Yin, Fang; Sun, Zhanli; You, Liangzhi; Mülle , Daniel
A icle — Published Ve sion
De e minan s o changes in ha es ed a ea and yields o
majo c ops in China
Food Secu i y
P o ided in Coope a ion wi h:
Leibniz Ins i u e o Ag icul u al De elopmen in T ansi ion Economies (IAMO), Halle (Saale)
Sugges ed Ci a ion: Yin, Fang; Sun, Zhanli; You, Liangzhi; Mülle , Daniel (2024) : De e minan s o
changes in ha es ed a ea and yields o majo c ops in China, Food Secu i y, ISSN 1876-4525,
Sp inge Na u e, Be lin, Vol. 16, Iss. 2, pp. 339-351,
h ps://doi.o g/10.1007/s12571-023-01424-x ,
h ps://link.sp inge .com/a icle/10.1007/s12571-023-01424-x
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/289498
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0
Vol.:(0123456789)
Food Secu i y (2024) 16:339–351
h ps://doi.o g/10.1007/s12571-023-01424-x
ORIGINAL PAPER
De e minan s o changes inha es ed a ea andyields o majo c ops
inChina
FangYin1,2 · ZhanliSun2 · LiangzhiYou3,4 · DanielMülle 2,5,6
Recei ed: 2 Ma ch 2023 / Accep ed: 27 No embe 2023 / Published online: 25 Janua y 2024
© The Au ho (s) 2024
Abs ac
Global ag icul u al p oduc ion has isen subs an ially in ecen decades and needs o ise u he o mee he e e -g owing ood
demand. While highe p oduc ion can be di ec ly a ibu ed o ag icul u al expansion and in ensi ica ion, he unde lying ac o s
behind he changes in cul i a ed a eas and yields can be complica ed and ha e no been well unde s ood. China has d ama ically
inc eased i s ood p oduc ion in pas decades, especially du ing he ini ial app oxima ely 30yea s ollowing he commencemen
o he u al e o m in he la e 1970s. The ag icul u al land use, including c opland a eas, he composi ion o di e en c ops and
hei spa ial dis ibu ions, and c op yields ha e expe ienced subs an ial changes. In his esea ch, we quan i a i ely analysed he
changes in he ha es ed a eas and yields o he ou mos widely cul i a ed c ops in China ( ice, whea , maize, and soybean) a
he coun y le el om 1980 o 2011. We used spa ial panel eg essions o quan i y he de e minan s o he obse ed changes in
ha es ed a ea and yields o he majo cul i a ion egion o each o he ou c ops. Resul s showed ha g ow h in popula ion, g oss
domes ic p oduc , and u banisa ion a e posi i ely associa ed wi h ha es ed a eas. Highe usage o machine y and e ilise inpu s
inc eased yields o he h ee ce eal c ops, while he ha es ed a ea o soybean dec eased, pa icula ly a e China’s accession o he
WTO. Ou indings e eal how domes ic u banisa ion and changes in consump ion pa e ns, coupled wi h he ising globalisa ion
o ag icul u al ma ke s, shaped China’s ag icul u al p oduc ion and land use o e he h ee decades. These insigh s shed ligh on
he de e minan s o long- e m ag icul u al dynamics and hus in o m e idence-based decision-making.
Keywo ds Ag icul u al p oduc ion· Land-use in ensi y· C op p oduc i i y· Land-use change· Food secu i y· Spa ial
panel eg ession
1 In oduc ion
Food secu i y con inues o be a majo conce n o human-
i y and is an in insic elemen o sus ainable de elopmen
(God ay, Bedding on e al.,2010). Fu u e ag icul u al
p oduc ion mus inc ease o mee he e e -inc easing
ood demand due o popula ion g ow h, highe demands
o plan -based ene gy p oduc ion, and mo e esou ce-
demanding die s. While he inc eased ood p oduc ion
in he pas , pa icula ly a e he g een e olu ion s a ed
* Zhanli Sun
[email p o ec ed]
Fang Yin
[email p o ec ed]
Liangzhi You
l.you@cgia .o g
Daniel Mülle
[email p o ec ed]
1 Shandong Academy o Ag icul u al Sciences, Gongyebei
Road 202, Jinan250100, China
2 Leibniz Ins i u e o Ag icul u al De elopmen
inT ansi ion Economies (IAMO), Theodo -Liese -S . 2,
06120Halle(Saale), Ge many
3 In e na ional Food Policy Resea ch Ins i u e (IFPRI), 1201
Eye S ee , NW, Washing on, DC20005, USA
4 Mac o Ag icul u e Resea ch Ins i u e, College o Economics
andManagemen , Huazhong Ag icul u al Uni e si y,
Wuhan430070, Hubei, China
5 Geog aphy Depa men , Humbold -Uni e si ä Zu Be lin,
Un e Den Linden 6, 10099Be lin, Ge many
6 In eg a i e Resea ch Ins i u e On T ans o ma ions
o Human-En i onmen Sys ems (IRI THESys),
Humbold -Uni e si ä Zu Be lin, Un e Den Linden 6,
10099Be lin, Ge many
340 F.Yin e al.
in he 1950s, was mainly achie ed by ag icul u al land
expansion and in ensi ica ion, u u e p oduc ion inc eases,
howe e , mus be achie ed a lowe en i onmen al cos s
(Alexande e al., 2015; Ga ne e al., 2013; Popp e al.,
2016). A ew c ops play a pa icula ly impo an ole in
ood secu i y as hey p o ide he majo i y o he ene gy
and essen ial nu ien s (Khou y e al., 2014). O e all,
humani y ob ains 50% o i s daily calo ies om ce eals,
and mo e han 40% o hese calo ies a e om only h ee
majo s aple c ops: ice, whea , and maize (FAO, 2019;
Kea ney, 2010). Mo eo e , he g owing consump ion o
li es ock p oduc s, which a e inc easingly p oduced in
indus ial p oduc ion sys ems, equi es la ge amoun s o
eed and odde , which a e mainly sou ced om maize,
as he sou ce o ene gy, and soybean, which p o ides he
p o eins o animal g ow h (Cassidy e al., 2013). In 2017,
hal o he 1.42 billion hec a es (Bha) o he global ha -
es ed a ea we e cul i a ed wi h only ou c ops: whea ,
maize, ice, and soybean (FAO, 2019). Unde s anding
changes in he a ea dedica ed o hese majo c ops and
hei yields p o ides c i ical insigh s o be e unde -
s anding he ag icul u al dynamics ha shape he salien
changes in he ood sys em.
While he o e all sha e o he ou majo c ops in he
global ha es ed a ea and p oduc ion olume has emained
s able o e he las 50yea s, he con ibu ion o he indi-
idual c ops o he o e all p oduc ion quan i ies has sub-
s an ially changed o e ime. The p opo ion o whea in
he o al ha es ed a ea dec eased om 22% in 1967 o
15% in 2017. A he same ime, he a ea occupied by maize
inc eased om 11% in 1967 o 14% in 2017, equi alen o
an absolu e inc ease o 85 million ha (Mha). The inc ease
in soybean cul i a ion has been especially d as ic, wi h an
inc ease g ea e han 95 Mha, o om 3 o 9% o he global
ha es ed a ea du ing his 50-yea pe iod (FAO, 2019).
The sweeping changes in he global pa e ns o c op
cul i a ion ha e been d i en by in e linked poli ical, socio-
economic, clima e change, and biophysical ac o s. Whe e
hese c ops a e p oduced depends on loca ional ac o s
ha shape land en s, including clima e, soil, and acces-
sibili y, while changes in economic, ins i u ional, poli i-
cal, and demog aphic cha ac e is ics d i e he changes in
cul i a ion pa e ns (Mey oid , 2016). Fo example, he
spa ial pa e ns o ood p oduc ion and consump ion ha e
been ans o med by he globalisa ion o he ood sys em,
mani es ed by he shi om locally p oduced ood o an
inc easing eliance on ag icul u al commodi ies ha a e
sou ced om dis an ma ke s (Le e s & Mülle , 2019).
While he o e all c opland a ea only inc eased mode -
a ely, much o he ecen global inc ease in c op p oduc ion
has been due o highe c op yields, mainly as a esul o
highe inpu in ensi y pe uni a ea (Magliocca e al., 2015;
Rudel e al., 2009). The in ensi ica ion o p oduc ion has
g ea ly bene i ed global ood secu i y because i has sa ed
subs an ial land esou ces om being con e ed in o ag i-
cul u al p oduc ion (Bo laug, 2007; Bu ney e al., 2010).
Howe e , he in ensi ica ion may also p o oke ebound
e ec s by aising p o i s om p oduc ion and lowe ing ood
p ices, he eby incen i ising u he expansion (Lambin &
Mey oid , 2011; Rudel e al., 2009). Unde s anding he pa -
e ns and de e minan s o in ensi ica ion p ocesses emains
impo an because o he ad e se side e ec s ha can esul
om highe inpu in ensi y, such as nu ien leaching, wa e
pollu ion, ai pollu ion, and nega i e e ec s on human heal h
(God ay, C u e e al., 2010; Tilman e al., 2011).
China is a pa icula ly in e es ing case because he coun-
y’s apid economic de elopmen has had subs an ial e ec s
on land use (Deng e al., 2015; Jiang e al., 2013; Sun e al.,
2018). China has expe ienced d as ic changes in ag icul u al
p oduc ion and land sys ems in he pas yea s. Ag icul u al
p oduc ion has inc eased d ama ically in China ollow-
ing he e o ms ini ia ed in 1978 ha shi ed a mland use
igh s om communes o millions o a m households—
he so-called household esponsibili y sys em (Huang &
Rozelle,2018; Lin e al., 2022). Ag icul u al p oduc ion has
aken o e e since. China’s eal ag icul u al ou pu alue
g ew a an annual a e o 5.3 pe cen be ween 1978 and 2017
(Sheng e al., 2020).
A p esen , China is by a he la ges p oduce and con-
sume o ice and whea in he wo ld; he go e nmen iews
sel -su iciency in he p oduc ion o hese c ops as c i i-
cal o secu ing China’s “ ice bowl” (i.e., p oducing su i-
cien ood o China) (Zhang, 2019). Main aining domes ic
ood secu i y, de ined by he Chinese go e nmen as a 95%
deg ee o g ain sel -su iciency, emains a op policy p io -
i y and is a s a egic goal on China’s ood secu i y agenda
(Huang & Yang, 2017). The linge ing COVID pandemic and
ongoing Russia-Uk ain con lic ha e pushed ood secu i y
e en highe on he Chinese go e nmen ’s agenda (Hellege s,
2022; Pu & Zhong, 2020).
To achie e he policy goals, he Chinese go e nmen has
implemen ed a s ic policy o p o ec i s a able land, he
so-called “a able land ed line policy”, which aims o main-
ain a leas 1.8 billion mu (i.e., 120 Mha) o ag icul u al
land in p oduc ion. In addi ion, China se up a gua an eed
g ain p ocu emen p ice o whea , ice, and maize, among
o he s, o os e g ain p oduc ion. While China become he
la ges ood impo e , pa icula ly o soybeans and maize,
in he wo ld, he key s aple c ops, whea and ice, con inue
o be la gely p oduced domes ically, mainly o educe eli-
ance on impo s and o gua an ee domes ic ood secu i y
(Huang e al., 2017). Howe e , China aces daun ing chal-
lenges in s i ing o he en isaged domes ic ood secu i y.
The domes ic demand o ag icul u al p oduc s has been
apidly inc easing, mainly because o popula ion g ow h,
and a die a y s uc u al shi om mainly plan -based ood
341
De e minan s o ha es ed a ea and yields in China
owa ds die s wi h a highe eliance on animal p o eins—
d i en by u banisa ion and g owing a luence (Huang e al.,
2015; Jiang e al., 2015; Zhao e al., 2021).
Achie ing p oduc ion inc eases is hampe ed by he
small a m s uc u e— he a e age a m size in China was
app oxima ely 0.5ha in 2015 (Wu e al., 2018) and he high
agmen a ion o a ms, which hinde s highe capi al inpu s,
ag icul u al mode nisa ion, and ealisa ion o economies o
scale (Lai e al., 2020; Zhang e al., 2013). C opland pe
capi a amoun ed o only 0.09ha pe capi a in 2016 in China,
and he sca ci y o land esou ces becomes e en mo e se e e
wi h he loss o a able land caused by u banisa ion, land
deg ada ion, and soil con amina ion (B en d’Amou e al.,
2016; Deng & Li, 2016). The con as be ween he low
income om ag icul u e and he ising wage le els om
u ban employmen oppo uni ies d i es he massi e mig a-
ion om u al o u ban a eas. Since he 2000s, many u al
a eas ha e s a ed o depopula e and a e su e ing om
labou sho ages because o he e e -inc easing mig a ion
om u al a eas o he g owing ci ies (Liu e al., 2017). The
aging socie y, pa ially due o he p e ious one-child policy,
u he augmen s u al labou sca ci y and aises he ques-
ion o who will cul i a e China’s ag icul u al land in he
u u e. The undamen al changes in he Chinese coun yside
ha e p o ound e ec s on he ex en and inpu in ensi y o
ag icul u e and, consequen ly, on ag icul u al p oduc ion
ou pu s. I is i al o unde s and he impac o hese changes
on ag icul u al p oduc ion s a egies and, hus, on ag icul-
u al p oduc i i y.
The spa ial composi ion o c op cul i a ion in China has
changed no ably in he las h ee decades. Rice, a guably he
mos impo an s aple ood o he Chinese, has adi ionally
been cul i a ed in sou he n China. Howe e , in ecen yea s,
ice cul i a ion has expanded in No heas China because
he japonica ice g own in he no h is in high demand due
o i s supe io nu i ional alue and good as e (Sun e al.,
2018). In con as , he ha es ed a ea o ice in he sou h
has dec eased due o mul iple ac o s such as u banisa ion,
c opland abandonmen , and educed cul i a ion in ensi y
e lec ed in he mul i-c opping index (Jin & Zhong, 2022;
Liu e al., 2013). Combined wi h clima e change, he cen-
oid o he ice plan a ion a ea has al eady shi ed 230km
o he no heas (Hu e al., 2019; Li e al., 2015; Liu e al.,
2013). Whea is he main s aple c op in no he n China.
App oxima ely 126 million me ic ons o whea we e p o-
duced on 24 Mha in 2011, and mos o his whea was cul i-
a ed ex ensi ely and o a ed wi h maize. Mo eo e , China
has become he second-la ges maize p oduce in he wo ld.
The ha es ed a ea o maize inc eased om 20 Mha in 1980
o 36 Mha in 2011, mainly in esponse o he inc easing
demand o maize as a eed c op o China’s g owing li e-
s ock popula ion. Maize p oduc ion is concen a ed in he
plain egions s e ching om he no heas o he sou hwes
(Li, 2009; Yin e al., 2018). O e all, ice, whea , and maize
accoun o 80% o he o al ha es ed a ea in China, wi h an
inc easing end. Con e sely, he a ea cul i a ed wi h soy-
beans dec eased sligh ly, and yields ha e emained s able
since app oxima ely 1980 (Sun e al., 2018).
The changes in ha es ed a ea, c op s uc u es, and yields
ha e s a k implica ions o ood secu i y and he en i onmen .
Howe e , he exis ing li e a u e on he pa e ns, de e minan s,
and d i e s o changes in c opland s uc u es in China has
ocused on indi idual c ops, such as ice (Hu e al., 2019; Jin
& Zhong, 2022; Li e al., 2015; You, 2012; Yu e al., 2022)
o maize (Li, 2009). To he bes o ou knowledge, a holis ic
assessmen o he spa ial changes in c opping s uc u es and
p oduc i i y, how hese ha e changed in all o China, and wha
ac o s ha e de e mined hese changes is s ill lacking. He e we
analyse he de e minan s o he changes in ha es ed a eas and
yields o ou majo c ops ( ice, whea , maize, and soybean)
a he coun y le el wi h spa ial panel eg essions ha accoun
o spa ial and se ial au oco ela ions in he da a. We aim o
answe wo key esea ch ques ions:
1. How did he ha es ed a eas and yields o he ou majo
c ops change be ween 1980 and 2011?
2. Wha we e he main de e minan s o he changes in he
ha es ed a ea and yield o each o hese c ops?
2 Da a
We u ilised spa ial panel da a om 2,354 coun ies and
om each yea om 1980 o 2011— he i s h ee dec-
ades a e he u al e o m when he mos d as ic changes in
c opland a ea and c op yields happened. The u al e o m
o China s a ed om a g ass oo s ini ia i e in Fengyang
Coun y, Anhui p o ince, in 1978 and soon sp ead o o he
coun ies in Anhui and Sichuan p o inces. Bu only in 1980,
he household esponsibili y sys em (HRS) was o icially
endo sed by he cen al go e nmen and implemen ed
na ionwide (Lin, 1988). The e o e, we pu posely chose his
pe iod when he sweeping changes in ag icul u al p oduc ion
happened, mainly d i en by ins i u ional e o ms. All da a
we e sou ced om he s a is ical yea books o he Chinese
go e nmen . The panel se up allows o con ol o a iables
ha canno be obse ed o measu ed, such as cul u al ac o s
o di e ences in ag icul u al p ac ices ac oss obse a ions,
and o con ol o a iables ha change o e ime bu no
ac oss obse a ions (Hsiao, 2007). We ocused ou analysis
on he ou majo c ops, i.e., ice, whea , maize, and soy-
bean, which co e ed 90% o he cul i a ed a eas and 92% o
he g ain p oduc ion quan i y in China in 2011. In o al, we
es ima ed eigh c op-speci ic eg essions wi h he annual
ha es ed a eas o each o he ou c ops and hei yields as
he esponse a iables.
342 F.Yin e al.
C op p oduc ion in China is spa ially clus e ed in speci ic
egions (Sheng e al., 2017; Yin e al., 2018). We con ined
he ou c op-speci ic eg essions o he main cul i a ion
egion o each c op. To de ine hese egions, we selec ed he
p o inces wi h he la ges ha es ed a ea in 2011 in descend-
ing o de un il mo e han 90% o each c op’s ha es ed a ea
was included. The main cul i a ion egions isualise he
mos impo an cen es o p oduc ion o each c op (Fig.1).
Rice clus e s in No heas and Sou h China; whea is mos ly
loca ed in he no he n pa ; maize domina es in a bel om
he no heas o sou hwes ; soybean is concen a ed in he
no heas . In con as o he ha es ed a eas o each o he
ou c ops, he c op yields did no show ob ious spa ial clus-
e s (Supplemen a y Fig.S1).
The choice o he explana o y a iables was based on a
ho ough li e a u e e iew and p io knowledge abou land
use in China (Shi e al., 2013; Tong e al.,2003; Yu e al.,
2016) bu was also cons ained by da a a ailabili y. The ha -
es ed a ea and yields o he majo c ops can be in luenced
by many ac o s, including poli ical, socioeconomic, man-
agemen , echnological, and biophysical ac o s. We es ed
he model wi h di e en combina ions o he a iables o
a ain he mos plausible and gene alizable model. The socio-
economic a iables, which we selec ed om he Chinese s a-
is ical yea books, include he g oss domes ic p oduc (GDP)
pe coun y as a p oxy o economic pe o mance, oad leng h
pe coun y o cap u e ma ke accessibili y, and he popula-
ion pe coun y (in housands) o measu e he local demand
o ag icul u al p oduc s, which may a ec he ex en and
pa e ns o cul i a ion. The GDP alue was adjus ed o
in la ion by employing he consume s' p ice index (CPI)
wi h he e e ence yea se as 2010. We hypo hesized ha
he GDP alues in p eceding yea s migh impac he ha es
a ea and c op yields. Consequen ly, we used lagged GDP,
speci ically GDP( -1), in he model. We accoun ed o he
ag icul u al labou inpu (in housands), he ho sepowe o
machine y (in 1000kw) used in ag icul u al p oduc ion, and
he use o e ilise in ag icul u al p oduc ion o ep esen
Rice Whea
Maize Soybean
Fig. 1 Main cul i a ion egion o each c op ( hick black ou line). These egions ha bou ed mo e han 90% o he en i e ha es ed a ea o each
c op in 2011
343
De e minan s o ha es ed a ea and yields in China
land-use in ensi y. We u he employed wo ime- a ian
biophysical ac o s ha we hypo hesised o be impo an
spa ial de e minan s o he loca ion o c op cul i a ion: he
g owing deg ee days as he accumula ed empe a u e o e
10 deg ees and he o al ain all o e e y yea (De yng e al.,
2014; Hu e al., 2019). Finally, we included a dummy a i-
able ha cap u es he admission o China o he Wo ld T ade
O ganiza ion (WTO) in 2001 as a po en ially impo an a i-
able a ec ing he amoun o c op impo s, which may de e -
mine he domes ic pa e ns o ag icul u al p oduc ion.
All explana o y a iables om he s a is ical yea books
a e a ailable a he coun y le el o e e y yea om 1980 o
2011; we esampled o agg ega ed he biophysical a iables
o he coun y le el. As he s udy a eas o each c op di e
om each o he , he s a is ics o he explana o y a iables
also di e o each c op. While he applica ion o capi al
inpu s o ag icul u e (i.e., machine y, e ilise ), as well
as GDP and oad leng h, ha e inc eased subs an ially in
he main cul i a ion egions o all c ops, labou inpu has
dec eased (Fig.2). The popula ion inc eased by 35% o e
he s udy pe iold, bu ag icul u al labou dec eased by 20%.
As expec ed, ain all and g owing deg ee days luc ua ed
o e ime, wi h a sligh ly inc easing end in he numbe o
g owing deg ee days.
3 Me hods
Gi en he spa ial dynamic na u e o he changes in he c op-
land a ea and yields, we chose he spa ial panel eg ession
model, a s a e-o - he-a s a is ical model, in his esea ch.
Reg ession models mus co ec o he p esence o spa ial
au oco ela ion in he dependen a iable because i io-
la es he s anda d assump ion o independen obse a ions
in eg ession analysis, simila o se ial co ela ions in ime
se ies da a. To es o he p esence o spa ial au oco e-
la ion in he dependen a iables, we calcula ed Mo an’s I
sepa a ely o ha es ed a eas and yields o all cul i a ion
egions o he ou c ops. We consis en ly ound s a is i-
cally signi ican spa ial clus e ing o he ha es ed a eas
and yields, implying he exis ence o posi i e au oco ela-
ion. We co ec o he au oco ela ions o e ime and space
wi h spa ial panel models (Belo i e al., 2017; Elho s , 2010,
2012). We es ed di e en models ha con olled o ime
lags, spa ial lags, spa ial e o s, o a combina ion he eo .
To selec he app op ia e model, we used he Akaike in o -
ma ion c i e ia (AIC). Finally, we used he Hausman es o
decide be ween andom and ixed e ec o mula ions. The
esul s sugges ed he ixed e ec s o mula ion as app op i-
a e and, in consequence, all ime-in a ian a iables cancel
ou o he eg essions. The spa ial au o eg essi e model is
as ollows:
W
is an
n×n
(n is he numbe o he spa ial obse a ions, in
ou case, coun ies) spa ial weigh s ma ix ha desc ibes he
spa ial neighbo ing ela ionship:
wij
akes a posi i e alue i
coun y i is a neighbo o coun y j, o he wise 0.
wij
is he (
i,j
) h elemen o W, whe e
iandj=(1, …,n)
. The spa ial lag
e m
Wy
is he spa ially weigh ed a e age o he alue o y
in he neighbou ing loca ions, i.e., coun ies.
𝜌
is he spa ial
au o eg essi e coe icien , and
ε
is he e o e m, and
X
he
(1)
y =𝜌Wy +𝛽X+𝜏 +𝜀
Fig. 2 T ends o coun y-le el
explana o y a iables in he
main cul i a ion egions o he
majo c ops
344 F.Yin e al.
ec o o he explana o y a iables. To cap u e he in lu-
ence o agg ega e empo al ends (e.g., he echnological
p og ess), we included he yea alues in he eg ession wi h
τ
deno ing he ime coe icien .
The spa ial weigh s ma ix,
W
, accoun o he spa ial
au oco ela ion. The e a e no uni e sal ules o he choice
o he neighbou hood s uc u e, size, and weigh o indi-
idual neighbou s. We, he e o e, es ed se e al ealisa ions
o he spa ial weigh s ma ix o assess he sensi i i y o he
esul s o he choice o he neighbou hood s uc u e. He e
we epo ed he esul s wi h he i s -o de ook con igui y
weigh s, which include immedia e neighbou s ha sha e a
common bo de wi h he obse a ion o in e es . The imple-
men a ions o he second-o de weigh s (i.e., he inclusion
o neighbo s o i s -o de neighbo s) and queen con igui y
(i.e., sha ed bo de and e exes) had mino e ec s on he
esul s and do no a ec hei in e p e a ion ( hese esul s can
be ob ained om he i s au ho upon eques ). We es ed o
mul icollinea i y among explana o y a iables wi h he a i-
ance in la ion ac o (VIF). As a ule o humb, a a iable
wi h a VIF g ea e han 10 may me i u he in es iga ion
(Miles, 2014). No majo co ela i e s uc u es o conce n
occu ed in ou da a, as judged by he VIFs.
Finally, we assessed i ou models su e om Simpson’s
pa adox, ha is, he coe icien ha ela es he explana o y
(x) o he esponse a iables (y) changes sign i ano he a i-
able is added o he model (Pea l, 2014). This allows es ing
he obus ness o he models o addi ional model o mula-
ions by s a ing wi h one a iable o in e es , hen consecu-
i ely adding o he a iables un il a i ing a he ull model
speci ica ion. We did no ind any change in he a iables’
sign when including addi ional co a ia es in he models.
4 Resul s
We p esen all eg ession esul s in log–log o m so ha we
can in e p e he a iable e ec s as elas ici ies, exp essed as a
pe cen change. Fo example, an elas ici y o wo implies ha
a 1% inc ease in he independen a iable would esul in a
2% inc ease in he dependen a iables. In addi ion o compa -
ing he size o he in luence o each coe icien o all c ops,
we also calcula ed he s anda dised e ec sizes ha acili a e
compa ing he s eng hs o in luence ac oss he explana o y
a iables, i espec i e o hei measu emen uni s.
4.1 Changes inha es ed a ea andyield
The a e age ha es ed a ea o maize and soybean inc eased
by 78% and 7%, espec i ely, while he a eas o ice and
whea declined by 14% and 9%, espec i ely, especially
be ween 1998 and 2004 (Fig.3; no e ha we p esen he
changes in he main cul i a ion egions; hence, he num-
be s may di e om he o icial s a is ics). The decline o
ice and whea du ing his pe iod e lec ed he c op plan -
ing s uc u e shi : many egions, pa icula ly in coas al
and wes e n China adjus ed he c op s uc u e by educing
he p opo ion o g ain c ops, such as whea and ice, and
inc easing cash c ops and ege ables (Liu e al., 2013). The
s eady and sha p inc ease o maize, especially a e yea
2000, e lec s he inc eased demand o animal eed due o
he die a y change(Wang e al., 2019; Zhao e al., 2021).
Go e nmen subsidies on maize also play a subs an ial
ole(Huang e al., 2013). The yields o ice, whea , maize,
and soybean inc eased by 40%, 46%, 61%, and 26% om
1980 o 2011, espec i ely. (A e c oss- e e encing he al-
ues wi h o he da a sou ces, we conclude ha he spikes in
maize and whea yields in 1983 a e likely a da a a i ac ;
his ou lie will no subs an ially a ec he esul s because
we ha e a long ime se ies).
The a ea ha es ed wi h ice inc eased in he no heas
and declined in sou he n China (maps wi h he spa ial
changes in ha es ed a eas om 1980 o 2011 a e in Sup-
plemen a y Fig.S2 and yield changes in Supplemen a y
Fig.S3). The main a eas o whea cul i a ion we e in no h-
e n China, al hough almos e e y p o ince has some whea
cul i a ion. The spa ial pa e ns o changes in whea cul i a-
ion om 1980 o 2011 show ew ob ious spa ial pa e ns
excep he inc easing spa ial clus e ing a ound he No h
China Plain, he adi ional whea cul i a ion egion. Maize
Fig. 3 Changes in ha es ed
a eas (le ) and yields ( igh ) pe
c op ype in he main cul i a ion
egions
345
De e minan s o ha es ed a ea and yields in China
cul i a ion co e ed la ge a eas s e ching om he no heas
o he sou hwes ; his a ea is known as he Chinese maize
bel (Meng e al., 2016). Soybean was inc easingly concen-
a ed in he no heas .
C op yields showed a much mo e he e ogeneous dis ibu-
ion han he a ea ha es ed. The yields o ice, whea , and
maize inc eased in mos egions in China, while he yield
o soybean dec eased, especially be ween 1980 and 1990.
4.2 De e minan s o hechanges
The spa ial lag o he dependen a iables has s ong and
posi i e e ec s, which mi o he spa ial concen a ion o
he ha es ed a eas (see Supplemen a y TableS4 o de ailed
eg ession esul s).
The impac s o ime- a ying explana o y a iables quan-
i y he e ec s in pe cen changes (Fig.4 isualises changes
in ha es ed a eas; Fig.5 shows changes in yields). Popula-
ion is posi i ely associa ed wi h he ha es ed a ea o c ops;
a 1% inc ease in popula ion is associa ed wi h subs an ial
inc eases in ha es ed a eas o all ou c ops: 0.16% o ice
and soybean, 0.33% o whea , and 0.27% o maize. Rising
ag icul u al labou pe coun y also has posi i e in luences
on all c ops excep whea , wi h he s onges e ec on ice,
a 0.29% pe one pe cen inc ease. Road leng h has a mixed
e ec on he ha es ed a eas: inc easing oad leng h ends
o be associa ed wi h less ha es ed a eas o whea , ice and
soybean, bu he a eas o whea and maize end o expand
in coun ies wi h a dense oad ne wo k. GDD has posi i e
e ec s on he ha es ed a eas o all c ops excep whea ,
possibly because mos whea cul i a ion in China is win e
whea , sown in au umn. WTO accession in 2001 had a uni-
e sal nega i e e ec on he ha es ed a eas o all c ops, pa -
icula ly whea and soybean. GDP, ag icul u al machine y,
and ain all ha e negligible in luences on ha es ed a eas.
In e ms o yield changes, GDP has posi i e, albei mino
e ec s on c op yields: a 1% g ow h in GDP is posi i ely
associa ed wi h he yields o ice (0.02%), whea (0.02%),
maize (0.03%), and soybean (0.01%) (Fig.5). As o he
p oduc ion inpu s, mo e ag icul u al labou has li le in lu-
ence on yields; machine y and e ilise ha e posi i e e ec s
on all c ops aside om soybean. The inc ease in oad leng h
was posi i ely co ela ed wi h he yield g ow h o soybean,
and nega i ely co ela ed wi h ice and maize. Wea he
condi ions we e, unsu p isingly, impo an de e minan s o
yields. GDD is posi i ely associa ed wi h c op yields, wi h
pa icula ly s ong e ec s o whea and soybean a a ound
0.15%. Highe ain all is associa ed wi h highe yields o
whea and maize, while i has minimal e ec s on ice and
soybean. The WTO accession is associa ed wi h an inc ease
in he yields o whea , maize, and soybean, bu i has a nega-
i e e ec on ice yield.
The compa ison o he a iables wi h he s anda dised
eg ession coe icien s e eals he ela i e in luence o each
co a ia e in he c op-speci ic eg ession (Fig.6; see Sup-
plemen a y TableS5 o he de ailed eg ession esul s).
The popula ion is he mos impo an de e minan o all
ou c ops, wi h mo e people being associa ed wi h mo e
ha es ed a eas o all c ops. This e lec s he spa ial o e -
lapping o popula ion and majo ag icul u al a eas. The ha -
es ed a ea o ice is posi i ely a ec ed by mo e ag icul u al
labou and popula ion and s ongly nega i ely in luenced by
mo e oads. The ice p oduc ion a ea seems o be associa ed
Fig. 4 Pe cen age change in
ha es ed a eas o each c op
wi h a 1% inc ease in he
explana o y a iables (o in
he yea s a e he accession o
WTO o he WTO dummy);
ma ke s a e coe icien es i-
ma es, and whiske s deno e he
95% s anda d e o s
346 F.Yin e al.
wi h less dense oad ne wo ks. This may be due o he ac
ha la ge a eas o ice cul i a ion expanded in No heas
China, whe e he oad ne wo k is qui e low. Simila ly, soy-
bean also exhibi s a simila pa e n, as he majo expansion
and plan ing egions we e concen a ed in he emo e a eas
o No heas China.
The whea a ea is nega i ely ela ed o he accession o
WTO and highe g owing deg ee days, bu posi i ely ela ed
o popula ion and oad. The ha es ed a ea o soybean is
nega i ely associa ed wi h g ea e oad leng h and, unsu -
p isingly, accession o he WTO. The impo o soybeans
by China su ged signi ican ly ollowing i s accession o he
WTO. P esen ly, China s ands as by a he la ges impo e
o soybeans, accoun ing o app oxima ely wo- hi ds o he
o al aded olume (Gale e al.,2019).
As expec ed, he wea he a iables a e impo an o he
yields o all c ops, albei wi h a ying deg ees o in lu-
ence (Fig.7). The empe a u e indica ed by GDD is he
Fig. 5 Pe cen age change in
yields o each c op wi h a 1%
inc ease in he explana o y a i-
ables (o a e he accession o
WTO as WTO is a dummy a i-
able); ma ke s a e coe icien
es ima es, and whiske s deno e
he 95% s anda d e o s
Fig. 6 Va iable impo ance o
changes in ha es ed a eas o
each c op; do s a e s anda d-
ised coe icien s, and whiske s
deno e he 95% s anda d e o s
