K a iuk, Vasyl; Bukin, Edua d; He z eld, Thomas
A icle — Published Ve sion
“Fo whoe e has will be gi en mo e”? Land en al
decisions and echnical e iciency in Uk aine
Land Use Policy
P o ided in Coope a ion wi h:
Leibniz Ins i u e o Ag icul u al De elopmen in T ansi ion Economies (IAMO), Halle (Saale)
Sugges ed Ci a ion: K a iuk, Vasyl; Bukin, Edua d; He z eld, Thomas (2024) : “Fo whoe e has will
be gi en mo e”? Land en al decisions and echnical e iciency in Uk aine, Land Use Policy, ISSN
1873-5754, Else ie , Ams e dam [u.a.], Vol. 146,
h ps://doi.o g/10.1016/j.landusepol.2024.107336 ,
h ps://www.sciencedi ec .com/science/a icle/pii/S0264837724002898
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/302221
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/4.0/
“Fo whoe e has will be gi en mo e”? Land en al decisions and echnical
e iciency in Uk aine
Vasyl K a iuk
a,*
, Edua d Bukin
b
, Thomas He z eld
a,c
a
Leibniz Ins i u e o Ag icul u al De elopmen in T ansi ion Economies (IAMO), Theodo -Liese -S . 2, Halle (Saale) 06120, Ge many
b
Jus us-Liebig-Uni e si ¨
a Giessen, Ins i u e o Ag icul u al Policy and Ma ke Resea ch, Senckenbe gs aße 3, Gießen 35390, Ge many
c
Ma in-Lu he -Uni e si ¨
a Halle-Wi enbe g, Uni e si ¨
a spla z 10, Halle (Saale) 06108, Ge many
ARTICLE INFO
Keywo ds:
Land en al ma ke
E iciency
Land concen a ion
Land e o m
Uk aine
ABSTRACT
Land en al ma ke s had played a c i ical ole in p o iding a ms access o Uk aine’s ag icul u al land be o e he
ban on land sales was li ed in 2021. This pape examines whe he en al-based land ela ions can p omo e land
use by mo e p oduc i e a ms in he con ex o impe ec ins i u ions. In pa icula , we examine whe he
Uk ainian a ms’ decisions o en land a e linked o hei ag icul u al abili y. U ilizing a ich panel o mo e han
16,000 Uk ainian comme cial ag icul u al p oduce s o 2005–2015, we analyze demand-side de e minan s o
pa icipa ion in he land en al ma ke . The e idence sugges s ha a ms’ o al ac o p oduc i i y is discon-
nec ed om hei decisions o en land. Land accumula ion appea s o be d i en by o he con ex - ela ed ac o s,
including exis ing local land concen a ion and o ien a ion owa d cash c op p oduc ion. Resul s call o
launching a land sales ma ke and imp o ing en al ma ke in as uc u e because hese measu es can align land
en al p ices wi h he alue o he ma ginal p oduc o land.
1. In oduc ion
Al hough op imal ag icul u al o ganiza ion and espec i e land
policies ha e been puzzling schola s o mo e han h ee decades and
s ill p esen an ac i ely deba ed issue, ela i ely li le a en ion has been
paid o a pos -socialis ansi ion con ex . Li e a u e has ex ensi ely
conside ed land ma ke s as a mechanism o c ea ing oppo uni ies o
land ans e s om less o mo e p oduc i e a ms and as a po e y-
educing ool (de Jan y e al., 2001; Deininge , 2003). Al hough land
owne ship is widely conside ed o induce in es men s and imp o e
p oduc i i y (Besley, 1995; Koi ala e al., 2016; Place, 2009), land sales
ma ke s a e o en di icul o implemen wi h impe ec ins i u ions.
Because many ansi ion economies a e known o high ansac ion cos s
in land pu chases, hin land sales ma ke s, and c edi cons ain s
(Le man e al., 2004), en ing land is o en he p ima y way o ans-
e ing land-use igh s. In hese en i onmen s, land en al ma ke s may
play an impo an ole in enhancing ag icul u al e iciency because hey
may enable land o low owa ds mo e e icien p oduce s (Sadoule
e al., 2001) and p o ide access o land o he poo (Deininge and
Binswange , 2001). The idea is ha land en al ma ke s may minimize
ansac ion cos s o land exchange and a e, hus, mo e e icien han
o he o ms o land ela ions (Deininge e al., 2009; Deininge and
Binswange , 2001; V anken and Swinnen, 2006). Howe e , e y li le
esea ch has been conduc ed on how en al ma ke s ac ually wo k in
se ings wi h impe ec ins i u ions. Ren al ma ke s may be hin, local
eli es may be able o exe ma ke powe , dep essing p ices and accu-
mula ing land, and landowne s’ lack o nego ia ing capaci y may dis up
land low owa ds mo e capable p oduce s. This s udy explo es whe he
land en al ma ke s con ibu e o an e ec i e land exchange in a
pos -So ie se ing wi h impe ec ins i u ions.
Exis ing li e a u e on he ela ionship be ween en al decisions and
ag icul u al abili y almos exclusi ely ocuses on a m households whe e
non-sepa abili y and he exis ence o mul iple income sou ces migh
acili a e access o land (Huy and Nguyen, 2019; Rahman, 2010; Tan
e al., 2018). Regionally, exis ing s udies ha e mainly ocused on
Sou heas Asia and A ica. Fo me So ie s a es di e subs an ially in
ins i u ional se up and land owne ship pa e ns. In Russia, Uk aine, and
Kazakhs an, o me wo ke s o collec i e a ms ecei ed land plo s as
pa o he de-collec i iza ion du ing he 1990s. Mos en hese plo s ou
o la ge comme cial a ms ins ead o s a ing a amily a m. Na u ally,
land en al decisions in comme cial a ms a e expec ed o di e om
household decisions, as he e a e no addi ional income sou ces, and new
challenges a ise wi h hi ed labo supe ision and emune a ion. I is
unclea whe he land en al ma ke s can acili a e
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (V. K a iuk).
Con en s lis s a ailable a ScienceDi ec
Land Use Policy
jou nal homepage: www.else ie .com/loca e/landusepol
h ps://doi.o g/10.1016/j.landusepol.2024.107336
Recei ed 11 Augus 2020; Recei ed in e ised o m 2 July 2024; Accep ed 22 Augus 2024
Land Use Policy 146 (2024) 107336
A ailable online 4 Sep embe 2024
0264-8377/© 2024 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license ( h p://c ea i ecommons.o g/licenses/by/4.0/ ).
p oduc i i y-enhancing land dis ibu ion in hese ci cums ances. We
add ess hese esea ch gaps by examining he link be ween a ms’ abili y
and hei p opensi y o en land.
Uk aine ep esen s an in e es ing case because land en al has been
an almos exclusi e way o access ag icul u al land. The eason is ha by
adop ing he new Land Code in 2002, he Uk ainian go e nmen
in oduced p i a e p ope y igh s on land, ollowed by an immedia e
mo a o ium o bidding land sales. Land-in ensi e ag icul u al p oduc-
ion wi h la ge comme cialized p oduce s pu s Uk aine in one g oup
wi h coun ies wi h simila na u al and en i onmen al condi ions as
well as compa able e o m pa hs: Russia and Kazakhs an. The Uk ainian
land en al ma ke had been in place o o e wo decades, leading o
subs an ial land use concen a ion, gi ing ise o la ge a ms ope a ing
on se e al hund eds o housands o ha (Keyze e al., 2017; Mamono a,
2015). One explana ion o he eme gence o la ge-scale a ms and local
land concen a ion could be he pe sis ence o economies o scale in c op
p oduc ion and a highe e u n o land in la ge p oduc ion uni s. How-
e e , Deininge e al. (2018) ha e no been able o p o e ha la ge
a ms a e mo e p oduc i e in Uk aine. This aises he ques ion o how
en al decisions a e made and whe he p oduc ion e iciency plays a
ole.
The main objec i e o his pape is o analyze whe he he Uk ainian
land en al ma ke acili a es he con eyance o ag icul u al land o
mo e e icien ag icul u al p oduce s. Fi s , we e iew he li e a u e on
he ac o s a ec ing a me ’s en al decisions. The heo y sugges s ha
he alue o he ma ginal p oduc o land should mo i a e he decision o
en land, implying ha mo e p oduc i e a ms should ope a e la ge
a eas. Howe e , ini ial land concen a ion s emming om he imes o
collec i e a ming may dis o his ela ionship. Second, we use unique
a m-le el panel da a om Uk aine o ace he de elopmen s o he land
en al ma ke o e 11 yea s be ween 2005 and 2015. Then, we es ima e
he de e minan s o pa icipa ion in he land en al ma ke o com-
me cial ag icul u al p oduce s, analyzing he connec ion be ween
a ms’ ag icul u al abili y and he likelihood o en ing land. In doing so,
we explici ly conside he ini ial land concen a ion and how la ge a ms
may exploi hei ad an ageous posi ion. Based on he esul s, we
discuss he implica ions o he cu en land ela ions o he g ow h o
he Uk ainian ag icul u al sec o .
Ou esul s do no suppo he hypo hesis ha mo e p oduc i e a ms
en mo e land. O he con ex - ela ed ac o s appea o be he d i e s o
en al decisions. In pa icula , ini ial ayon-le el land concen a ion
appea s o gene a e a m size pola iza ion wi h u he g ow h o la ge
a ms and u he sh inkage o hei smalle coun e pa s. Mo eo e , he
a ailabili y o cheap land may gene a e incen i es o land-in ensi e
business models ocusing on la ge-scale cul i a ion o cash c ops wi h
a low added alue pe hec a e. Finally, he g ow h o he ag icul u al
sec o appea s o ha e been b ough abou no by he size expansion bu
by he en y o mo e p oduc i e and he exi o less p oduc i e a ms.
The es o his pape is o ganized in he ollowing ashion. Sec ion 2
a icula es ou li e a u e e iew ocusing on he e icien use o land as a
p oduc ion ac o and ini ial land concen a ion. Sec ion 3 p o ides a
b ie excu se o he ins i u ional con ex o Uk ainian land ela ions.
Da a and me hods a e p esen ed in Sec ion 4, whe eas we p esen he
esul s in Sec ion 5. Finally, Sec ion 6 concludes and discusses he im-
plica ions o ou indings.
2. Concep ual amewo k
Ou esea ch ques ions a e embedded in an old deba e abou he
a m size-p oduc i i y ela ionship because land en al is by a he mos
impo an way o accessing land in Uk aine. An obse a ion ha la ge
a ms end o be less p oduc i e has u ned in o a s ylized ac wi h a
coined e m – “in e se a m size – p oduc i i y ela ionship” (Chayano ,
1926; Eas wood e al., 2010; Eswa an and Ko wal, 1986; Lip on, 2009).
Well-known p incipal-agen p oblems o a m owne s’ supe ision o he
hi ed labo ha e been he p ima y explana ion o he in e se size –
p oduc i i y ela ionship (Ba e e al., 2010; Fede , 1985). Howe e ,
mode n echnologies (e.g., p ecision a ming) and managemen ap-
p oaches may a enua e he e ec o labo supe ision challenges
(Deininge and Bye lee, 2012; Rada and Fuglie, 2019). In his case, labo
cons ain s may be elaxed, unleashing po en ial d i e s o land accu-
mula ion i o he non-land ac o ma ke s a e unhinde ed. Because land
en al ma ke s ep esen a lexible way o accessing land (Sadoule e al.,
2001), hey could allow new a me s o g ow o con ibu e o land
accumula ion by la ge a ms. Unde s anding a a me ’s decision o en
land will p o ide a heo e ical ounda ion o es ing whe he mo e
e icien a me s en in mo e land.
2.1. Land en al decisions
The agg ega e p oduc i i y o an ag icul u al sec o may be
imp o ed ia an e ec i e land exchange mechanism. A unc ional land
en al ma ke can gene a e a low o land om less o mo e able a me s
(Sadoule e al., 2001). Unde u ilized land imposes oppo uni y cos s on
mo e able a me s and gene a es incen i es o land exchange, pu ing
land o he mos e icien use. As a esul , land en al should acili a e
e icien ac o alloca ion, imp o ing he agg ega e e iciency o land
cul i a ion.
Basic mic oeconomic heo y in o ms a a m’s decision o en land.
The depa u e poin is ha a p o i -maximizing a m will add uni s o a
p oduc ion ac o un il he ac o ’s p ice equals he alue o i s ma ginal
p oduc (Dasgup a e al., 1999; Lackman, 1977). Because he alue o
he ma ginal p oduc o land depends on he p oduc ion p ocesses, i s
alue will be highe o he a ms ha use all p oduc ion ac o s mo e
e icien ly. In o he wo ds, he decision o en in an addi ional land plo
depends on he a me ’s abili y o use i in he p oduc ion p ocess
e ec i ely. Thus, a a me wi h low ag icul u al abili y will no be able
o compe e in he en al ma ke , acing high oppo uni y cos s o holding
land, which gene a es incen i es o educe he po olio o en ed land.
Consequen ly, we expec a ms wi h supe io ag icul u al abili y o en
mo e land. This o ms a basis o ou es , wi h he cen al hypo hesis
s ipula ing a posi i e co ela ion be ween a a m’s ag icul u al abili y
and i s ac i eness on he land en al ma ke .
Fa me ’s en ing decisions may be a ec ed by o he incen i es i
en al ma ke s o o he ac o ma ke s show impe ec ions. Fo ins ance,
impe ec ac o ma ke s may incen i ize he sel -selec ion o business
models ocusing on ce ain c ops. On he one hand, household a ms
acing challenges en ing land and lacking al e na i e o - a m
employmen oppo uni ies may o e use labo because hey canno
adjus hei a m size (Sen, 1966). Con e sely, comme cial a ms may be
in a be e posi ion o o e come land en al ansac ion cos s, making i
easie o hem o expand hei sizes. In his case, he ma ginal p oduc
o land will exceed he en al p ice and lead o land o e u iliza ion
(Deininge e al., 2018). Relaxed cons ain s o la ge comme cial a ms
may incen i ize less in ensi e cul i a ion. Fo ins ance, in he Chinese
con ex , less labo -in ensi e g ain p oduc ion was ound o be associa ed
wi h less cons ained access o land (Min e al., 2017; Qiu e al., 2020).
The lowe demand o labo pe uni o land in g ain p oduc ion
compa ed o, o ins ance, li es ock o ege able p oduc ion allows o
deal wi h labo sho ages and/ o high supe ision cos s. A he same
ime, mo e mechanized p oduc ion sys ems a e expec ed o bene i om
la ge cul i a ed a eas (Luo, 2018; Pingali, 2007). Thus, in he Uk ainian
con ex , land-in ensi e business models o ien ed owa ds expo c ops
(co n, whea , and sun lowe ) should be associa ed wi h a highe land
en al ac i i y.
2.2. Land concen a ion ia en al ma ke s
Ini ial dis ibu ion o land on a local le el may a ec he ansac ion
cos s o land exchange, impeding he achie emen o e iciency-
enhancing dis ibu ion o land. In pa icula , loca ions wi h subs an ial
amoun s o land con olled by la ge a ms may be subjec o p ohibi i e
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
2
ansac ion cos s o land access o smalle a ms o new en an s.
Fu he mo e, en ing la ge amoun s o land may equi e subs an ial
adminis a i e capaci ies o managing he con ac s. Thus, la ge a ms
a e ypically be e posi ioned and will ha e an ad an age in nego ia ing
wi h local landowne s. As a esul , new o smalle a me s may ace
addi ional cos s o land access and may be disad an aged in he
compe i ion o land access compa ed o la ge coun e pa s. Addi-
ionally, a me s cul i a ing small a eas may be excluded om local
poli ical p ocesses (Acemoglu e al., 2004; Binswange and Deininge ,
1997). This, in u n, may limi hei poli ical in luence on he
decision-making o local au ho i ies wi h espec o land dis ibu ion.
Fo ins ance, in luen ial a ms may ha e an ad an age in en ing
s a e-owned ag icul u al land
1
due o hei supe io ba gaining powe
ela i e o hei small coun e pa s.
La ge a ms may also ha e incen i es o cemen hei dominan po-
si ion and expand hei en ed land i he land p ices a e lowe han he
alue o he plo s’ ma ginal p oduc . The logic is ha se e al dominan
land use s may be able o de e po en ial en an s (Balmann e al., 2021;
Ma inelli, 2014). A si ua ion when a ew a ms con ol a subs an ial
sha e o land may discou age smalle a ms om en ing in land wi hin
his pa icula loca ion because o limi ed en al op ions. Fo ins ance,
dominan a ms may s a egically en speci ic plo s, inc easing he cos s
o land consolida ion o po en ial en an s (Ha igsen, 2015) o play a
ole o p ice leade s (G aubne e al., 2021), making en ing land wi hin
a gi en egion una ac i e.
As a esul , compe i ion o land based on e iciency conside a ions
may be hinde ed, gene a ing misalloca ion o land and labo . Small
a ms may be s uck wi h insu icien land en al possibili ies, whe eas
la ge coun e pa s may use hei poli ical and s a egic posi ion o
expand cul i a ed a eas u he . Because Uk ainian la ge a ms eme ged
on he basis o he o me collec i e a ms (Swinnen and V anken,
2007), hei dominan posi ion may had been a ained exogenously,
gi ing ise o ini ial inequali y in en ed land dis ibu ion. I a e age
en al p ices a e a below he ma ginal p oduc o land,
2
we expec all
ypes o a ms o compe e o land. Howe e , because o he lowe
ansac ion cos s, we expec a ms wi h a dominan posi ion in he local
en al ma ke o ein o ce hei posi ion and en in mo e land. On he
o he hand, smalle a ms ailing o compe e wi h la ge coun e pa s
may sh ink o exi ag icul u al p oduc ion, con ibu ing o a m size
pola iza ion. Wi h inc easing land concen a ion, we expec a
sel -selec ion p ocess whe e each emaining a m’s ma ke and poli ical
powe is expec ed o inc ease, and hus a ms migh en in mo e land.
Thus, we hypo hesize ha a ms loca ed in egions wi h ela i ely un-
equally dis ibu ed land should en in mo e land.
3. Land ela ions in Uk aine
A e he collapse o he So ie Union, Uk aine se he cou se o
libe al land e o ms in he ea ly 1990s bu hen go apped in he
ansi ion pe iod. Fig. 1 demons a es he e olu ion o he majo legis-
la i e ini ia i es. The i s impo an miles one ha signi ied he launch
o he land e o m was he 1995 P esiden ial Dec ee
3
which launched
he dis ibu ion o he so-called “condi ional land sha es” (CLS) – s akes
in he o me collec i e a ms. Employees o collec i e a ms and ce ain
g oups o he u al popula ion we e eligible o ecei e CLS. La e on, he
1999 P esiden ial Dec ee
4
ga e a chance o he CSL holde s o con e
hei sha es in o physical plo s (Le man e al., 2007), subsequen ly
c ea ing ca. 7 million land owne s, who owned 27.6 million ha. How-
e e , he ins i u ions necessa y o es ablishing a ma ke -based ex-
change o owne ship i les we e no de eloped hen. The 2002 Land
Code ha pa ed he way o new land ela ions in Uk aine could be
conside ed a majo b eak h ough because i clea ly de ined p ope y
igh s ela ed o ag icul u al land. Howe e , he subsequen ly adop ed
Mo a o ium on land sales ook away he basic igh o he landowne s o
sell land. Ini ially planned as a empo a y measu e, he Mo a o ium has
been p olonged en imes since i s adop ion. Only in 2020, he Uk ainian
pa liamen adop ed a law s ipula ing he launch o a es ic ed sales
ma ke in 2021.
In he mean ime, access o land o ag icul u al p oduce s was almos
exclusi ely based on land en al. Land en al p ices we e o med based
on a so-called ‘no ma i e mone a y alue’ o land, a e e ence alue
a i icially se based on soil quali y and expec ed e enues. Se e al ob-
se e s claim ha he ‘no ma i e mone a y alue’ o land was se way
below he alue o he ma ginal p oduc o land and ha he gap
inc eased o e he yea s (K a iuk and He z eld, 2019; Ni ye skyy,
2019; Ni ye skyy and Nizalo , 2016). Apa om being low, Kuns
(2017) sugges s ha land p ices we e o en paid in he o m o ag i-
cul u al p oduce. Uk ainian landowne s we e disad an aged because o
hei low ba gaining powe and, hus, hei abili y o exe upwa d
p essu e on p ices. Ren al p ices below he alue o he ma ginal p oduc
o land incen i ize land accumula ion and lead o o e u iliza ion o his
p oduc ion ac o (Deininge e al., 2018). In addi ion, we obse e
anecdo al e idence o illegal exp op ia ion o land and hos ile o e akes.
The combina ion o low p ices and enu e unce ain ies gene a es he
incen i es o business models o expo -o ien ed annual c op cul i a-
ion on la ge a eas (K a iuk and He z eld, 2019). These ins i u ional
se ings ha e con ibu ed o he eme gence o Uk aine as one o he
wo ld’s leading p oduce s and expo e s o g ains and oilseeds (Keyze
e al., 2017).
Wi hin hese ins i u ional se ings, h ee majo ypes o ag icul u al
p oduce s eme ged in Uk aine: ag icul u al en e p ises (including s a e
en e p ises), indi idual a ms, and households. Fig. 2 b eaks down land
use by each ype o ag icul u al p oduce and by land-use ype (owned
s. en ed). We see ha comme cial a ming on owned land is a e. A
ligh g ey colo ep esen s he small amoun o land owned by indi idual
a ms and ag icul u al en e p ises ha managed o acqui e small a eas
be o e he 2002 sales ban. Impo an ly, households p edominan ly a m
on owned land and accoun o ca. 4.3 million small ag icul u al p o-
duce s ha cul i a ed abou 15.7 million ha (43.1 % o he o al ag i-
cul u al land) in 2016. They mainly p oduce o subsis ence
consump ion, o ien ing excess o he local ma ke s. The S a e S a is ics
Se ice o Uk aine (SSSU) es ima ed ha households accoun ed o
38.7 % o he alue o ag icul u al c ops p oduced in 2016. Despi e he
households’ signi ican ole in Uk ainian ag icul u e, his s udy ocuses
on comme cial ag icul u al p oduce s: indi idual a ms and ag icul u al
en e p ises.
Family a ming has been p omo ed in Uk aine o acili a e he
comme cializa ion o he households ha en ed land in addi ion o
wha hey owned. Fo his pu pose, indi idual a ms we e de ined by he
law in 2003
5
and ha e been g an ed p e e en ial condi ions on es ab-
lishmen , axa ion, and access o land. Indi idual a ms ha e been
gaining impo ance and eached 4.4 million ha in o al land use in 2016
(12 % o he o al ag icul u al land), ep esen ed by abou 32,000 p o-
duce s. They a m p edominan ly on en ed land (89 % o he o al land
used by his ca ego y).
1
Roughly a qua e o all ag icul u al land was owned by he Uk ainian s a e
as o 2016.
2
We explain why ha is he case in he Uk ainian con ex in Sec ion 3.
3
Dec ee o he P esiden o Uk aine No. 720/95 om 08.08.1995 “On he
o de o dis ibu ion o lands ans e ed o a collec i e owne ship o ag icul-
u al en e p ises and o ganiza ions”.
4
Dec ee o he P esiden o Uk aine No. 1529/99 om 03.12.1999 “On u -
gen measu es o accele a ing he e o ms in ag icul u al sec o o he
economy”.
5
Law No. 973 om 19.06.2003 “On indi idual a ming”.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
3
Ag icul u al en e p ises ep esen legal en i ies ypically un by hi ed
manage s and o en owned by indi iduals no di ec ly in ol ed in
ag icul u e o by he s a e. Many o hese en e p ises we e c ea ed on he
basis o o me collec i e a ms bu almos exclusi ely ope a ed on land
en ed om he CLS holde s: hey owned only 1.5 % (200,000 ha) o he
o al land in 2015. The sha e o land ope a ed by en e p ises declined
a e es uc u ing bu s abilized in he mid-2000s. As o 2016, hey
accoun o oughly 16.3 million ha o 47 % o he o al ag icul u al land
while being ep esen ed by 9796 legal en i ies.
A dis inc i e ea u e o Uk ainian ag icul u e is a ela i ely la ge
ope a ional scale in compa ison o he es o he wo ld. Uk aine hos s
some o he la ges ag icul u al en e p ises in he wo ld. The a e age
u ilized land pe a m amoun s o 460 ha bu he la ges a ms ope a e
up o 500,000 ha (Deininge e al., 2018). A subs an ial sha e o land is
used by ag icul u al en e p ises be ween 500 ha and 4000 ha in size,
which likely ep esen en e p ises o ganized a ound o me collec i e
a ms. In addi ion, many a ms a e a ilia ed wi h a cen al holding
o ming mul ib anch a ms – so-called ‘ag iholdings’ (Deininge e al.,
2018; G aubne e al., 2021). Apa om be e manage ial and
adminis a i e capaci ies, hese a ms ypically enjoy be e access o
c edi and o he o ms o ou side capi al, con ibu ing o subs an ial land
concen a ion wi h o al land holdings spanning up o hal a million ha.
En e p ises ope a ing mo e han 10,000 ha accoun o 17.7 % o he
cul i a ed land (SSSU, 2017). Because he majo i y o hese a ms
inhe i ed CLS-land om collec i e a ms (Le man e al., 2007), land
concen a ion was a gi en condi ion ac oss Uk aine in he ea ly 2000s.
Thus, he unequal ini ial dis ibu ion o land, which gene a ed poli ical
and ma ke powe o la ge a ms, in combina ion wi h low en al p ices
may ha e con ibu ed o ex eme land concen a ion ou comes.
4. Da a and me hods
4.1. Da a
To es ou hypo heses ou lined abo e, we use a unique a m-le el
panel da ase co e ing he pe iod om 2005 o 2015. Da a has been
collec ed annually by he S a e S a is ics Se ice o Uk aine (SSSU).
Obse a ions a e sampled om a ms egis e ed as legal en i ies, i e-
spec i e o a m ype, and exceeding one ou o i e size h esholds
de ined by SSSU.
6
App oxima ely 8000 a ms a e co e ed by he sample
each yea accoun ing o 17 million ha o ag icul u al land (48 % o all
ag icul u al land). O e all, ou sample co e s 16,950 ag icul u al en-
e p ises and indi idual a ms dis ibu ed ac oss 625 ayons
7
(dis ic s)
o Uk aine, which p oduce abou hal o he domes ic ag icul u al
ou pu . This makes i a unique da a sou ce o medium- and la ge-scale
en e p ises in Uk aine.
8
Howe e , i is impo an o no e ha his da ase
Fig. 1. Timeline o he majo miles ones in he Uk ainian land e o m.
Fig. 2. Dis ibu ion o land use by a m ypes. Sou ce: SSSU (2017).
6
Su ey pa icipa ion h esholds emained s able be ween 2005 and 2015. In
gene al, any legal-en i y p oducing ag icul u al commodi ies has been sampled
i i we e quali ying o any o he ollowing h esholds: size abo e 200 ha;
numbe o ull- ime wo ke s abo e 20; numbe o li es ock abo e 50 heads o
ca le, o swine o sheep o 500 o poul y; o annual e enue abo e 150
housand UAH pe yea (ca. 6000 USD in 2020)
7
SSSU’s s anda d epo ing o m “50SG” co e s om 85 % in 2005 o 95 %
in 2015 o land ope a ed by all ag icul u al en e p ises (s a e and p i a ely
owned) as well as 79–60 % o land ope a ed by indi idual a ms egis e ed as
legal en i ies (based on da a om (SSSU, 2017)).
8
Fo he yea s 2014–2015, no da a is a ailable o he e i o ies occupied by
he Russian Fede a ion: C imea and pa o e i o ies o he Donbass egion.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
4
does no include small p oduce s ha do no pass he minimum c i e ia
o being included in he sampling popula ion se by SSSU. Fu he mo e,
i is an unbalanced panel wi h ela i ely high a i ion whe e only 24 %
o a ms a e p esen each yea (see Appendix A o u he de ails).
A ailable a iables include s anda dized a m managemen in o ma ion
and pe o mance s a is ics. I con ains da a on a ms’ expendi u es on
inpu s, p oduced c ops, and li es ock as well as olumes and alues o
sales in each calenda yea . Speci ically, he inpu da a dis inguishes
be ween land a ea, labo in ull- ime equi alen (FTE), and expendi u es
on land en , hi ed labo , seeds, e ilize s, pe ol, se ices, machine y,
and o he i ems.
9
We de la ed all mone a y da a o he 2018 p ice le el
using he Consume P ice Index (CPI) and hen con e ed he alues in o
US dolla s.
To ensu e da a consis ency and o na ow he scope o he analysis,
we es ic he da ase in a numbe o ways. Fi s , we ocus on a ms ha
specialize in c op p oduc ion, es ic ing he sample o he a ms ha
mee bo h o he ollowing c i e ia in each yea a a m is p esen in he
sample: 1) a sha e o cos s ela ed o c op p oduc ion in hei o al cos s
is g ea e han 25 % and 2) ope a ional size is abo e 5 ha.
10
Mo eo e ,
as we a e in e es ed in changes in en ed land, we exclude all he a ms
ha appea ed in he panel only once ( hose a ms ha exis ed o one
yea o we e quali ied o be co e ed by he su ey only once).
Table 1 p esen s he desc ip i e s a is ics. The i s obse a ion is
ha be ween 2005 and 2015, he a e age amoun o owned land
dec eased and en ed land inc eased o all ypes o p oduce s excep o
he s a e a ms ha con inued elying on he cul i a ion o s a e-owned
land. Impo an ly, he a e age a m size inc eased o all ca ego ies o
p oduce s. Secondly, we obse e conside able g ow h in land concen-
a ion. Obse ing he He indahl-Hi schman Index (HHI) o he land
used wi hin a gi en ayon,
11
we see a mode a e coun y-wide inc ease
om 0.22 o 0.31 h oughou ou panel. Fig. 3 highligh s he spa ial
di e ences in land concen a ion and hei changes o e he decade.
Main egions o g ain p oduc ion s e ching om he sou h h ough he
cen e and o he no heas o Uk aine expe ienced mode a e le els o
land concen a ion. On he o he hand, ag icul u al land appea s o be
mo e concen a ed in he no h-wes e n and wes e n egions. These
pa e ns la gely coincide wi h he ope a ion domains o la ge a ms.
Mo eo e , we obse e signs o he business models’ e-o ien a ion o-
wa ds la ge-scale g ain p oduc ion. Fi s , en e p ises appea o ha e
subs an ially inc eased he sha es o co n, soybeans, and oilseeds a he
expense o ba ley, suga bee s, and oa s. These g ains ha e been among
he majo expo commodi ies du ing 2005–2015.
Because he SSSU collec s he da a on a calenda yea basis, olumes
o c ops sold in each speci ic yea a e a ely he same as he p oduc ion
le els. Consequen ly, he amoun o ou pu sold may be highe o lowe
compa ed o he ac ual p oduc ion. In addi ion, he e may be a po en ial
bias due o he ans e p icing o c op ou pu wi hin e ically in e-
g a ed c op-li es ock a ms. We ollow Deininge e al. (2018) in dealing
wi h hese challenges and calcula e he alues o he p oduced c ops in
2018 USD o each a m based on he median c op p ices a he ayon
le el and indi idual a ms’ p oduc ion quan i ies.
Examining he da a on a ms’ pe o mance e eals clea di e ences
be ween he h ee ypes o ag icul u al p oduce s. Al hough a ms’
ou pu pe ha inc eased conside ably on a e age ac oss all a m ypes,
hese gains a e less p onounced o s a e en e p ises. Thei le el o
ou pu pe ha was only hal o he espec i e alue o ag icul u al en-
e p ises in 2015 al hough i also g ew oughly 50 % du ing he decade
o 2005–2015. These di e ences a e also aceable in he cos s uc u e,
indica ing ha s a e-owned ag icul u al en e p ises a e undamen ally
di e en om o he ypes o p oduce s. To al p oduc ion cos s wen up
2.7 imes o all a m ypes, whe eas s a e a ms expe ienced only a 1.4-
old inc ease. This may e lec he ac ha he manage s o s a e a ms
may u ilize ou da ed echnologies due o a lack o in es men s. Fo
ins ance, s a e a ms spen h ee imes less on e ilize pe ha in 2015,
and pe ol cos s almos did no change and we e only 50 % o he le el o
o he a m ypes. Impo an ly, al hough he land en al p ices almos
ipled o all a m ypes, s a e a ms, on a e age, paid only 19 % o he
a e age en al p ices in 2005 and 12 % in 2015. This illus a es hei
compe i i e ad an age in accessing cheape s a e-owned land.
4.2. Me hodology
Following ou heo e ical amewo k, we aim o explain ag icul u al
p oduce s’ pa icipa ion in he Uk ainian land en al ma ke s. In
pa icula , ou goal is o es whe he p oduce s wi h be e ag icul u al
abili y end o en in mo e land. This will help unde s anding whe he
cu en land ela ions s imula e a low o land owa d mo e e icien
ag icul u al p oduce s.
Fi s , we explain he amoun o land en ed by each a m i
Ren al inali in he las pe iod whe e i is obse ed in he sample. Ou key
explana o y a iable is a measu e o he a m i’s ini ial ag icul u al
abili y
α
ini i a he ime o i s i s appea ance in he sample. In line wi h
ou i s hypo hesis, we expec i o be posi i e should mo e p oduc i e
a ms en mo e land. In addi ion, we include he ini ial a ea o land
owned (Owned ini i), he le el o ini ial land concen a ion in ayon j
HHI ini ij a he ime when he a m en e ed ou panel, and o he a m-
speci ic cha ac e is ics as a ec o o con ol a iables Xi:
Ren al inali=β0+β1
α
ini i+β2Owned ini i+β3HHI ini j+β4Xi+
ε
i
(1)
As he en ed a ea a he end o he obse a ion pe iod o each a m
alls in an in e al be ween ze o and in ini y, we u ilize Tobi e-
g essions.
12
We minimize he isk o endogenei y be ween he en ed
land and he a ms’ economic pe o mance because we explain he inal
land en al wi h he ini ial ag icul u al abili y es ima ed a he ime o
he a ms’ en y in o he sample. This ensu es ha in mos cases, we
ha e la ge lags. The same logic is applied o ou p oxies o land con-
cen a ion and owned land.
Second, we explain he changes in land use o e he obse ed pe iod
be ween 2005 and 2015.
13
To achie e ha , we cons uc a new
dependen a iable o he model ep esen ed by Eq. 1: he di e ence in
en ed land a ea be ween he las and he i s a m-speci ic obse a ion
ollowing he yea o he p oduc ion unc ion es ima ion ΔA eai=
A ea inali−A ea ini iali. No e ha he di e ence in en ed land ex-
cludes he yea used o calcula e ag icul u al abili y o minimize endo-
genei y. The a iable ΔA eai cap u es only land use changes achie ed
h ough en al ma ke s, excluding land sales, which was p ohibi ed a
he ime. In addi ion, we es ima e his model sepa a ely o se e al sub-
samples: a ms wi h dec eased and inc eased u ilized land as well as o
a ms ha s ayed du ing he whole pe iod, exi ed, and en e ed he land
ma ke . We expec ha in he sample o he a ms wi h dec easing land
use mo e p oduc i e a ms demons a e a smalle decline in hei land
9
As he e is no in o ma ion on capi al and machine y pe se, we assume
machine y cos s being app oxima ed by he annual amo iza ion expendi u es.
Land cos s a e only eco ded o he c ops p oduc ion, while he li es ock
p oduc ion con ains eco ds o he eed cos s.
10
Reco ds on a ms smalle han 5 ha may be biased because o measu emen
e o s conside ing SSSU’s sampling s a egy ou lined abo e.
11
We ollow he li e a u e and de ine HHI as he sum o he squa ed sha es o
he land use s in a gi en ayon. As a esul , i anges om ze o o one. Because
owned ag icul u al land ep esen s only oughly one en h o he land used, we
ocus on he concen a ion o he land used.
12
In he Uk ainian con ex , a ms ypically do no en ou land as hey ha e
miniscule amoun s o owned land acco ding o he da a o S a e Se ice o
Geodesy, Ca og aphy, and Cadas e . Thus, we canno obse e en ing ou
pa e ns which ep esen s le -censo ing o ou dependen a iable a ze o.
13
We use he las pe iod o which da a is a ailable o hose a ms ha exi ed
he panel.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
5
Table 1
Desc ip i e s a is ics.
All a m ypes Ag icul u al en e p ises Indi idual a ms S a e a ms
2005 2015 2005 2015 2005 2015 2005 2015
Ou pu pe ha a ayon med. p ices* (2018 USD) 242.8
(212)
623.4
(488)
245.8
(211.4)
650.8
(504.1)
230.3
(170.3)
483.4
(320.8)
193.1
(255.5)
326.5
(257.4)
To al u ilized land, ha 2156.2
(2432.8)
2230.8
(4758.7)
2148.6
(2484.2)
2231.1
(5021.7)
1949.5
(1561.7)
2084.5
(1657.4)
2672.3
(2084.4)
2890.7
(4578.6)
Owned land, ha 176
(823.4)
94.5
(877.3)
93.3
(598.5)
28.7
(313)
49.4
(208.8)
31.5
(160.8)
2532.4
(2067.3)
2740.6
(4600.8)
Ren ed land, ha 1980.2
(2395.3)
2136.3
(4714.4)
2055.3
(2450.3)
2202.4
(5021)
1900
(1562.7)
2052.9
(1651)
139.9
(539.1)
150.1
(467.1)
Labo , a e age numbe o ull- ime wo ke s 101.3
(116.8)
47.4
(112.9)
100.6
(116.5)
48.7
(119.2)
73.1
(64.8)
30.7
(31.9)
162.9
(157.6)
78.9
(108.9)
Land concen a ion a ayon (HHI) 0.221
(0.238)
0.305
(0.297)
- - - - - -
To al cos s*, 2018 USD pe ha 175.9
(417.3)
486.6
(1783)
176.7
(433.2)
503.1
(1909.4)
170.5
(124.2)
425.6
(437.8)
166.7
(258.2)
229.3
(180.3)
Labo cos s*, 2018 USD pe ha 27.3
(127.4)
33.8
(172.9)
26.6
(130.6)
34.9
(185.1)
24
(23.2)
24.2
(37.7)
46.7
(125.9)
33.2
(59)
Pe ol cos s*, 2018 USD pe ha 30.9
(23.3)
57.5
(381.4)
30.9
(23.4)
58.4
(409.8)
33.2
(20.1)
57.7
(31.6)
28.9
(25.6)
32.6
(25.7)
Seed cos s*, 2018 USD pe ha 20.4
(22.4)
57
(181.4)
20.6
(22.4)
59
(183.8)
21.1
(16.3)
52.1
(187.3)
16.1
(27)
19.9
(18)
Fe ilize cos s*, 2018 USD pe ha 23.8
(39.9)
91.8
(113.4)
24.3
(41.2)
94
(116.5)
23.8
(24.5)
92.2
(93.7)
13.3
(19.7)
27.1
(42.2)
Machine y/capi al cos s*, 2018 USD pe ha 22.7
(55.7)
57.4
(129.3)
22.3
(57.3)
58.2
(134.5)
24.8 (27.8) 61.6 (98) 27.8 (43.5) 22.6 (30.2)
Se ice cos s*, 2018 USD pe ha 18.9
(56.8)
92.3
(513.1)
19.1
(58.4)
96.2
(550.4)
13
(41.7)
64.2
(78.8)
20.6
(31.8)
75.7
(87.8)
O he cos s**, 2018 USD pe ha 31.9
(153.6)
96.8
(638.5)
32.8
(160.9)
102.5
(685.8)
30.6
(22.6)
73.5
(57.4)
13.2
(34.5)
18.2
(28.7)
Land en al p ice, 2018 USD pe ha 22.2
(63.6)
58.2
(363.2)
22.5
(58.3)
60.9
(390.1)
20.7
(138.5)
50.6
(27.6)
4.3
(17.8)
6.8
(24.3)
Whea sha e in ha es ed a ea, % 37
(15.1)
31.5
(16.8)
36.9
(15.1)
31.1
(16.9)
35.6
(14)
33.1
(15.2)
40.1
(16.1)
35.3
(15.8)
Oilseeds sha e in ha es ed a ea, % 18.2
(14.5)
26.9
(17.2)
18.2
(14.5)
26.4
(17.2)
20.5
(14.8)
32
(15.7)
15.8
(13)
26.3
(18.5)
Co n sha e in ha es ed a ea, % 6.2
(10)
17.9
(19.9)
6.4
(10.3)
18.9
(20.5)
6.1
(7.7)
12.4
(13.5)
3.3
(4.2)
8.7
(14.1)
Soybean sha e in ha es ed a ea, % 2.2
(6.1)
10.2
(15.9)
2.2
(6.2)
10.6
(16.1)
1.7
(4.1)
7.5
(13.8)
2.1
(5.6)
7.3
(15.7)
Ba ley sha e in ha es ed a ea, % 19.1
(10.9)
8.2
(9.6)
19
(11)
7.6
(9.5)
19.8
(9.9)
11.3
(9.8)
20.9
(11.2)
12.4
(9.8)
Suga sha e in ha es ed a ea, % 3.7
(6.8)
1.3
(4.9)
3.8
(6.9)
1.4
(5.2)
3.6
(6.3)
0.5
(2)
2
(4.3)
0.6
(2.8)
Rye sha e in ha es ed a ea, % 0
(0)
0.6
(4)
0
(0)
0.6
(4.1)
0
(0)
0.3
(1.9)
0
(0)
1.6
(6)
Oa sha e in ha es ed a ea, % 2.9
(6.8)
0.6
(2.8)
2.9
(6.8)
0.5
(2.8)
2.2
(5.6)
0.4
(1.5)
3.4
(7.3)
2.6
(5.2)
Numbe o obse a ions 6136 7034 5594 6088 328 776 214 170
No es: S anda d de ia ions a e gi en in b acke s. Fo all a iables (excep o “Land concen a ion a ayon, (HHI)”) means and s anda d de ia ions a e weigh ed by he
a m size. * a iables a e denomina ed pe hec a e o land use (owned and en ed). ** Va iable “O he cos s” includes he cos s o access o land.
Sou ce: Au ho s’ calcula ions based on he SSSU da a.
Fig. 3. Spa ial dis ibu ion o land concen a ion (HHI index) a he ayon le el.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
6
holdings (i.e., posi i e sign o β
1
because he dependen a iable is
nega i e o his sub-sample). On he o he hand, mo e able a ms
should also g ow mo e in he sample wi h a ms ha expe ienced an
inc ease in a m size (i.e., posi i e sign o β
1
). Mo eo e , exi -en y
analysis should p o ide clues abou he ac o s in luencing he de-
cisions o en e , exi , o s ay on he en al ma ke .
Na u ally, he key o his analysis is o es ima e unobse able ag i-
cul u al abili y
α
i. We employ h ee di e en app oaches o es ima e
abili ies. Fi s , we ollow Schmid and Sickles (1984) in calcula ing a
a m-speci ic ixed e ec (wi hin ans o ma ion) using panel eg ession
(FE). Second, we calcula e echnical e iciency sco es based on he
S ochas ic F on ie Analysis (SFA) wi h ime-in a ian e iciency sco es
dis ibu ed acco ding o he unca ed no mal dis ibu ion ollowing
(Ba ese and Coelli, 1992). Thi d, we de i e To al Fac o P oduc i i y
(TFP) om he Solow esidual o he pooled-OLS p oduc ion unc ion
(Deininge e al., 2009; Deininge and Jin, 2005; Schmid and Sickles,
1984). To calcula e each o he abo e, we es ima e a Cobb-Douglas
p oduc ion unc ion o he ollowing gene al o m:
Yi =γ+βXim +δZin +
α
i+
ε
i (2)
whe e Yi −is he log o he alue o c op ou pu o he a m i in ime ;
Xim −is he ec o o logs o m inpu cos s used by a m i in ime ; Zin −
is he ec o o he n con ol a iables such as e e se dummy a iables
o ze o inpu use ollowing (Ba ese, 1997), linea end, and c op-yea
speci ic dummy a iables indica ing ha a gi en c op was p oduced on
he a m. Wi h he FE o SFA es ima o s, we de i e
α
i which is a
ime-in a ian , a m-speci ic ixed e ec s o ine iciencies, espec i ely.
Then we ans o m SFA ine iciencies in o echnical e iciencies (e−
α
i),
which we use as a i s p oxy o ag icul u al abili y. FE es ima es
ep esen ano he p oxy o he abili y. Finally, in he speci ica ion wi h
pooled c oss-sec ions, a m-speci ic TFP is calcula ed as TFPi =∑
ε
i /ni,
whe e ni is he numbe o yea s ha he a m is p esen in he sample,
and is used as an al e na i e p oxy o ag icul u al abili y. All hese
pa ame e s a e es ima ed along wi h he coe icien s γ,β,and δ.
5. Main esul s
5.1. Es ima ing ag icul u al abili y
Be o e es ima ing he model wi h he de e minan s o land en al, i
is in o ma i e o examine he esul s o he model wi h he p oduc ion
unc ion es ima ions (Table 2). Models (1) o (3) in Table 2 es ima e he
abili ies on a ime-in a ian subsample o a ms pooled om hei i s
yea o hei appea ance in he panel o a oid endogenei y issues in ou
main es ima ions below. Models (4)-(5) exploi he whole a ia ion o
he panel and es ima e he p oduc ion unc ions based on he whole
sample. Because we deal wi h an unbalanced panel wi h 38.6 % o he
a ms p esen du ing he whole sample (see Appendix A o de ails on
panel s uc u e), i is impo an o conside he a ms en e ing and
exi ing du ing ou pe iod. Thus, we es ima e he e iciencies o he
“en an s” a he ime o hei i s appea ance in he sample. To ensu e
he obus ness o ou esul s, we compa e ou es ima es wi h hose ob-
ained om a balanced panel (see Appendix B) which appea o be e y
simila .
All he coe icien s o he inpu cos s show he expec ed signs and a e
s a is ically signi ican ly di e en om ze o, p edic ing he ou pu wi h
ela i ely high explana o y powe . We ind land o be he mos impo -
an p oduc ion ac o wi h an elas ici y anging om 20.1 % o he OLS
o 23.1 % o FE es ima ion. These igu es a e way abo e he a e age
sha e o land en al expendi u es in o al cos s (12.6 %). Models (4)-(6)
sugges ha hese igu es inc eased u he o e ime. This indica es an
uncompe i i e alloca ion o land as we would obse e a close ma ch
Table 2
P oduc ion unc ion es ima ions.
Fi s yea Whole sample
(1) (2) (3) (4) (5) (6)
Va iables OLS FE SFA Pooled OLS FE SFA
U ilized land (ha) 0.226***
(0.021)
0.231***
(0.022)
0.233***
(0.038)
0.232***
(0.004)
0.343***
(0.006)
0.268***
(0.009)
Labo cos s 0.043***
(0.010)
0.041***
(0.010)
0.051**
(0.018)
0.067***
(0.002)
0.064***
(0.003)
0.09***
(0.003)
Seeds cos s 0.186***
(0.013)
0.187***
(0.014)
0.192***
(0.019)
0.162***
(0.003)
0.142***
(0.003)
0.156***
(0.019)
Fe ilize s cos s 0.134***
(0.007)
0.134***
(0.008)
0.132***
(0.014)
0.129***
(0.002)
0.100***
(0.002)
0.116***
(0.003)
Pe ol cos s 0.154***
(0.012)
0.148***
(0.013)
0.141***
(0.019)
0.113***
(0.003)
0.080***
(0.003)
0.086***
(0.014)
Machine y/capi al cos s 0.053***
(0.007)
0.053***
(0.007)
0.056***
(0.014)
0.050***
(0.002)
0.034***
(0.002)
0.048***
(0.003)
Se ices cos s 0.091***
(0.005)
0.092***
(0.006)
0.094***
(0.010)
0.100***
(0.002)
0.089***
(0.002)
0.107***
(0.002)
O he cos s 0.100***
(0.011)
0.100***
(0.011)
0.099***
(0.018)
0.139***
(0.002)
0.109***
(0.003)
0.127***
(0.003)
Linea end 0.102***
(0.002)
0.082***
(0.002)
0.086***
(0.002)
Cons an 1.267
(0.862)
2.313***
(0.360)
0.310***
(0.025)
0.732***
(0.055)
Scale elas ici y 0.987***
(0.0097)
0.987**
(0.0101)
0.999**
(0.0234)
0.993***
(0.0023)
0.961***
(0.0051)
0.997***
(0.0053)
N 11,710 11,710 11,710 80,245 80,245 80,245
Adj. R
2
/ % o al a iance due o ine iciency (SFA) 0.842 0.798 57.3 % 0.877 0.571 68.2 %
Mean Technical E iciency (s . de .) 0.632
(0.148)
0.7066
(0.165)
No e: p- alues a e * <0.05; ** <0.01; *** <0.001. Clus e ed, he e oscedas ici y, and au oco ela ion obus s anda d e o s (A ellano, 1987) a e p esen ed in he
pa en hesis o pooled OLS and FE speci ica ions, while egula s anda d e o s a e epo ed o he SFA. All con inuous dependen and independen a iables a e in
loga i hmic o m; inpu cos s a e exp essed in housands o cons an 2018 USD; e e se dummy a iables a e used o compensa ing ze o inpu use ollowing Ba ese
(1997). Mul iple c op-speci ic dummy a iables a e in oduced o con ol o he c op composi ion a he a m le el each yea . Model (2) u ilizes ayon ixed e ec s
because each a m is obse ed only once.
Sou ce: Au ho s’ calcula ions based on he SSSU da a.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
7
be ween hese alues o he wise (Dobbelae e and Mai esse, 2013).
In e es ingly, labo has a e y low con ibu ion o he ou pu in all h ee
speci ica ions, which is e lec ed by he sha e o labo in he cos
s uc u e. This sugges s ha business models ocusing on cash c ops’
cul i a ion hea ily ely on non-labo p oduc ion ac o s whe e land
ep esen s he mos impo an p oduc ion ac o . Finally, a ious ma-
e ial inpu s ha e a subs an ial con ibu ion o ou pu . The models
(1)-(3) es ima ed on he i s yea a e simila o he models (4)-(6)
es ima ed on he whole panel sample, sugges ing obus ness o he e-
sul s. Ob ained ag icul u al abili ies ac oss he h ee app oaches a e also
close o each o he (co ela ion coe icien s: 0.90–0.97), sugges ing a
high obus ness o he es ima ions.
Es ima ion esul s sugges ha e u ns o scale a e close o cons an .
Es ima ed scale elas ici ies a y be ween 0.988 and 1.002. Al hough in
mos o he models, we canno ejec he hypo hesis ha he scale
elas ici y is di e en om one. Scale elas ici ies in models (4) and (5) a e
signi ican ly di e en om one sugges ing minimal diminishing e u ns
o scale.
To explo e he ela ionship be ween a m size and he h ee measu es
o a a m’s abili y u he , we use Loess smoo hing depic ed in Fig. 4.
14
Ac oss he plo s based on di e en e iciency p oxies, we see a sligh
inc ease in he e iciency up o he size o 1000–1500 ha wi h a conse-
quen sligh dec ease as we mo e along he x-axis. Fo a ms la ge han
10,000 ha, he a iance is oo la ge o d aw any meaning ul conclusions
abou he a m size s. e iciency ela ionship. Because he SFA model
assumes a s ochas ic na u e o he on ie analysis, he con idence in-
e als o he g een line plo ed based on SFA es ima ions a e na owe
compa ed o he lines based on FE and pooled OLS es ima es. The e o e,
we ely on he SFA measu es o echnical e iciency as an app oxima ion
o a a m’s abili y in ou u he analysis.
We obse e subs an ial spa ial a ia ion in ou ag icul u al abili y
p oxies. The a e age echnical e iciency is 70.7 % o he possible p o-
duc ion le el (s anda d de ia ion - 16.5 %). Fig. 5 p esen s he spa ial
dis ibu ion o he e iciency sco es on he ayon le el a he beginning
and a he end o ou sample ( he cho ople h map is based on deciles o
espec i e dis ibu ions o compa abili y easons). We obse e a clus e
o high e iciency in he eas e n pa s which we e hinde ed by he
Russian mili a y in asion in 2014. Al hough we see some e iciency
imp o emen s in he sou he n egions, he cen al pa s wi h he mos
in ensi e ag icul u al p oduc ion do no show signs o imp o emen .
5.2. De e minan s o land en al
Table 3 p esen s he es ima ions o he de e minan s o land en al.
Mos impo an ly, we ind he coe icien s o he ini ial ag icul u al
abili y o be nega i e and signi ican ac oss all he speci ica ions. This
implies ha mo e e icien a ms end o en in less land in he las
obse ed pe iod, and con e sely, hose a ms ha en in mo e land
appea o be less e icien . Quan i a i ely, his means ha a a m wi h a
10 % highe ag icul u al abili y h oughou ou pe iod was likely o end
up wi h 141.1 ha less en ed land on a e age. Ou inding is p ecisely he
opposi e o wha we would expec in a well- unc ioning land en al
ma ke whe e mo e able a ms should en in mo e land. In pa icula ,
en al-based land ela ions in Uk aine up un il 2015 appea o ha e
acili a ed a low o land owa d less e icien ag icul u al p oduce s.
Tes ing ou hypo hesis ela ed o land concen a ion, we ind ha
ini ial egional land dis ibu ion ma e s o u he a ming modes.
Thus, a ms ha we e ope a ing in a ayon wi h a high deg ee o land
concen a ion in 2005 we e mo e likely o ha e en ed mo e land a he
end o he obse a ion pe iod. The e ec appea s o be ela i ely la ge,
as a 10 % inc ease in he HHI index wi h espec o a e e ence ayon is
associa ed wi h 132.6 ha mo e en ed land in he las obse ed yea . We
also ind ha he coe icien o he in e ac ion e m be ween ayon land
concen a ion and ag icul u al abili y does no di e om ze o in a
s a is ical sense.
Ini ially owned land appea s o be posi i ely associa ed wi h he land
en ed a he end o he obse a ion pe iod, con a y o ou expec a ions
and he esul s o simila s udies (e.g., Deininge and Jin, 2005; V anken
and Swinnen, 2006). Owned land was mos ypically held in he o m o
he CLS, ep esen ing negligible amoun s compa ed o he en ed a eas.
As he a m g ows, i may be be e posi ioned o a ac CLS om he
indi iduals in i s icini y. As a esul , we may obse e a complemen-
a i y be ween he CLS and en ed land.
A a m’s legal o m appea s o play a signi ican ole in i s ac i i y in
he land en al ma ke . We ind he coe icien s o he dummy o an
indi idual a m o be posi i e and s a is ically signi ican , sugges ing
ha his ype o p oduce was mo e likely o accumula e land a he end
o he obse a ion pe iod compa ed o he ag icul u al en e p ises,
which a e accoun ed o in he in e cep . We obse e an opposi e pic u e
wi h he dummy o s a e a ms, which appea s o exe a pe sis en
nega i e e ec h oughou he speci ica ions. This implies ha s a e
en e p ises end o en less land han hei p i a e coun e pa s
Ou e idence sugges s ha a m size expansion is closely associa ed
wi h he ocus on cash c ops cul i a ion. Mos o he coe icien s o he
a iables e lec ing he pe cen age o a ea cul i a ed wi h hese c ops
a e posi i ely and signi ican ly ela ed o he dependen a iable. Fo
ins ance, an addi ional 10 % o land ini ially alloca ed o co n was
associa ed wi h ca. 365.1 ha mo e en ed land a he end o he pe iod.
Simila ly, 10 % mo e land alloca ed o whea in 2005 is associa ed wi h
ca. 60.6 ha o addi ionally en ed land in 2015. This means ha ag i-
cul u al en e p ises cul i a ing c ops ha a e sui able o expo s we e
mo e likely o en in mo e land. This is in line wi h ou hypo hesis abou
annual expo -o ien ed c op cul i a ion d i ing a m expansion ia land
en al ma ke s.
5.3. De e minan s o changes in land en al
We now examine he ac o s ha may impac he changes in land
en al explici ly conside ing he ypes o a ms du ing he obse ed
pe iod. In pa icula , we di ide he sample in o subsamples acco ding o
he ollowing c i e ia: a ms ha s ayed o e he whole pe iod, exi ed,
and en e ed he land ma ke (Table 4). We use he same explana o y
a iables as in Table 3 abo e. As h ee in e ac ion e ms a e p esen in
he model, we also es ima e ma ginal e ec s o each eg esso a he
mean o he co esponding in e ac ion e ms.
Ou es ima ions do no e eal con incing e idence o en al ma ke s
acili a ing he low o land owa ds mo e e icien a me s as he co-
e icien s o he ag icul u al abili y a e insigni ican in mos o he
speci ica ions. Among he a ms ha ope a ed du ing he whole pe iod
o 2005–2015 (s aye s), mo e able ones we e mo e likely o en in mo e
land. Mo eo e , hose s aye s ha cul i a ed co n and suga bee s
appea o ha e been ypical expo -o ien ed a ms ope a ing la ge a eas
as he coe icien s o he sha es o espec i e c ops a e posi i e, signi -
ican , and la ge in magni ude. These a ms, o en es uc u ed om
collec i e a ms, inhe i ed subs an ial land en al con ac s. In con as ,
mo e e icien en an s en ed less land on a e age, sugges ing highe
p oduc ion in ensi y o di icul ies in accessing land. No ably, o he
en an s, we ind he coe icien o he in e ac ion e m be ween he
abili y and land concen a ion o be nega i e and signi ican implying
ha hey we e mo e p oduc i e in he a eas wi h highe land
concen a ion.
I a ms’ e iciency d i es hei land en al decisions, we should
obse e his in he subsamples wi h sh inking and g owing a ms in a
mo e p onounced way. We hus e-es ima ed ou models on he sub-
samples wi h a ms ha g ew and sh ank by a leas 5 % and 25 %
(Table 5). Howe e , ma ginal e ec s o ag icul u al abili y do no
demons a e signi ican e ec s. In con as , ini ial land concen a ion
appea s o exe a pola izing e ec : a ms wi h dec easing land holding
14
All he smoo he s a e based on ull-sample es ima ions wi h he excep ion
o he ini ial echnical e iciency.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
8
No e: Technical e iciency dis ibu ions a e s anda dized wi h he help o he a s anda d no mal dis ibu ion wi h a ze o mean and s anda d
de ia ion o 1. This allows compa ibili y o hese dis ibu ions as hey a e es ima ed based on di e en samples.
Appendix D. Rayon-le el es ima ions o he sha es o land eed by "exi e s"
Sha e o land unde exi e s
Coe icien s
Ag icul u al abili y ( ayon-le el mean) −1.164***
(0.345)
Ini ial land concen a ion in a ayon (HHI) −0.763*
(0.310)
Ag icul u al abili y ( ayon-le el mean) * Ini ial land concen a ion in a ayon (HHI) 0.718
(0.489)
In e cep 1.066***
(0.221)
Ma ginal e ec s
Ag icul u al abili y ( ayon-le el mean) −0.978***
(0.251)
Land concen a ion in a ayon in 2005 (HHI) −0.309***
(0.058)
Numbe o independen a iables 4
88
Numbe o obse a ions 605
No es: *p<0.1; **p<0.05; ***p<0.01. S anda d E o s a e p esen in pa en hesis. Ma ginal e ec s a e e alua ed o selec ed
a iables a means o he co esponding in e ac ion e ms. The Del a me hod is used o es ima ing he s anda d e o s.
Sou ce: Au ho s’ calcula ions based on he SSSU da a.
Re e ences
Acemoglu, D., Johnson, S., Robinson, J., 2004. Ins i u ions as he undamen al cause o
long- un g ow h (No. 10481), NBER Wo king Pape Se ies. Camb idge, MA.
A ellano, M., 1987. PRACTITIONERS ’ CORNER Compu ing Robus S anda d E o s o
Wi hin-g oups. Ex o d Bull. Econ. S a . 49, 431–434.
Balmann, A., G aubne , M., Mülle , D., Hü el, S., Sei e , S., Odening, M., Plogmann, J.,
Ri e , M., 2021. Ma ke powe in ag icul u al land ma ke s: Concep s and empi ical
challenges. Ge . J. Ag ic. Econ. 70, 213–235. h ps://doi.o g/10.30430/
gjae.2021.0117.
Ba e , C.B., Bellema e, M.F., Hou, J.Y., 2010. Reconside inG Con en ional
Explana ions O The In e se P oduc i i y-size Rela ionship. Wo ld De . 38, 88–97.
h ps://doi.o g/10.1016/j.wo ldde .2009.06.002.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
15
Ba ese, G., 1997. A no e on he es ima ion o Cobb-Douglas p oduc ion unc ions when
some explana o y a iables ha e ze o alues. J. Ag ic. Econ. 48, 250–252. h ps://
doi.o g/10.1177/109114219302100405.
Ba ese, G.E., Coelli, T.J., 1992. F on ie p oduc ion unc ions, echnical e iciency and
panel da a: Wi h applica ion o paddy a me s in India. J. P oduc . Anal. 169,
153–169.
Besley, T., 1995. P ope y igh s and in es men incen i es: Theo y and e idence om
Ghana. J. Poli . Econ. 103, 903–937. h ps://doi.o g/10.1086/262008.
Binswange , H.P., Deininge , K., 1997. Explaining Ag icul u al and Ag a ian Policies in
De eloping Coun ies. J. Econ. Li . 35, 1958–2005.
Chayano , A., 1926. The heo y o peasan coope a i es. Ohio S a e Uni e si y P ess,
Columbus.
Dasgup a, S., Knigh , T.O., Lo e, H.A., 1999. E olu ion o ag icul u al land leasing
models: a su ey o he li e a u e. Re . Ag ic. Econ. 21, 148–176.
de Jan y, A., Pla eau, J.P., Go dillo, G., Sadoule , E., 2001. Access o land and policy
e o ms. In: de Jan y, A., Go dillo, G., Pla eau, J.P., Sadoule , E. (Eds.), Access o
Land, Ru al Po e y, and Public Ac ion. Ox o d Uni e si y P ess, New Yo k,
pp. 1–26.
Deininge , K., 2003. Land policies o g ow h and po e y educ ion. A Wo ld Bank
Resea ch Repo . New Yo k.
Deininge , K., Binswange , H., 2001. The e olu ion o The Wo ld Bank’s land policy., in:
De Jan y, G. Go dillo, J.-P. Pla eau, & E.S. (Ed.), Access o Land, Ru al Po e y,
and Public Ac ion. Ox o d Uni e si y P ess, New Yo k, pp. 406–440.
Deininge , K., Bye lee, D., 2012. The ise o la ge a ms in land abundan coun ies: do
hey ha e a u u e? Wo ld De . 40, 701–714. h ps://doi.o g/10.1016/j.
wo ldde .2011.04.030.
Deininge , K., Jin, S., 2005. The po en ial o land en al ma ke s in he p ocess o
economic de elopmen : e idence om China. J. De . Econ. 78, 241–270. h ps://
doi.o g/10.1016/j.jde eco.2004.08.002.
Deininge , K., Jin, S., Naga ajan, H.K., 2009. De e minan s and consequences o land
sales ma ke pa icipa ion: panel e idence om India. Wo ld De . 37, 410–421.
h ps://doi.o g/10.1016/j.wo ldde .2008.06.004.
Deininge , K., Nizalo , D., Singh, S.K., 2018. De e minan s o p oduc i i y and s uc u al
change in a la ge comme cial a m en i onmen : E idence om Uk aine. Wo ld
Bank Econ. Re . 32, 287–306. h ps://doi.o g/10.1093/wbe /lhw063.
Dobbelae e, S., Mai esse, J., 2013. Panel da a es ima es o he p oduc ion unc ion and
p oduc and labo ma ke impe ec ions. J. Appl. Econom. 28, 1–46. h ps://doi.o g/
10.1002/jae.
Eas wood, R., Lip on, M., Newell, A., 2010. Fa m size, in: Pingali, P., E enson, R.E.
(Eds.), Handbook o Ag icul u al Economics. No h-Holland, Ams e dam, pp.
3323–3397.
Eswa an, M., Ko wal, A., 1986. Access o capi al and ag a ian p oduc ion o ganiza ion.
Econ. J. 96, 482–498.
Fede , G., 1985. The ela ion be ween a m size and a m p oduc i i y. J. De . Econ. 18,
297–313.
G aubne , M., Os apchuk, I., Gagalyuk, T., 2021. Ag oholdings and land en al ma ke s: a
spa ial compe i ion pe spec i e. Eu . Re . Ag ic. Econ. 48, 158–206.
Huy, H.T., Nguyen, T.T., 2019. C opland en al ma ke and a m echnical e iciency in
u al Vie nam. Land Use Policy 81, 408–423. h ps://doi.o g/10.1016/j.
landusepol.2018.11.007.
Keyze , M.A., Me bis, M.D., Halsema, A., Heye s, V., Bo odina, O., P okopa, I., 2017.
Unlocking Uk aine’s p oduc ion po en ial. In: Paloma, S.G.Y., Ma y, S., Lang ell, S.,
Ciaian, P. (Eds.), The Eu asian Whea Bel and Food Secu i y. Sp inge , Swi ze land,
pp. 141–155.
Koi ala, K.H., Mish a, A., Mohan y, S., 2016. Impac o land owne ship on p oduc i i y
and e iciency o ice a me s: he case o he Philippines. Land Use Policy 50,
371–378. h ps://doi.o g/10.1016/j.landusepol.2015.10.001.
Kuns, B., 2017. Beyond Coping: Smallholde In ensi ica ion in Sou he n Uk aine. Sociol.
Ru al. 57, 481–506. h ps://doi.o g/10.1111/so u.12123.
K a iuk, V., He z eld, T., 2019. Wel a e e ec s o land ma ke libe aliza ion scena ios in
Uk aine: E idence-based economic pe spec i e (No. 186), IAMO Discussion Pape s.
Halle (Saale).
K a iuk, V., Pe ick, M., 2021. Libe al land e o m in Kazakhs an? The e ec on land
en al and c edi ma ke s. Wo ld De . 138, 105285.
Lackman, C., 1977. The mode n de elopmen o classical en heo y. Am. J. Econ.
Sociol. Inc. 36, 51–63.
Le man, Z., Csaki, C., Fede , G., 2004. Ag icul u e in ansi ion: Land policies and
e ol ing a m s uc u es in pos -So ie coun ies. Lexing on Books, Lanham, MD.
Le man, Z., Sedik, D.J., Pugacho , N., Goncha uk, A., 2007. Re hinking ag icul u al
e o m in Uk aine. Halle (Saale).
Lip on, M., 2009. Land e o m in de eloping coun ies. P ope y igh s and p ope y
w ongs. Rou ledge, London.
Luo, B., 2018. 40-yea e o m o a mland ins i u ion in China: a ge , e o and he
u u e. China Ag ic. Econ. Re . 10, 16–35.
M. Ha igsen Land e o m and land consolida ion in Cen al and Eas e nEu ope a e
1989: Expe iences and pe spec i es 2015 doi: 10.5278/ bn.phd.engsci.00019.
Mamono a, N., 2015. Resis ance o adap a ion? Uk ainian peasan s’ esponses o la ge-
scale land acquisi ions. J. Peasan S ud. 42, 607–634. h ps://doi.o g/10.1080/
03066150.2014.993320.
Ma inelli, P., 2014. La i undia e isi ed: Ma ke powe , land inequali y and ag icul u al
e iciency. E idence om in e wa I alian ag icul u e. Explo . Econ. His . 54,
79–106. h ps://doi.o g/10.1016/j.eeh.2014.05.003.
Min, S., Waibel, H., Huang, J., 2017. Smallholde pa icipa ion in he land en al ma ke
in a moun ainous egion o Sou he n China: Impac o popula ion aging, land enu e
secu i y and e hnici y. Land Use Policy 68, 625–637. h ps://doi.o g/10.1016/j.
landusepol.2017.08.033.
Ni ye skyy, O., Nizalo , D., 2016. Economic e u n o a mland in Uk aine and i s
incidence. Vox Uk . Ap il.
Ni ye skyy, O., 2019. Black gold. Why ag icul u alis s ea n mo e and he u al a eas a e
dying? Vox Uk . Janua y..
Pe ick, M., 2015. Compe i ion o land and labo among indi idual a ms and
ag icul u al en e p ises: e idence om Kazakhs an’s g ain egion, in: Kimhi, A.,
Le man, Z. (Eds.), Ag icul u al T ansi ion in Pos -So ie Eu ope and Cen al Asia
a e 25 Yea s. IAMO, Halle (Saale), pp. 117–139.
Pe ick, M., Wandel, J., Ka s en, K., 2011. Fa m es uc u ing and ag icul u al eco e y
in Kazakhs an’s g ain egion: An upda e (No. 137), IAlMO Discussion Pape s. Halle
(Saale).
Pingali, P., 2007. Ag icul u al mechaniza ion: Adop ion pa e ns and economic impac .
In: E enson, R., Pingali, P. (Eds.), Handbook o Ag icul u al Economics. FAO, Rome,
I aly, pp. 2779–2805. h ps://doi.o g/10.1016/S1574-0072(06)03054-4.
Place, F., 2009. Land enu e and ag icul u al p oduc i i y in A ica: A compa a i e
analysis o he economics li e a u e and ecen policy s a egies and e o ms. Wo ld
De . 37, 1326–1336. h ps://doi.o g/10.1016/j.wo ldde .2008.08.020.
Qiu, T., Choy, B., Li, S., He, Q., Luo, B., 2020. Does land en ing-in educe g ain
p oduc ion? E idence om u al China. Land Use Policy 90, 104311. h ps://doi.
o g/10.1016/j.landusepol.2019.104311.
Rada, N.E., Fuglie, K.O., 2019. New pe spec i es on a m size and p oduc i i y. Food
Policy 147–152. h ps://doi.o g/10.1016/j. oodpol.2018.03.015.
Rahman, S., 2010. De e minan s o ag icul u al land en al ma ke ansac ions in
Bangladesh. Land Use Policy 27, 957–964. h ps://doi.o g/10.1016/j.
landusepol.2009.12.009.
Sadoule , E., Mu gai, R., Jan y, A.De, 2001. Access o land ia land en al ma ke s. In:
de Jan y, A., Go dillo, G., Pla eau, J.P., Sadoule , E. (Eds.), Access o Land, Ru al
Po e y, and Public Ac ion. Ox o d Uni e si y P ess, New Yo k, pp. 196–229.
Schmid , P., Sickles, R.C., 1984. P oduc ion on ie s and panel da a. 1. J. Bus. Econ. S a .
2, 367–374.
Sen, A., 1966. Peasan s and dualism wi h o wi hou su plus labo . J. Poli . Econ. 74,
425–450.
SSSU, 2017. Ag icul u e o Uk aine. S a is ical Yea book, Kyi .
Swinnen, J., V anken, L., 2007. Pa e ns o land ma ke de elopmen s in ansi ion (No.
179), LICOS Discussion Pape .
Tan, S. hao, Zhang, R. xin, Tan, Z. chun, 2018. G assland en al ma ke s and he de
echnical e iciency: abili y e ec o esou ce equilib a ion e ec ? Land Use Policy
77, 135–142. h ps://doi.o g/10.1016/j.landusepol.2018.05.030.
Uzun, V., Le man, Z., 2017. Ou comes o Ag a ian Re o m in Russia. In: Paloma, S.,
Ma y, S., Lang ell, S., Ciaian, P. (Eds.), The Eu asian Whea Bel and Food Secu i y.
Sp inge In e na ional Publishing, Swi ze land, pp. 81–102.
V anken, L., Swinnen, J., 2006. Land en al ma ke s in ansi ion: heo y and e idence
om Hunga y. Wo ld De . 34, 481–500. h ps://doi.o g/10.1016/j.
wo ldde .2005.07.017.
V. K a iuk e al.
Land Use Policy 146 (2024) 107336
16
