quan i ies o chemical e ilise use a e aken om FAO(2020a), and uses o o al
pes icides a e aken om COMTRADE UN(2020) and Minis y o Finance(2020).
|
63
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
TABLE 1 Inpu – ou pu cos s uc u e o ield ope a ions
Inpu s
(million Nu.a)Nu se y
Land p epa a ion
Sowing /
Plan ingc
Fe ilise applica ion
O he
Ope a ions I iga ion
Plan p o ec ion
C op
Gua dingeHa es ing
MechanicalbD augh Manual O ganic Chemical O ganicdChemical
Fuel 24.5 2.4 11.3
Bulls 304.3
Manu e 194.9
Chemical
e ilise
68.4
Pes icides 53.5
Labou 113.3 23.3 222.4 11.5 180.0 216.9 6.0 87.6 90.0 345.8 6.8 320.9 648.9
Capi al 132.3 43.5
To al cos s 113.3 180.1 526.8 11.5 180.0 411.9 74.4 99.0 90.0 345.8 60.3 320.9 703.7
Physical
labou
inpu
(in 1000
pe son-
days)
643 132 1263 65 1022 1232 34 497 511 1964 38 1822 3685
Nu ien use
(in ons
o NPK
elemen s)
6986 954
Use by
c opping
ac i i y
Paddy ice,
ege ables
All c ops All c ops All c ops All c ops All c ops Con en ional
c opping
Veg’es,
po a oes,
spices,
ui s
Paddy ice,
o he
ce eals,
eg’es,
spices.
ui s
All c ops Con en ional
c opping
All c ops All c ops
a1 US$ = 53.4 Bhu anese ngul um (Nu.).
bMainly using powe ille s.
cIncl. ansplan ing.
dMainly manual weeding.
eAgains wildli e.
Incl. ha es machine y.
70
|
FEUERBACHER and LUCKMANN
specula ion since, acco ding o ou knowledge, he e a e no empi ical es ima ions o subs i u-
ion elas ici ies o hese wo ield ope a ions.
In he ieldops model, he no malised e o is 48% o e en 67% lowe han in he bench-
ma k_Leon ie and benchma k_CES models. This demons a es ha he ieldops model ou -
pe o ms bo h benchma k se - ups. Ye , his is only one indica o used o compa e he model
pe o mance o he di e en model se - ups. Fo ins ance, he benchma k_Leon ie equi es
only minimal changes o exis ing model s uc u es o achie e a easonably good esul . In
addi ion, he plausibili y o b oade esul s in he ag icul u al sec o also needs o be assessed,
which we do in he ollowing.
5.2 | Compa ing ag icul u al sec o esul s ac oss model se - ups
This sec ion compa es he ag icul u al sec o esul s o he ieldops and he wo benchma k
model se - ups using he calib a ed elas ici y alues (see Table3). A he mac o- le el, he h ee
se - ups epo simila posi i e e ec s on Bhu an's economy (Table4) ollowing he Re e ence
scena io, which e lec s he s uc u al change be ween 2012 and 2018 (see Table2). Howe e ,
he e a e subs an ial di e ences in how he ag icul u al sec o is a ec ed.
Ag icul u al wages inc ease be ween 35% in he benchma k_CES se - up and 46% in he iel-
dops se - up, e lec ing di e en deg ees o subs i u abili y o ag icul u al labou , while e u ns
o c opland dec ease (Table4). In he benchma k_CES se - up, e ilised land and labou a e
subs i u able in he alue- added nes (Figu e2), which allows o (pa ially) o se he educ ion
in he wo k o ce wi h inc easing land supply. The e o e, he benchma k_CES model is he leas
cons ained, and i s inc ease in ag icul u al wages and dec ease in c opland en s a e he lowes
among he h ee se - ups. F om an ag onomic pe spec i e, i is di icul o assess o wha deg ee
he land- labou subs i u ions epo ed by he benchma k_CES se - up a e plausible, as unlike in
he ieldops model, i is unknown o which ope a ions labou equi emen s a e educed.
FIGURE 4 Compa ison o model esul s o empi ical changes in key indica o s obse ed o es ima ed
om ends occu ing be ween 2012 and 2018 using (a) li e a u e- g ounded base elas ici ies and (b) calib a ed
elas ici ies. The empi ically obse ed changes in land use a e based on o icial ag icul u al s a is ics
(MoAF,2013b, 2019a). Changes in ag ochemical use a e es ima ed by applying a linea end analysis o 5- yea
mo ing a e ages. The annual quan i ies o chemical e ilise use a e aken om FAOSTAT (FAO,2020a), and
da a on he use o o al pes icides a e aken om COMTRADE (UN,2020) and Bhu an T ade S a is ics (Minis y
o Finance,2020).

|
71
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
The ieldops and benchma k_Leon ie se - ups do no allow o a di ec subs i u ion be ween
land and labou . This esul s in a mo e s able labou in ensi y compa ed wi h he benchma k_
CES se - up, as shown in he lowe pa o Table5 (and in AppendixF o single c ops). In
he ieldops se - up, he e is a mo e p onounced di e ence be ween con en ional and o ganic
c opping. As con en ional c opping allows o mo e labou - sa ing ield ope a ions, such as
chemical e ilisa ion and PP, he labou in ensi y o con en ional c opping declines s onge .
Also, a me s swi ch o he p oduc ion o less labou - in ensi e c ops. Due o bo h e ec s, con-
en ional c opping expands, while o ganic c op p oduc ion declines (AppendixG).
TABLE 4 Mac o- le el changes ela i e o he base
GDP componen s ( alued a base
p ices)
Base sha e o
GDP (%)
Change compa ed wi h base (%)
Benchma k_CES Benchma k_Leon ie Fieldops
GDP 100.0 34.4 34.4 34.6
Abso p ion (C + I + G) 134.2 21.1 21.0 21.1
Consump ion (C) 44.6 48.3 48.2 49.5
Ag icul u al households 16.4 29.3 29.3 30.3
Nonag icul u al households 28.2 56.7 56.6 58.2
In es men (I) 71.9 2.6 2.6 2.2
Go e nmen (G) 17.8 26.0 26.0 25.5
Expo s (E) 36.2 72.2 71.7 71.1
Impo s (M) 70.4 28.0 27.7 27.3
T ade balance (E- M) −34.2 −18.8 −18.9 −19.1
O he mac o- le el indica o s
Exchange a e (Domes ic cu ency/ o eign cu ency)−2.1 −2.0 −1.7
A e age household wel a ea37.8 37.6 38.7
Ag icul u al households 27.5 27.4 28.4
Nonag icul u al households 42.3 42.2 43.3
Fac o p ices
Base (USD
pe mon h)
Labou Ag icul u al 49.7 35.0 39.2 45.9
Unskilled 263.0 −0.9 −0.9 −0.6
Skilled 395.5 5.4 5.5 6.1
Land Paddy −23.7 −66.2 −47.7
Rain ed −5.3 −8.3 −5.2
Pas u e land 25.1 26.6 −16.3
Capi al Powe ille s 0.0 10.8 1.3
Ca le 5.8 4.0 −22.8
Bulls 43.9 58.7 −25.3
O he animals 19.0 20.2 −24.6
Uninco po a ed 6.7 6.7 7.2
Public 18.4 18.3 18.7
In o mal −0.8 −1.3 −0.7
a Measu ed as he equi alen a ia ion as a sha e o base income.
72
|
FEUERBACHER and LUCKMANN
TABLE 5 Land use and inpu in ensi y
Land use
Land ype Base (hec a e)
Change compa ed wi h base (%)
Benchma k_CES Benchma k_Leon ie Fieldops
To al c opped land TOTAL 76,017 −4.3 −9.1 −6.3
I iga ed 16,873 −10.6 −8.8 −10.8
Rain- ed 59,144 −2.5 −9.2 −5.1
Con en ional TOTAL 14,107 −3.8 −5.5 −6.7
I iga ed 4,368 −5.8 1.8 −4.9
Rain- ed 9,738 −3.0 −8.8 −7.5
O ganic TOTAL 61,911 −4.4 −9.9 −6.2
I iga ed 12,505 −12.2 −12.5 −12.8
Rain- ed 49,406 −2.4 −9.3 −4.6
Fallow land 21,291 15.3 32.6 22.6
Inpu in ensi y
Land ype Inpu Base- quan i y
Change compa ed wi h base (%)
Benchma k_CES Benchma k_Leon ie Fieldops
To al c opped land Labou 10.0 days/100 USD ou pu −6.3 −2.6 −1.5
To al nu ien s 104.4 kg NPK/ha −2.6 −4.6 5.3
Manu e 91.9 kg NPK/ha −7.1 −8.7 −1.2
Chemical e ilize 12.5 kg NPK/ha 30.8 25.1 53.3
Pes icides –
a5.6 18.2 37.0
Con en ional Labou 8.7 days/100 USD ou pu −11.1 −4.7 −7.4
Chemical e ilize 67.6 kg NPK/ha 30.2 20.3 53.9
Pes icides –
a5.1 13.8 37.5
O ganic Labou 10.4 days/100 USD ou pu −4.3 −1.2 0.8
aThe composi ion o pes icides is unknown bu assumed o emain unchanged.
|
73
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
O e all, mo e land is le allow wi h all se - ups (uppe pa o Table5). The ac eage unde
o ganic a ming is especially educed due o i s compa ably high labou in ensi y (lowe pa
o Table5, AppendixC) and gi en ha he e is no p ice p emium o o ganic p oduce in
Bhu an, while con en ional a ming e en expands somewha . The decline in land use is highes
o i iga ed (paddy) land due o he high labou in ensi y o paddy p oduc ion, causing a m-
e s o swi ch o less labou - in ensi e c ops.
Despi e he o e all educ ion in c opped ac eage and inc easing a ailabili y o powe ille s
as pa o he s uc u al change simula ed, he p ice o powe ille s does no decline (Table4).
In he benchma k_CES se - up, he CES elas ici y a he agg ega e alue- added nes allows
o a eac ion o he mo e expansi e capi al– labou agg ega e by subs i u ing i wi h land
(Figu e2). In o he wo ds, simila o labou , powe ille s could be eplaced o some ex en by
he expansion o ag icul u al a eas, which balances ou he subs i u ion e ec a he labou –
capi al nes . In he benchma k_Leon ie se - up, his is no possible, which is why he e he sub-
s i u ion be ween labou and powe ille s domina es, esul ing in inc easing en al p ices o
ille s. In he ieldops se - up, a di ec subs i u ion be ween ille s/labou and land is no pos-
sible by he design o he p oduc ion s uc u e. Ins ead, he exis ence o mo e labou - e icien
ploughing echniques is acknowledged by he di e en land p epa a ion ope a ions, which
can be subs i u ed o one ano he (Figu e2). Despi e he educ ion in land use, his esul s
in a highe demand o mechanical ploughing and hus a mode a e inc ease in ille p ices.
In con as , labou - in ensi e manual ploughing is especially educed (−69.0%), while d augh
bull ploughing emains cons an (Figu e5). The ex en o which demand o ield ope a ions
and hei p ices change is a use ul in o ma ion o iden i y whe e ag icul u al echnology in e -
en ions (e.g. mechanising ha es ing; educing labou needs o c op gua ding) may be mos
wo hwhile.
A c ucial di e ence be ween he model se - ups is ha he ieldops se - up cap u es he ech-
nological de elopmen away om animal- based manu e and d augh powe owa ds hei
mode n coun e pa s, chemical e ilise and mechanised ploughing. The lowe demand
o manu e and d augh powe educes hei p oduce p ices o such an ex en ha ac o
en s speci ic o li es ock all s ongly (Table4). These ela ionships a e no cap u ed in he
FIGURE 5 Supply (use) and p ice (cos ) changes o ield ope a ions.
74
|
FEUERBACHER and LUCKMANN
benchma k se - ups whe e manu e and d augh powe demand inc eases as he e is no link o
he ac ual labou equi ed o use hese inpu s in c opping.
Fu he mo e, he ieldops se - up allows o a mo e de ailed analysis o inpu use in he c op-
ping sec o . Fo mos ield ope a ions, labou is he main inpu . Thus, he associa ed cos s
inc ease a simila a es as he wages o ag icul u al wo ke s (Figu e5). Chemical e ilisa ion
and PP depend o a la ge ex en on nonlabou inpu s, namely impo ed chemical e ilise s
and pes icides. Fo hese, he wo ld ma ke p ice d ops by 6.7% and inc eases by 16.9%. Since
hey become ela i ely cheape compa ed wi h ag icul u al wages, mo e a me s a e incen i -
ised o use chemical inpu s, ha is a swi ch om o ganic o con en ional c opping ac i i ies.
In he ieldops se - up, his swi ch can be obse ed by he use o o ganic e sus chemical e -
ilise s and PP. As shown in he lowe sec ion o Table5, his change in inpu in ensi y is much
less p onounced in he benchma k se - ups. On agg ega e, wi h he wo benchma k se - ups, he
o al nu ien supply is dec easing, possibly esul ing in soil mining, while i inc eases wi h he
ieldops se - up, poin ing a in ensi ica ion.
5.3 | Analysis o policies mi iga ing ag icul u al labou sho ages
In his sec ion, he ieldops app oach (wi h calib a ed model pa ame e s) is applied o analyse
he e ec i eness o di e en policy measu es in mi iga ing he ad e se e ec s o labou sho -
ages in he ag icul u al sec o caused by he Re e ence scena io. The esul s a e summa ised
in Table6. The i s column p esen s he e ec s o he Re e ence scena io wi hou any policy
esponse and he ollowing columns epo he ou comes o h ee policy scena ios as de ia-
ions om he Re e ence scena io.
The ambi ious inc ease in he supply o public powe ille s (Scena io Public_Tille ) has
a e y limi ed e ec on mac oeconomic indica o s. In alignmen wi h expec a ions, he
highe a ailabili y o powe ille s leads o an expansion o c op p oduc ion and hus land
use since he p ice o ploughing (powe ille s and bulls) and animal manu e declines. As
p ices o chemical inpu s emain cons an , o ganic p oduc ion expands ela i ely mo e.
O e all, ag icul u al p oduc ion inc eases by 0.6% compa ed wi h he Re e ence scena io.
This policy esul s in an inc ease in ood sel - su iciency, especially o ice. Ye , i is by a
no enough o o se he ini ial educ ion caused by he Re e ence scena io. An in e es ing
aspec he e is ha a me s in Bhu an p edominan ly e ain om slaugh e ing animals
because o hei Buddhis o Hindi belie s (Samdup e al.,2010). Hence, he ‘phasing ou ’
o using bulls o ploughing would equi e much mo e ime, while less u ilised bulls s ill
equi e ca e and eed.
The emo al o impo axes on chemical e ilise s and pes icides (Scena io: Lib_AgChem)
educes hei pu chasing p ices by 9.4% and 9.3% espec i ely. This only bene i s he con en-
ional c op p oduc ion sec o , whe e cul i a ed land inc eases by 0.7%. Wi h la gely una ec ed
o ganic p oduc ion, his leads o a negligible inc ease in ag icul u al p oduc ion and sel -
su iciency. The labou in ensi y declines in alignmen wi h much highe le els o ag ochem-
ical use pe hec a e, bu he e is s ill a sligh inc ease in land use o 0.1%. Manu e p ices all,
e lec ing he subs i u ion o labou - in ensi e manu e wi h chemical e ilise in con en ional
c opping. The o e all low magni ude o his shock wi h espec o mac oeconomic indica o s
is due o he low in ensi y o ag ochemical usage in Bhu an. The esul s show ha e en o only
a sligh inc ease in a m household wel a e (0.3%), chemical applica ion a es need o inc ease
conside ably (Table6). The po en ially ad e se en i onmen al impac s o such a policy a e
unknown and no e lec ed in his s udy. They migh be negligible, gi en he low in ensi y
o ag ochemical use in Bhu an. Howe e , o ins ance, Bu achlo , a he bicide widely used by
paddy a me s in Bhu an, is known o esul in widesp ead nega i e impac s on amphibians
(Liu e al.,2011).
|
75
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
TABLE 6 Main esul s o policy scena ios (% changes)
Re e ence
scena io
Public_
Tille
Lib_
AgChem
Fo eign_
Lab
Compa ed
wi h base Compa ed wi h e e ence scena io
Mac o- le el indica o s
GDP 34.6 0.0 0.0 0.4
P i a e consump ion 49.5 0.0 0.0 0.6
T ade balance −19.1 0.0 0.0 −1.4
Exchange a e (domes ic pe o eign cu ency uni s) −1.7 0.1 0.0 0.2
Wel a e U ban
householdsaSkilled labou 43.0 0.2 0.1 2.5
Unskilled labou 38.9 0.3 0.1 2.8
O he income 63.0 0.4 0.1 3.7
Ru al
householdsaSkilled labou 35.2 0.2 0.1 2.6
Unskilled labou 36.6 0.3 0.1 3.1
Fa m 27.9 0.2 0.3 −1.0
Landless 35.9 0.6 0.1 −3.1
Income U ban
householdsaSkilled labou 35.4 0.1 0.0 0.6
Unskilled labou 28.8 0.1 0.0 0.5
O he income 68.3 0.2 0.0 1.4
Ru al
householdsaSkilled labou 30.1 0.1 0.0 0.5
Unskilled labou 29.8 0.1 0.0 0.6
Fa m 28.0 0.0 0.1 −0.4
Landless 37.6 0.2 0.0 −0.9
Wages Ag icul u al 45.9 0.4 0.0 −5.3
Unskilled −0.6 0.0 0.0 0.3
Skilled 6.1 0.1 0.0 0.3
P oduce p ices Ag icul u e 15.4 −0.9 −0.2 0.0
Food
p ocessing
2.4 −0.1 0.0 −0.9
Rice 3.9 −0.7 −0.1 −0.9
Manu ac u ing −2.7 0.1 0.0 0.4
O he
indus ies
−3.0 0.1 0.0 0.4
Se ices −4.6 0.1 0.0 0.3
Domes ic
p oduc ion
To al 37.6 0.0 0.0 0.5
Ag icul u e 3.3 0.6 0.2 2.3
Food
p ocessing
18.1 0.6 0.1 4.7
Rice 2.3 1.8 0.4 3.2
Manu ac u ing 78.9 −0.2 0.0 0.3
O he
indus ies
19.1 −0.1 0.0 −0.2
Se ices 49.1 0.0 0.0 0.3
(Con inues)

76
|
FEUERBACHER and LUCKMANN
Bo h he Public_Tille and Lib_AgChem scena ios come a a cos o he go e nmen 's bud-
ge . Inc easing he p o ision o powe ille s equi es he go e nmen o educe in es men s
in o he sec o s. Libe alising ag ochemical impo s educes a i e enues, which is o se by
negligible endogenous changes in he di ec income ax. By con as , in oducing a quo a o
he employmen o o eign wo ke s (Scena io Fo eign_Lab) only en ails adminis a i e cos s,
which a e a guably low bu no e lec ed in he model. The Fo eign_Lab policy would ha e
he g ea es po en ial o mi iga e labou sho ages. Unlike he o he wo op ions, his sce-
na io has some p onounced mac oeconomic e ec s. I inc eases GDP by 0.4% and p i a e con-
sump ion by 0.6%, while ag icul u al and o al p oduc ion ise by 2.3% and 0.5% espec i ely.
Ag icul u al wages all subs an ially by 5.3%, ye skilled and unskilled wages sligh ly inc ease,
mos ly o he bene i o nonag icul u al households. Despi e he d op in ag icul u al wages,
p oduce p ices o ag icul u al emain cons an since he inc ease in land use (+4.2%) is ac-
companied by inc easing land p ices (+2.7%). Food sel - su iciency inc eases (+2.1%), al hough
he educ ion expe ienced due o he Re e ence scena io is no e e ed. The highe a ailabili y
o labou p edominan ly bene i s o ganic p oduc ion (+4.6%). Land cul i a ed unde con en-
ional ag icul u e inc eases by 2.7%, while he use o chemical e ilise s and pes icides pe
hec a e d ops by 3.4% and 6.0% espec i ely. No su p isingly, he decline in ag icul u al wages
leads o a decline in he wel a e o ag icul u al and especially landless households, which is
mino compa ed wi h hei o iginal gains om he Re e ence scena io. By con as , he wel a e
o all nonag icul u al household g oups ises.
Re e ence
scena io
Public_
Tille
Lib_
AgChem
Fo eign_
Lab
Compa ed
wi h base Compa ed wi h e e ence scena io
Ag icul u al sec o indica o s
Sel - su iciency Base a e
Food 67.4 −11.4 0.4 0.1 2.1
Ce eals 62.9 −22.2 1.2 0.3 0.8
Rice 56.5 −13.8 1.6 0.3 2.8
Inpu p ices/
en s
C opland −10.2 0.8 0.1 2.7
Tille 1.3 −18.4 0.2 3.9
Bulls −25.3 −28.0 0.2 26.1
Manu e −15.1 −2.0 −1.3 4.6
Chemical e ilise −8.2 0.1 −9.4 0.4
Pes icide 13.8 0.1 −9.3 0.4
Inpu use Land To al −6.3 1.2 0.1 4.2
Con en ional −6.7 1.0 0.7 2.7
O ganic −6.2 1.3 −0.02 4.6
Labou (pe alue o ou pu ) −1.5 −0.9 −0.4 1.6
To al nu ien s (pe ha) 5.3 −0.6 1.8 0.2
Manu e (pe ha) −1.2 −0.4 −1.5 1.0
Chemical e ilise (pe ha) 53.3 −1.4 17.2 −3.4
Pes icides (pe ha) 37.0 0.0 7.7 −6.0
a Rep esen a i e household g oups a e disagg ega ed by loca ion and hei main sou ce o income. See Feue bache e al.(2017) o
mo e de ails.
TABLE 6 (Con inued)
|
77
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
6 | DISCUSSION
6.1 | Modelling labou - sa ing echnologies by inco po a ing ield ope a ions
Compa ing economy- wide model esul s o eal- wo ld obse a ions (o ‘baseline o ecas ing’)
is an elemen o model alida ion (Dixon & Rimme ,2013). I may o e guidance on how good
deduc i e me hods such as simula ion modelling app oaches depic ends obse ed in eali y
(Sa gen ,2013). The calib a ion app oach applied in his s udy may be a aluable al e na i e
as o en no da a a e a ailable o es ima e adequa e unc ional o ms and hei unde lying pa-
ame e s econome ically. Ye , a disclaime is wa an ed ha he calib a ed pa ame e s a e no
ans e able o o he con ex s, as hey we e i ed o he Bhu anese con ex . Full- scale model
alida ion is a non i ial unde aking. The mani old ac o s a ec ing an economy a e nei he
known no obse ed in hei en i e y, and i measu ed, hey a e p one o measu emen e o
and biases. As done in his s udy, compa a i e- s a ic model simula ions a e mos ly conduc ed
in a ce e is pa ibus ashion, ha is, only exogenous pa ame e s o in e es a e changed, while
o he s emain cons an . These wo ds o cau ion should be heeded, bu hey apply o all h ee
model se - ups equally and may inspi e u he esea ch in his ield o model alida ion.
We show ha common model app oaches (which we label benchma k se - ups) ail o ad-
equa ely e lec he labou - sa ing po en ial o mode n ag icul u al echnologies and, mos
no ably, he use o ag ochemicals, while he ieldops app oach allows o mo e a ge ed ag i-
cul u al scena ios. By inco po a ing ield ope a ions in o he ag icul u al p oduc ion s uc-
u e, we can explici ly model echnological means o subs i u e labou wi h he adop ion o
machine y, pes icides o chemical e ilise . Impo an ly, we sepa a e hese subs i u ion ela-
ionships since he adop ion o a single echnology (e.g. applying he bicides) has only limi ed
labou - sa ing po en ial.
Gene ally, he compa ison o he ieldops model o he wo benchma k se - ups sc u inises
he plausibili y o model esul s and, mo e speci ically, ac o subs i u ions. Modelling land as
an impe ec subs i u e o capi al, as in he benchma k_CES se - up, is e y likely o o e es i-
ma e he ag icul u al sec o 's capaci y o compensa e o he educ ion in labou a ailabili y,
which is co obo a ed by he weak eplica ion o he empi ically obse ed changes in ag o-
chemical and land use. Fu he , he ac ha he applica ion o addi ional e ilise and chemi-
cals also equi es labou is no (explici ly) ecognised in he benchma k se - ups, whe e e ilise
ep esen s a subs i u e o land and pes icide use is bound o he p oduc ion le el (Figu e2)
(No e ha o al agg ega e labou use is iden ical ac oss all model se - ups.).
Including ield ope a ions makes subs i u ion ela ionships much mo e explici , allowing us
o de e mine subs i u ion elas ici ies o labou e sus o he inpu s o each speci ic ask a he
han ha ing only one o m o land- labou elas ici y, ep esen ing he agg ega ed subs i u ion
possibili ies o he g owing season. Wi h he benchma k se - ups, i would be di icul o ex
an e de e mine a single subs i u ion elas ici y ha co ec ly cap u es all speci ic subs i u ion
ela ionships.
While his s udy demons a es he possibili y o cap u ing a high le el o echnical de ail
using a CGE model, i does no pe se imply ha he ieldops app oach is a sil e bulle o
modelling (smallholde ) ag icul u e, as i equi es subs an ially mo e da a. In case o limi ed
da a a ailabili y in he con ex o smallholde ag icul u e and less ocus on ag icul u al sec-
o al de ails, a se - up wi h a ixed sha es ela ionship be ween land and labou such as he
benchma k_Leon ie model may be conside ed he nex bes al e na i e, as i a oids un eal-
is ic subs i u ion possibili ies. In ac , unning he policy scena ios analysed he e wi h he
benchma k_Leon ie model yields simila ou comes o mac oeconomic indica o s, ye qui e
subs an ial di e ences emain in he ag icul u al sec o (AppendixH). Howe e , acco ding o
ou e iew o he li e a u e, he as majo i y o s udies ely on a CES labou – land subs i u-
ion ela ionship. P epa ing he li e a u e e iew o his s udy, we examined 30 CGE model
78
|
FEUERBACHER and LUCKMANN
s udies, ou o which only A güello and Valde ama- Gonzalez(2015) used a Leon ie nes
be ween land and labou .
6.2 | Fu he model de elopmen op ions
The inco po a ion o ield ope a ions wi hin he p oduc ion s uc u e o a CGE model is a con-
ibu ion owa ds imp o ing he depic ion o ag icul u al p oduc ion sys ems in economy- wide
simula ion models. The app oach equi es he necessa y da a (see AppendixD) o es ima e he
cos s uc u e o ield ope a ions. This in es men is pa icula ly wo hwhile in coun ies wi h
a labou - in ensi e ag icul u al sec o , which holds o mos low- income coun ies in he wo ld.
Wi h be e da a, echnological ade- o s could be modelled a he c op and ope a ion
le el. While he cos s uc u e o many ield ope a ions is simila ac oss c ops (e.g. applica ion
o manu e o c op p o ec ion), he mechanisa ion po en ial is c op speci ic and allows o
labou sa ings in mos ield ope a ions. This applies pa icula ly o ha es ing, which com-
pa ed o all o he ield ope a ions in Bhu an, equi es mos labou - days (Table1). Ye , so a ,
ag icul u al machine y is p edominan ly used o land p epa a ion in Bhu an. The e o e, in
his applica ion, land p epa a ion is he only ield ope a ion, which allows o mechanisa ion.
Howe e , he ield ope a ions model is lexible o in eg a e mo e and o he echnologies (such
as di e en ha es ing echnologies), depending on hei ele ance o he speci ic coun y con-
ex and he a ailabili y o an app op ia e da abase.
Field ope a ions a e pe o med in di e en pe iods o he c opping season, which na u-
ally calls o also conside he seasonali y o labou since labou sho ages mainly occu in
speci ic seasons such as du ing he plan ing o ha es ime. Seasonal labou has been al eady
cap u ed in CGE models (Feue bache e al.,2020). In his way, seasonal o eign employmen
could be in eg a ed as well. Howe e , his would equi e a conside ably mo e complex p o-
duc ion s uc u e. Along he same lines, labou accoun s could be disagg ega ed by gende ,
which would allow us o ocus on he wel a e implica ions o gende dispa i ies in ag icul u e
and o he sec o s. These addi ions may be add essed in u u e esea ch. The inco po a ion
o seasons and ield ope a ions could p o ide he basis o model sequen ial decision- making
in ag icul u e (An le,1983). Ye , his would equi e swi ching o ecu si e dynamic mode and
ideally spli ing he model in o in a- annual pe iods (see, e.g. Dixon & Rimme ,2021).
6.3 | Policy implica ions o mi iga ing labou sho ages in ag icul u e
As in many o he low- income coun ies, Bhu an's ag icul u al sec o su e s om labou sho -
ages (e.g. Leona do e al.,2015). Howe e , as his s udy shows, he po en ial o p omo e labou -
sa ing echnologies is limi ed. Bhu an's opog aphy and small- scale subsis ence- ocussed
ag icul u al sec o do no pe mi he use o la ge machine y, and e en he use o single- axle
powe ille s in ields may be p oblema ic in some a eas. The policy o expanding public powe
en al se ices by 150% can be conside ed ambi ious, while i s po en ial o mi iga e labou
sho ages emains limi ed. The same holds o he libe alisa ion o ag ochemical impo s, al-
hough his scena io esul s in conside ably inc eased p oduc ion in ensi y. Fos e ing he use
o ag ochemicals would be a qui e con o e sial policy in Bhu an, which has ambi ions o
become he i s coun y wi h a 100% o ganic ag icul u al sec o (Feue bache e al.,2018).
Only he in lux o o eign ag icul u al labou shows conside able e ec s o mi iga e he
labou sho age in ag icul u e and he associa ed decline in ag icul u al p oduc ion. The sim-
ula ed quo a o allowing 5.9% o Bhu an's ag icul u al labou o ce in 2012 o wo k in ag icul-
u e is equi alen o app oxima ely 12,000 wo ke s. Despi e he associa ed ansac ion cos ,
his scena io is ealis ic gi en he high wage di e en ial be ween Bhu an and India. As a ming
|
79
AN ECONOMY-WIDE MODEL WITH FIELD OPERATIONS
in Bhu an is s ill p edominan ly semisubsis ence ocussed wi h small a e age landholdings, he
scope o wo ke hi ing is p ima ily limi ed o comme cial a me s wi h abo e- a e age a m
sizes. Howe e , allowing Indians o wo k in ag icul u e is subjec o con o e sy (Ch is ensen
e al.,2012), as a ming and u al adi ions a e pe cei ed o be an in eg al pa o Bhu an's cul-
u e. This could possibly be add essed by imposing es ic ions, o example by limi ing wo k
pe mi s o empo al employmen o o Sou he n egions close o he Indian bo de .
7 | CONCLUSIONS
This s udy shows ha he selec ion o an adequa e p oduc ion s uc u e is no i ial o
economy- wide analyses o ag icul u al policies in coun ies whe e he ag icul u al sec o is
s ill domina ed by labou - in ensi e smallholde a ming sys ems. We p opose a model se - up
ha explici ly inco po a es ield ope a ions (e.g. land p epa a ion, weeding o ha es ing) in o
he p oduc ion s uc u e and hus depic s echnological ade- o s such as he use o labou -
sa ing echnologies and labou - in ensi e p ac ices. We conduc a de ailed benchma k analysis
o he p oposed no el model s uc u e by compa ing i o obse ed empi ical e idence such
ha di e ences can be igo ously aced. The ield ope a ions app oach allows us o eplica e
empi ically obse ed changes in ag ochemical and land use, while he common app oaches
ound in he li e a u e ail o do so. I a oids opaque and un ealis ic adjus men pa e ns when
modelling scena ios o s uc u al change, making beha iou al adjus men s in he ag icul u al
sys em mo e explici and hence aceable.
We use he ield ope a ions model o in es iga e h ee policy op ions o mi iga e labou
sho ages in he ag icul u al sec o o Bhu an. We ind ha pe mi ing he employmen o
Indian wo ke s in ag icul u e has he highes sho - e m po en ial in his espec . Subsidising
ag icul u al machine y hi ing se ices and emo ing impo a i s on ag ochemical inpu s
a e ound o be less e ec i e.
Modelling ield ope a ions and hus he po en ial o labou - sa ing echnologies ade-
qua ely in an economy- wide model is o high ele ance o esea che s and policymake s
conce ned wi h ag icul u al and u al de elopmen . Many in e en ions, such as conse -
a ion ag icul u e, clima e- sma ag icul u e o sus ainable in ensi ica ion, imply highe
labou equi emen s, and hei success is o en impeded by labou sho ages. On he con-
a y, new echnologies such as machine y- sha ing pla o ms, imp o ed c op a ie ies and
ag icul u al ex ension se ices may boos he adop ion o labou - sa ing echnologies. These
new echnologies and p ac ices a e being scaled up and dissemina ed, while p ocesses o
s uc u al change and policy e o m a e aking place. The e y na u e o economy- wide
models allows us o cap u e hese p ocesses and economic linkages beyond he ag icul u al
sec o . Wi h he con ibu ion o his a icle, such models may also be imp o ed o be e
ep esen he ag icul u al sec o and he speci ic eali ies o labou - in ensi e, smallholde
a ming sys ems.
ACKNO WLE DGE MENTS
A nd Feue bache acknowledges suppo om he ia panis ounda ion (Ulm, Ge many) o
inancing ield esea ch in Bhu an. We hank Sco McDonald, Ha ald G e he and pa icipan s
a he In e na ional Ag icul u al T ade and De elopmen semina a Humbold - Uni e si ä zu
Be lin. All emaining e o s a e ou own. The au ho s decla e ha he e a e no con lic s o
in e es . Open Access unding enabled and o ganized by P ojek DEAL.
DATA AVAILABILITY STATEMENT
The da a ha suppo he indings o his s udy a e a ailable om he co esponding au ho
upon easonable eques .