Cha e jee, Somdeep; Ka ewa, Nee aj
Wo king Pape
D ough s: Lea ning con inui y and eco e y
ADBI Wo king Pape , No. 1466
P o ided in Coope a ion wi h:
Asian De elopmen Bank Ins i u e (ADBI), Tokyo
Sugges ed Ci a ion: Cha e jee, Somdeep; Ka ewa, Nee aj (2024) : D ough s: Lea ning con inui y and
eco e y, ADBI Wo king Pape , No. 1466, Asian De elopmen Bank Ins i u e (ADBI), Tokyo,
h ps://doi.o g/10.56506/IWFG1305
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/305428
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-nc-nd/3.0/igo/
ADBI Wo king Pape 1466 Cha e jee and Ka ewa
ADBI Wo king Pape Se ies
DROUGHTS: LEARNING CONTINUITY
AND RECOVERY
Somdeep Cha e jee and
Nee aj Ka ewa
No. 1466
July 2024
Asian De elopmen Bank Ins i u e
ADBI Wo king Pape 1466 Cha e jee and Ka ewa
The Wo king Pape se ies is a con inua ion o he o me ly named Discussion Pape se ies; he
numbe ing o he pape s con inued wi hou in e up ion o change. ADBI’s wo king pape s e lec
ini ial ideas on a opic and a e pos ed online o discussion. Some wo king pape s may de elop
in o o he o ms o publica ion.
Sugges ed ci a ion:
Cha e jee, S., and N. Ka ewa. 2024. D ough s: Lea ning Con inui y and Reco e y. ADBI
Wo king Pape 1466. Tokyo: Asian De elopmen Bank Ins i u e. A ailable:
h ps://doi.o g/10.56506/IWFG1305
Please con ac he au ho s o in o ma ion abou his pape .
Emails: [email p o ec ed], [email p o ec ed]
Somdeep Cha e jee is an Assis an P o esso a he Indian Ins i u e o Managemen ,
Calcu a, Kolka a, Wes Bengal, India. Nee aj Ka ewa is a Doc o al S uden a he Indian
Ins i u e o Managemen Lucknow, U a P adesh, India.
The iews exp essed in his pape a e he iews o he au ho and do no necessa ily e lec
he iews o policies o ADBI, ADB, i s Boa d o Di ec o s, o he go e nmen s hey ep esen .
ADBI does no gua an ee he accu acy o he da a included in his pape and accep s no
esponsibili y o any consequences o hei use. Te minology used may no necessa ily be
consis en wi h ADB o icial e ms.
Wo king pape s a e subjec o o mal e ision and co ec ion be o e hey a e inalized and
conside ed published.
The au ho s hank Daniel Su yada ma, Te sushi Sonobe, Albe Pa k, Ha y Pa inos, M. Niaz
Asadullah, and a ious con e ence pa icipan s a ADBI, Japan, o use ul sugges ions and
inpu s. Any emaining e o s a e ou own.
Asian De elopmen Bank Ins i u e
Kasumigaseki Building, 8 h Floo
3-2-5 Kasumigaseki, Chiyoda-ku
Tokyo 100-6008, Japan
Tel: +81-3-3593-5500
Fax: +81-3-3593-5571
URL: www.adbi.o g
E-mail: [email p o ec ed]
© 2024 Asian De elopmen Bank Ins i u e
ADBI Wo king Pape 1466 Cha e jee and Ka ewa
Abs ac
In his pape , we s udy he nuanced ela ionship be ween d ough , academic pe o mance,
and he e icacy o p e en i e measu es, ocusing on wo d ough -p one egions om he
s a es o Maha ash a and Ka na aka in India. U ilizing an e en s udy amewo k elying on
di e ence-in-di e ences eg essions, we add ess he ecu en and non-pe manen na u e o
d ough s, elucida ing hei impac on lea ning ou comes o e ime. In Maha ash a, which lacks
p e en i e schemes, successi e d ough s lead o a signi ican d op in s anda dized es sco es
o ma h and eading skills, ollowed by in e mi en eco e y and u he in e up ions. The
school d opou a e also exhibi s a simila pa e n. Con a ily, Ka na aka’s exis ing
go e nmen -led p e en i e p og ams demons a e esilience in academic achie emen du ing
d ough pe iods.
Keywo ds: clima e a iabili y, human capi al in es men , educa ional a ainmen
JEL Classi ica ion: I21, O12, Q54, Q58
1 In oduc ion
A majo consequence o clima e change is p olonged ex eme wea he e en s (Scoones, 1992; Angassa and
Oba, 2008), including he ecu ence o na u al disas e s such as d ough s (Pa y e al., 2010; IPCC, 2014;
Wo ld Bank, 2019). While his has di ec implica ions o global de elopmen in gene al (S ocke e al.,
2013), among he a ious ypes o na u al disas e s, d ough is pe cei ed as he mos se e e haza d a ec ing
ag icul u al p oduc i i y (Wilhi e and Buchanan-Smi h, 2005; A shad e al., 2017; A shad e al., 2018).
Gi en ha a conside able p opo ion o he popula ion in low- and middle-income coun ies depends on
ag icul u e ei he di ec ly o indi ec ly (Mahend a De , 2012; Meha e al., 2016), hese ad e se impac s can
p ecipi a e o ced mig a ion (Munshi, 2003; Hen y e al., 2003; Feng e al., 2010; Dillon e al., 2011; G ay and
Muelle , 2012; Mbaye, 2017) and a ec household incomes (Paxson, 1992; Miguel e al., 2004; Skou ias e al.,
2012; Kesha a z and Ka ami, 2013), he eby subsequen ly in luencing household consump ion expendi u es
and in es men s in human capi al, such as in es men s in educa ion (Benson and Clay, 2004; Fi zsimons,
2007; Du yea e al., 2007; Ba ios e al., 2008; Bjo kman-Nyq is , 2013; Agamile e al., 2021).
While his li e a u e seems o sugges ha he e is a nega i e impac o poo ain all on educa ion
ou comes, mo e gene ally, he e idence on he impac s o such ex eme wea he e en s on human capi al
ou comes is mixed (Randell and G ay, 2016; Shah and S einbe g, 2017). Fo ins ance, he e is some e idence
ha wa e sca ci y induced by d ough s could lead o highe oppo uni y cos s o en ollmen and esul s in
a ansi ion away om school pa icipa ion owa ds labo ma ke pa icipa ion (Glo y and Nsikak-Abasi,
2013). O he s udies ha e a emp ed o link ex eme wea he e en s o ou comes such as school d opou
a es (Khalili, 2020), educa ion expendi u es (Amja h-Babu e al., 2016), and he labo supply (G a Zi in
and Neidell, 2014). Mo eo e , he li e a u e highligh s he mul i ace ed impac o ea ly-li e ain all on
socioeconomic, heal h, and educa ional ou comes (Maccini and Yang, 2009). Howe e , he es ima es o
educa ional a ainmen among s uden s as a consequence o exposu e o hese shocks emain limi ed, wi h
he excep ion o Shah and S einbe g (2017), who u ilize es sco es as a p oxy o human capi al a ainmen
and demons a e ha posi i e ain all shocks esul in educed school en ollmen and a endance, along wi h
diminished o e all es sco es o school-aged child en.1
Ou pape is closely ela ed o his, and we s udy an analogous ou come a iable, i.e., s anda dized
es sco es. Howe e , unlike Shah and S einbe g (2017), who concen a e on gauging he a e age annual
1In e es ingly, an al e na i e iewpoin is p esen ed he e by Shah and S einbe g (2017), emphasizing he impo ance o human
capi al in es men s and he al e na i e alue o ime. A subs i u e o educa ion is engagemen in ag icul u al ac i i ies o
household esponsibili ies. Consequen ly, du ing yea s wi h a o able monsoons, households expe ience an inc eased oppo uni y
cos o keeping hei child en en olled in school, as he child en could con ibu e o he household’s ea nings i hey we e no
in school. Shah and S einbe g’s indings om India indica e ha child en a e mo e likely o d op ou o school and pe o m
poo ly du ing good ain all yea s, sugges ing ha highe ain all nega i ely a ec s hei academic pe o mance (o con e sely,
s uden s a end school and pe o m be e du ing poo ain all yea s).
2
p ecipi a ion and i s cu en o nex -yea e ec s, ou objec i e is wo- old. Fi s , we in end o analyze he
epe cussions o d ough e en s in egions p one o d ough o e an ex ended pe iod. We show ha d ough s
lead o lea ning losses as well, al hough in he con ex o ou s udy, his is imp ecisely es ima ed. Second, we
ake ad an age o his app oach o in es iga e eco e y imes and assess he e icacy o p e en i e measu es.
Essen ially, we compa e and con as wo dis inc egions di ec ly a ec ed by d ough s in he con ex o India.
Howe e , one o he egions had in place ce ain measu es a ge ed owa ds coping wi h he a e ma h o
he d ough e en , whe eas he o he did no . We use hese se ings as a na u al expe imen o analyze he
eco e y om d ough in hese egions in he con ex o he es ima ed lea ning losses, which in u n allows
us o es ima e he e icacy o public policy designed o cope wi h he d ough s.
The Indian Me eo ological Depa men (IMD) has es ablished me eo ological subdi isions, hence o h
e e ed o as subdi isions (dis inc om adminis a i e dis ic s, he eina e called dis ic s), based on
simila i ies in clima e and monsoon ain all pa e ns. These subdi isions o m he basis o ou iden i ica-
ion s a egy. Unlike con olled expe imen s o policy implemen a ions, d ough anscends adminis a i e
bounda ies, a ec ing b oade geog aphical a eas wi h di used exposu e. We acknowledge he non-pe manen
na u e o d ough and he ac ha impac egions may change wi h each d ough e en . The e o e, o ad-
d ess his issue, we ocus on subdi isions ha a e d ough -p one and aced pe sis en d ough in he decade
s a ing om he yea 2011. These d ough e en s, along wi h hei se e i y and he egions hey a ec ,
ha e been well documen ed in go e nmen epo s, news a icles, and o he esea ch pape s. Addi ionally, we
empi ically alida e he equi ed ele ance assump ion o ou iden i ica ion s a egy, con i ming ha hese
d ough -p one subdi isions aced se e e d ough compa ed o he es o hei espec i e s a es du ing he
s udy pe iod. Consequen ly, he chosen se ing o ou s udy is he semi-a id egions o he Deccan Pla eau,
speci ically Ma a hwada in Maha ash a and No h In e io Ka na aka (NIK), which ace he highes e-
quency o se e e d ough s in India. Al hough adminis a i ely bo h subdi isions a e in di e en s a es, hey
sha e me eo ological bo de s and belong o he same a id zone o he Deccan Pla eau, expe iencing simila
clima ic condi ions. Thus, hey encoun e ed compa able d ough inciden s in e ms o in ensi y and iming
du ing he s udy pe iod, as co obo a ed by ou da ase and seconda y sou ces.
While hese neighbo ing me eo ological subdi isions sha e clima ic simila i ies, hey also p o ide a com-
pelling backd op o ou compa a i e analysis o he eco e y and coping s a egies in he a e ma h o
d ough s. This is because, unlike Maha ash a, he s a e o Ka na aka had explici go e nmen p og ams
in place o assis wi h coping wi h hese e en s. Essen ially, his c ea es a na u al expe imen o s udying
ou ques ion. Ou iden i ica ion s a egy elies on sepa a ely compa ing d ough -p one a eas in he a ec ed
s a es o he non-d ough a eas be o e and a e he d ough e en s, o e a pe iod o ime. We pe o m
sepa a e e en s udies o Maha ash a and Ka na aka o ind sugges i e e idence o lea ning losses in bo h
3
s a es as a esul o he d ough , bu wi h di e en a es o eco e y, wi h Ka na aka ou pe o ming Ma-
ha ash a. This lends suppo o he hypo hesis ha public policy di ec ed owa ds eco e y om hese
ex eme wea he e en s is e ec i e in helping a egion o ecupe a e om lea ning losses. We use da a on
es sco es om he Annual S a us o Educa ion Repo (ASER) in wo s a es, Maha ash a and Ka na aka,
comp ising 0.45 million obse a ions be ween 2008 and 2018.
We ind ha , in Maha ash a, a signi ican decline in ma h and eading sco es occu ed in he a e ma h
o consecu i e d ough yea s in 2011–2012. Al hough he e was a g adual eco e y du ing he no mal yea o
2013, his p og ess was dis up ed by ano he d ough in 2014–2015. Con e sely, in Ka na aka, whe e p e en-
i e p og ams we e in place, academic sco es expe ienced a ela i ely modes decline ollowing consecu i e
d ough yea s. No ably, he eco e y o lea ning losses a e successi e d ough yea s was mo e p omp in
he d ough -p one egion o Ka na aka compa ed o Maha ash a. This obse a ion unde sco es ha he im-
plemen a ion o go e nmen -led d ough p e en i e p og ams can enhance he esilience o academic sco es,
indica i e o educa ional ou comes, agains he ad e se impac s o consecu i e me eo ological d ough s o a
ce ain ex en . Addi ionally, he indings ega ding d opou a es u he suppo his asse ion.
In Maha ash a, we compa e Ma a hwada wi h he es o he s a e o unde s and he impac o d ough
and he eco e y pe iod in he absence o p e en i e measu es (we call his ou baseline scena io). Du ing he
ele an s udy pe iod, he s a e did no ha e any speci ic scheme o p og am o p e en he ad e se e ec s o
ain all sho ages, unlike he neighbo ing s a e o Ka na aka, o he bes o ou knowledge. The Ka na aka
s a e go e nmen , howe e , implemen ed speci ic p e en i e ini ia i es such as he Sujala scheme and P ojec
Bhooche ana o add ess wa e sca ci y and he d ough impac . Wo k unde hese p ojec s was comple ed
be o e he consecu i e d ough e en s s a ing in 2011. Addi ionally, be o e he consecu i e d ough yea s o
2014–2015, he go e nmen had also olled ou he nex phase o hese ini ia i es, along wi h newe ini ia i es
suppo ed by he Wo ld Bank. These d ough p e en i e e o s we e expec ed o ha e esul ed in imp o ed
access o esou ces, enhanced ag icul u al p oduc i i y, and inc eased household income, demons a ing a
commi men o sus ainable wa e managemen . The e o e, he es ima ed impac s o d ough s on lea ning
o he s a e o Ka na aka, whe e we compa e NIK wi h he es o he s a e, would likely be masked
by he e ec i eness o public sec o p e en i e measu es in d ough -p one a eas. Consequen ly, based on
assump ions abou he coun e ac ual, we ail o ejec he hypo hesis ha he dampened lea ning losses in
Ka na aka and he phenomenal eco e y may be a ibu ed o hese public policy measu es.
The es o he pape is s uc u ed as ollows: Sec ion 2 o e s backg ound in o ma ion on he s udy’s
ocus a ea. In Sec ion 3, we de ail he sou ces o ou da a. Sec ion 4 examines he empi ical amewo k,
encompassing he iden i ica ion s a egy and me hodology. In Sec ion 5, we i s alida e iden i ica ion
assump ions and hen p esen he main indings, along wi h suppo ing e idence and obus ness es esul s.
4
The conclusion summa izes ou indings and ou lines u u e esea ch a enues.
2 Backg ound
India, cha ac e ized by i s ex ensi e geog aphical span and di e se clima es, is pa icula ly ulne able o he
impac s o clima e change (MOEFCC, 2018). Among hese, d ough eme ges as a pe sis en h ea wi h
widesp ead epe cussions (Dai, 2013). I is widely acknowledged ha d ough has a de imen al impac on he
economy, wi h he Global Assessmen Repo (GAR) e ealing ha se e e d ough s can lead o an es ima ed
2–5% educ ion in India’s g oss domes ic p oduc (GDP). The phenomenon o d ough is in ica e and elies
on in e ac ions wi h a ious hyd ological pa ame e s such as e apo a ion, p ecipi a ion, in il a ion, uno ,
and su ace wa e and g oundwa e s o age (Si da¸s and Sen, 2003; Es ahanian e al., 2017; Kim e al., 2017).
Based on hese pa ame e s, d ough s a e commonly classi ied in o me eo ological, ag icul u al, hyd ological,
and socio-economic ca ego ies (Heim, 2002). Me eo ological d ough is igge ed by high empe a u es and
low p ecipi a ion, and i esul s in wa e sca ci y. This, in u n, leads o ag icul u al d ough , causing s ess
on c ops and impac ing people’s li elihoods, he ecological en i onmen , and socio-economic condi ions
(U a uk and Laosuwan, 2017; Du e al., 2018).
2.1 D ough -p one Zones in India
D ough is a non-pe manen and ecu ing phenomenon caused by a p olonged de iciency in ain all compa ed
o long- e m a e age condi ions in a gi en a ea (Pa el and Yada , 2015; S eekesh e al., 2019). This is
especially p e alen in semi-a id ecosys ems (Hind and Ma wan, 2010; Mohammad e al., 2015; Hussein,
2018; Mohammad e al., 2018; Sandeep e al., 2021).
Wi hin India’s clima ic landscape, he sou he n semi-a id zone, spanning he s a es o Maha ash a,
Ka na aka, Telangana, and Andh a P adesh, aces clima ic ulne abili y due o i s loca ion on he leewa d
side o he Wes e n Gha s. The Wes e n Gha s, a o midable moun ain ange unning pa allel o he wes e n
coas , c ea es a ain shadow in he eas e n egion, known as he Deccan Pla eau (Mohamed e al., 2019).
Because o whe e i is loca ed, he Deccan Pla eau o en does no ge enough wa e , making i d y on
he leewa d side o he Wes e n Gha s, which block ain clouds coming om he A abian Sea. Consequen ly,
wa e sca ci y pe sis s as a signi ican challenge o he s a es in he Deccan Pla eau2.
2Re e o Figu e 2 o he case s udy i led “D ough Cha ac e is ics o e he Deccan Pla eau Region o India,” by S ini asan
e al., in he GAR Special Repo on D ough 2021 by he Uni ed Na ions O ice o Disas e Risk Reduc ion. The published
documen s o he case s udy can be accessed he e. In his case s udy, Figu e 2 p esen s he a eas showing composi e ain all
de ia ions om he clima ological no mal (1981–2010) om June–Sep embe o e y d y yea s (2002, 2004, and 2015). I
highligh s ha he mos a ec ed a eas in he wo s a es, Maha ash a and Ka na aka, all wi hin hei espec i e mos -d ough -
p one me eo ological subdi isions, i.e., Ma a hwada in Maha ash a and No h In e io Ka na aka (NIK) in Ka na aka.
5
Low and i egula ain all, along wi h ex eme empe a u es and in ense sola adia ion, makes hese semi-
a id egions in sou he n peninsula India highly ulne able (INECC, 2010). Wa e sca ci y is p e alen , wi h
low g oundwa e ables, minimal ain all, and high wa e uno . The p ima y wa e sou ce h oughou he
yea in hese egions is small and medium amoun s o s o ed wa e . Consequen ly, he Global Assessmen
Repo (GAR) on D ough 2021, eleased by he UN O ice o Disas e Risk Reduc ion, iden i ies he Deccan
egion as expe iencing he highes equency (mo e han 6%) o se e e d ough s in all o India.
The se e e epe cussions o d ough - ela ed c op ailu es a e i idly highligh ed by a conce ning s a is ic
in he mos ecen epo om he Na ional C ime Reco ds Bu eau (NCRB) in he yea 2021—49.6% o
a me suicides in India a e concen a ed in he s a e o Maha ash a, ollowed by Ka na aka (Acciden al
Dea hs & Suicides in India (ADSI), 2022). No ably, bo h s a es a e si ua ed in he geog aphical egion
benea h he Wes e n Gha s on he Deccan Pla eau. This egion, iden i ied as expe iencing he highes
equency (mo e han 6%) o se e e d ough s in all o India, acco ding o he Global Assessmen Repo
(GAR), u he emphasizes he co ela ion be ween geog aphical loca ion, d ough si ua ions, and hei
dis essing consequences.
Maha ash a is he hi d la ges s a e in India based on geog aphical a ea and he highes con ibu o o
India’s GDP a 14% (Deshpande, 2023). Ag icul u e plays an impo an ole in he economy o Maha ash a.
The o al popula ion o he s a e is 112 million acco ding o he 2011 census, o which 57.8% a e dependen
on ag icul u e. Abou 84% o he o al a ea unde ag icul u e in he s a e is di ec ly dependen on monsoon
ain all (Economic Su ey o Maha ash a 2012-13, 2013). A he same ime, a subs an ial po ion o he
s a e lies in he semi-a id egions in he ain shadow o he Wes e n Gha s (Todmal, 2019).
Thus, d ough is one o he majo na u al disas e s ha pose a g ea h ea o he s a e economy and
ag icul u al de elopmen . To s udy and o ecas wea he , he s a e has been di ided in o ou me eo ological
subdi isions3: Konkan (wes coas ), Madhya Maha ash a, Ma a hwada, and Vida bha. Me eo ological
subdi isions a e c ea ed by he Indian Me eo ological Depa men based on simila clima e and monsoon
ain all pa e ns (Kelka and S eeji h, 2020). His o ically, Maha ash a has been plagued by nume ous
d ough e en s, such as in he yea s 1972, 2000, and 2004, and mos ecen ly, Maha as a aced a dis essing
se ies o consecu i e d ough s om 2011–2012 (Maha ash a Ahead, 2013), 2014–2015 (Kulka ni e al., 2016),
and 2018–2019 (Singh e al., 2022). When a d ough e en occu s o wo o mo e consecu i e yea s, hen i
is conside ed o be a pe sis en d ough e en (Am i e al., 2018).
Ma a hwada, a egion in a id po ions o Maha ash a, has pa icula ly been he epicen e o such pe -
sis en d ough s. Nea ly 73.83% o he popula ion o he Ma a hwada egion depends on ag icul u e as hei
p ima y sou ce o income (Kelka , 2013). None heless, he i iga ion p o ision in his a ea is qui e de icien ,
3Re e o he Indian Me eo ological Depa men ’s websi e o mo e de ail: h ps://mausam.imd.go .in/.
6
o 2011, he subdi isions p one o d ough (Ma a hwada in Maha ash a and NIK in Ka na aka) ecei ed
signi ican ly lowe ain all compa ed o he es o he egions o hei espec i e s a es. Fo ins ance, in
Maha ash a, in he yea 2010, which was conside ed a no mal yea , his di e ence amoun ed o 97.734
mm. By 2011, his di e ence had inc eased o 168.6887 mm. In pe cen age e ms, d ough -p one egions
in Maha ash a expe ienced a d as ic 31.56% dec ease in ain all in he yea 2011 compa ed o he p e ious
yea , whe eas he es o Maha ash a only saw a modes 2.52% dec ease. Simila ly, in Ka na aka, d ough -
p one a eas expe ienced a no able 29.41% d op in ain all in 2011 om he p e ious yea , while he es o he
s a e ac ually saw an inc ease o 7.28% du ing he same pe iod. This subs an ial d op in ain all pe sis ed
in he ollowing yea , ma king wo consecu i e d ough yea s o Ma a hwada and NIK. While o he egions
in bo h s a es also aced a sho all in ain all in 2012, i is c ucial o no e ha e en be o e his dip, he
ain all in hese d ough -p one subdi isions was consis en ly lowe han ha in he es o hei espec i e
s a es.
The examina ion o he aw means o ain all da a, coupled wi h e idence om he backg ound sec ion,
ins ills con idence ha Ma a hwada and NIK we e indeed se e ely impac ed by d ough in Maha ash a and
Ka na aka s a ing om 2011. Howe e , we u he igo ously es his in an e en s udy se ing and he
esul s a e p esen ed in Figu es A1 and A2, wi h a de ailed discussion p o ided la e in subsec ion 5.1.
4.2 Me hodology
Using he o dina y leas squa es (OLS) me hod o e alua e he impac o d ough s is p oblema ic. The e
can be unobse able ac o s ha can lead o omi ed a iable bias. Fo ins ance, compa ing child en in
he d ough -p one di ision wi h hose in o he egions o he s a e migh be con ounded by p e-exis ing
dispa i ies be ween he wo a eas, possibly skewing ou comes. Simila ly, compa ing pos -d ough -pe iod
ou comes wi h p e-d ough -pe iod ou comes wi hin he d ough -p one di ision migh yield biased esul s
due o a ious ac o s, including shi s in he mac oeconomic landscape o e ime.
To add ess hese issues, we p opose an e en s udy amewo k using he di e ence-in-di e ences me hod-
ology. As a esul , we le e age spa io- empo al a ia ions a ising om pe sis en d ough occu ences o
iden i y hei causal impac on lea ning ou comes and he eco e y pe iod o lea ning losses. Ou analysis
ocuses on a se ies o d ough e en s s a ing in 2011 in d ough -p one subdi isions o Maha ash a and
Ka na aka. Gi en ha ou p ima y da ase s a s in 2008, a yea no ma ked by se e e d ough condi ions,
we ha e designa ed 2008 as he e e ence pe iod (base yea ) o ou analysis wi hin his amewo k.
To es ima e he causal impac o d ough on lea ning ou comes and he eco e y pe iod o lea ning
losses, we employ he ollowing eg ession equa ion o each child i om dis ic din s a e sdu ing he ime
13
pe iod :
T es Sco eids =αd+β1·(D ough ds ∗T ime ) + β2·(T ime ) + βi·(Xi) + µids .(1)
In Equa ion 1, T es Sco e ep esen s he s anda dized eading o ma h es sco es o child iin ime .
αd ep esen s dis ic ixed e ec s; i con ols o ime-in a ian dis ic speci ic cha ac e is ics. T ime is a
dummy a iable wi h eigh e sions, each designed o acili a e he cons uc ion o an e en s udy amewo k.
Ac oss hese e sions, i assumes a alue o 0 o he base yea 2008 and subsequen ly akes he alue 1 o
each yea om 2009 o 2018 (excluding 2015 and 2017, as ou p ima y da ase lacks da a o hese yea s).
D ough ds is a dummy a iable indica ing whe he he indi idual belongs o he d ough -p one subdi ision
o s a e s. In he con ex o Maha ash a, his subdi ision co esponds o Ma a hwada, while in Ka na aka,
i is NIK. As we know ha d ough is non-pe manen in na u e and is no es ic ed by adminis a i e
bounda ies, o a oid con amina ion in he con ol g oup, we excluded he dis ic s in s a e s ha aced
se e e d ough du ing he pe iod o s udy bu a e no a pa o he d ough -p one me eo ological subdi ision
in s a e s. Demog aphic con ols ep esen ed by Xencompass he child’s age, he child’s gende , he amily
size, and he mo he ’s educa ion. S anda d e o s a e adjus ed o dis ic -le el clus e ing.
The in e p e a ion o he coe icien β1yields he di e ence in mean es sco es o ime compa ed o
he base yea , di e en ia ing be ween child en om he d ough -p one subdi isions and hose om o he
egions o he s a e s. The iden i ying assump ion implies ha in he coun e ac ual scena io, he es ima ed
coe icien would be s a is ically insigni ican . This is demons a ed in he esul s sec ion.
5 Resul s
In his sec ion, we employ an e en s udy di e ence-in-di e ences amewo k, as desc ibed in he me hodology
sec ion (subsec ion 4.2), o p esen he ou comes o ou analysis ega ding d ough inciden s in he d ough -
p one a eas o Maha ash a and Ka na aka, s a ing om he yea 2011.
5.1 Valida ing he T ea men S a us
To con i m ha ou iden i ica ion o d ough -p one egions is easonable, we use ain all da a om ou
p ima y da ase as he dependen a iable in he main empi ical equa ion (Equa ion 1), discussed in he
me hodology sec ion (subsec ion 4.2). This is execu ed o bo h s a es, encompassing all yea s s a ing om
2009 wi h espec o he base yea 2008 in he e en s udy se up. Figu es A1 and A2 plo he es ima ed
coe icien s o he di e ence in he a e age annual ain all o each ime pe iod compa ed o he base yea ,
14
dis inguishing be ween he Ma a hwada egion and he es o he egions in Maha ash a, as well as he
NIK egion and he es o he egions in Ka na aka, espec i ely.
These igu es indica e ha p io o 2011, he di e ence in he a e age annual ain all be ween he d ough -
p one subdi isions and he es o he egions in hei espec i e s a es was no s a is ically signi ican when
compa ed o he base yea . Howe e , in 2011, a s a is ically signi ican decline in he a e age annual ain all
is obse ed in he d ough -p one subdi isions compa ed o he es o he egions, as indica ed by he nega i e
coe icien s. Howe e , in he yea 2012, a ma ginal posi i e coe icien is obse ed o he Ma a hwada egion
in compa ison o he es o Maha ash a when compa ed o he base yea . Howe e , his obse ed inc ease is
no s a is ically signi ican and does no indica e an ac ual ise in ain all. Ins ead, Ma a hwada expe ienced
a decline in ain all o he second consecu i e yea , and he posi i e coe icien is in luenced by o he egions
in Maha ash a also acing sho ages in ain all du ing ha yea ( e e o Figu e A2 o he online appendix).
A simila pa e n is no ed o Ka na aka in 2012, whe e a non-signi ican posi i e coe icien is obse ed, and
he aw means o he ain all da a align wi h he ac ual scena io (Figu e A3 o he online appendix).
5.2 Main Findings
As discussed in he backg ound sec ion (Sec ion 2), he Maha ash a go e nmen lacked p e en i e schemes
o add ess d ough . Despi e he implemen a ion o he Jalyuk Shi a scheme a e consecu i e d ough
yea s in 2014–2015, i s e ec i eness was ques ioned.
This ci cums ance o e s an oppo uni y o in es iga e he impac o d ough in he absence o p e en i e
measu es. To accomplish his, we employ he di e ence-in-di e ences eg ession o Maha ash a in an e en
s udy amewo k ou lined in Equa ion 1, as de ailed in he me hodology sec ion (subsec ion 4.2), o examine
he causal impac o d ough on lea ning ou comes. Equa ion 1 is sepa a ely applied o eading and ma h
sco es o he a ious yea s conside ed in ou s udy, wi h 2008 as he base yea . The esul s o child ma h
sco es a e p esen ed in Figu e 1, and hose o child eading sco es a e epo ed in Figu e 2 based on he
es ima ion o ou eg ession equa ion. Panels 1 and 2 o Table A1 o he appendix con ain he ull eg ession
able co esponding o hese igu es.
We obse e a s a is ically signi ican d op in ma h sco es ollowing consecu i e d ough yea s in 2011–
2012. In 2012, he decline in ma h sco es is app oxima ely 0.22 σpoin s compa ed o he base yea , as
compa ed o he base yea be ween he a e age child in Ma a hwada and he es o Maha ash a a e
accoun ing o ime-in a ian dis ic -speci ic e ec s. This lea ning loss in ma h sco es g adually begins o
eco e du ing he no mal yea o 2013, only o be ollowed by ano he consecu i e d ough a ec ing he
exposed g oup in 2014–2015. A simila impac is disce nible in eading sco es, wi h a no able d op in he
15
yea 2012. Subsequen ly, eco e y is obse ed un il he nex consecu i e d ough in 2014–2015 impac s he
exposed g oup.
The impac o d ough on he academic pe o mance o exposed g oups, speci ically in ma hema ics and
eading sco es, exhibi s a simila i y in magni ude. Howe e , gi en ha ma hema ics sco es a e lowe han
eading sco es in Maha ash a, and pa icula ly in Ma a hwada, he nega i e impac is mo e p onounced in
ma hema ics. Fu he mo e, he eco e y om he impac o d ough appea s o be swi e in eading sco es
han in ma hema ics sco es. This dispa i y indica es ha child en may al eady possess weake ma hema ical
skills compa ed o eading skills, ampli ying he de imen al e ec s o d ough on ma hema ics pe o mance.
Mo eo e , we obse e ha despi e d ough occu ing in bo h 2011 and 2012, he e ec s on academic
pe o mance begin o mani es om 2012 onwa ds. This sugges s a empo al lag o one yea be ween he
onse o d ough and i s disce nible in luence on lea ning ou comes. Gi en his hypo hesis, o he second
consecu i e d ough (which occu ed in 2014–2015), we would an icipa e obse ing i s subs an ial impac on
es sco es in he yea 2015. Howe e , since ASER da a o 2015 and 2017 a e no publicly a ailable, we we e
unable o comp ehensi ely assess he e ec s o he second consecu i e d ough . Despi e incomple e da a,
we s ill obse e a sligh dec ease in ma h sco es, hinde ing eco e y om p io d ough yea s, alongside a
signi ican decline in eading sco es.
5.2.1 E ec i eness o D ough P e en i e S a egies by he Public Sec o 16
As discussed ea lie , Ka na aka, being a neighbo ing s a e, sha es a simila clima e and clima ic challenges
wi h Maha ash a, especially in he neighbo ing me eo ological subdi isions (Ma a hwada and NIK). How-
e e , unlike Maha ash a, Ka na aka had exis ing d ough p e en i e p og ams be o e a se ies o consecu i e
d ough s s a ing in 2011. This p esen s an oppo uni y o assess he e ec i eness o Ka na aka’s p e en i e
s a egies. Wi h bo h s a es acing compa able clima ic condi ions, his analysis o he neighbo ing me eo-
ological di ision (NIK) allows us o e alua e he impac o d ough on lea ning ou comes wi h p e en i e
measu es in place.
To achie e his objec i e, we employ he di e ence-in-di e ences eg ession o Ka na aka in an e en
s udy amewo k ou lined in Equa ion 1, as de ailed in he me hodology sec ion (subsec ion 4.2), o assess
he in luence o d ough on lea ning ou comes wi h p e en i e measu es implemen ed. Equa ion 1 is inde-
penden ly applied o eading and ma h sco es o he di e en yea s unde examina ion in ou s udy o
Ka na aka, wi h 2008 se ing as he e e ence yea . The ou comes o child ma h sco es a e depic ed in
16We also examine he ole o he p i a e sec o in d ough -p one a eas o add ess wa e sca ci y; unlike go e nmen e o s,
he p i a e sec o o en a ge s localized a eas. Gi en he cons ain s o da a and he ime lag in educa ional ou comes, we
analyze expendi u e pa e ns. Howe e , he main ocus o ou pape on educa ional ou comes leads us o place his analysis o
p i a e sec o d ough p e en ion e ec i eness in he online appendix.
16
Figu e 4, and he esul s o child eading sco es a e illus a ed in Figu e 5, de i ed om he es ima ion o
ou eg ession equa ion. The ull eg ession able co esponding o hese igu es is p o ided in panels 1 and
2 o Table A2 o he appendix.
In con as o Maha ash a, he academic achie emen o an a e age child in he exposed g oup wi hin
Ka na aka, i.e., he NIK me eo ological subdi ision, appea s esilien o he impac o successi e me eo o-
logical d ough s. While he e is a sligh down u n in he eading sco es in 2011 ollowing consecu i e d ough
yea s in 2011–2012, his dip is nei he economically no s a is ically signi ican . Mo eo e , he eco e y om
his dip is swi when compa ed o he baseline scena io.
Howe e , he e could ha e been a po en ial issue wi h in e p e ing ou indings i Ka na aka inhe en ly
excelled in educa ion compa ed o Maha ash a, beyond d ough mi iga ion e o s. Howe e , ou summa y
s a is ics indica e ha Maha ash a sligh ly ou pe o ms Ka na aka in educa ion, wi h he e e se no hold-
ing ue. The e o e, Maha ash a canno be cha ac e ized as a poo pe o me in e ms o lea ning ou comes
compa ed o Ka na aka. In ac , Ma a hwada, wi hin Maha ash a, also ou pe o ms he NIK subdi ision o
Ka na aka. Addi ionally, Maha ash a and Ma a hwada ecei e sligh ly mo e ain all han Ka na aka and
NIK, espec i ely ( e e o Figu es A2 and A3 in he online appendix).
Fu he mo e, as he la e hal o he decade un olds, disce nible posi i e ends eme ge, indica ing
ha he ini ia i es implemen ed in d ough -p one a eas o he s a e a e yielding esul s. The academic
pe o mance o child en om he d ough -p one di ision (NIK) exhibi s an upwa d ajec o y compa ed o
he baseline yea o 2008.
5.3 E ec on School D opou Ra es
The p ima y ocus o his s udy is o assess he e ec s o d ough on lea ning ou comes. One signi ican
pa hway h ough which d ough may in luence lea ning ou comes is by compelling child en o wi hd aw
om school. In he con ex o clima e- ela ed disas e s such as d ough s, households acing wa e sca ci y
encoun e inc eased oppo uni y cos s associa ed wi h main aining hei child en’s en ollmen in school
(Glo y and Nsikak-Abasi, 2013).
To in es iga e his phenomenon and o p o ide addi ional suppo o ou p ima y indings abou lea ning
ou comes, we employ he e en s udy amewo k delinea ed in Equa ion 1 (see subsec ion 4.2) o examine
he causal impac o d ough on school d opou a es. We apply Equa ion 1 sepa a ely o d opou a es in
Maha ash a and Ka na aka, p esen ing he co esponding esul s in Figu es 3 and 6 o Maha ash a and
Ka na aka, espec i ely. The ull eg ession ables co esponding o hese igu es a e p o ided in panel 3 o
Appendix Tables A1 and A2.
17
Ou analysis e eals a s a is ically signi ican inc ease in d opou a es du ing consecu i e d ough yea s in
2011–2012 in he d ough -p one egions o Maha ash a, namely he Ma a hwada me eo ological subdi ision,
ollowed by a eco e y in subsequen yea s. In con as , he school pa icipa ion in he exposed g oup
wi hin Ka na aka, namely he NIK me eo ological subdi ision, appea s esilien o he impac o successi e
me eo ological d ough s. Mo eo e , posi i e ends eme ge in he la e yea s, indica ing he e ec i eness o
ini ia i es implemen ed in d ough -p one a eas o he s a e. These indings ega ding he impac o d ough
on school pa icipa ion complemen ou p ima y esul s on lea ning ou comes, p o iding addi ional suppo
and con idence in ou conclusions.
5.4 Resul s om an Al e na i e Iden i ica ion S a egy
In ou main e en s udy amewo k (Equa ion 1), we employed di e ence-in-di e ence eg ession equa ions
o assess he impac o d ough in compa ison o egions less p one o d ough wi hin a s a e o e ime.
Speci ically, we conduc ed sepa a e analyses o Maha ash a and Ka na aka o examine he e ec s o d ough
wi h and wi hou public sec o p e en i e measu es in place. By unning wo dis inc e en s udies o each
s a e, we a oided assuming ha Maha ash a and Ka na aka we e iden ical in hei coun e ac ual condi ions.
In his subsec ion, we p opose an al e na i e iden i ica ion s a egy u ilizing iple di e ence eg ession
equa ions wi hin he same e en s udy amewo k. This app oach allows o a mo e comp ehensi e com-
pa ison be ween s a es and acili a es he compa ison o he d ough -p one egions wi h hei co esponding
non-d ough -p one egions wi hin a single equa ion. Addi ionally, i allows us o es he obus ness o ou
es ima es.
The p oposed al e na i e iden i ica ion s a egy o each child i om dis ic din s a e sdu ing he ime
pe iod is as ollows:
Yids =αd+β1·(D ough d∗S a es∗T ime ) + β2·(D ough d∗S a es) + β3·(S a es∗T ime )
+β4·(D ough d∗T ime ) + β5·(S a es) + β6·(T ime ) + βi·(Xi) + µids ,
(2)
whe e Yids deno es he ou come a iable o child ia ime .T ime is a dummy a iable, a ying ac oss
eigh e sions o c ea e an e en s udy amewo k. I assumes a alue o 0 o he base yea 2008 and a alue
o 1 o each subsequen yea om 2009 o 2018, excluding 2015 and 2017 due o da a una ailabili y in ou
p ima y da ase . D ough dis a dummy a iable indica ing whe he a child belongs o he d ough -p one
subdi isions o no , aking he alue 1 o indi iduals om ei he Ma a hwada o NIK. S a esis a dummy
a iable ha is assigned he alue 1 o Maha ash a and 0 o Ka na aka. Con ol a iables consis en wi h
Equa ion 1 a e included, deno ed by Xi.αd ep esen s dis ic ixed e ec s, and s anda d e o s a e clus e ed
18
a he dis ic le el.
The esul s o he abo e-men ioned al e na i e iden i ica ion s a egy, depic ed in Figu es 7, 8, and 9
o he ma h sco e, eading sco e, and d opou a es, espec i ely, p o ide a combined pic u e o he ea lie
indings p esen ed sepa a ely o bo h s a es. The iple di e ence poin es ima es illus a ed in hese igu es
e eal ha , o e ime, p io o consecu i e d ough yea s, he e exis ed no signi ican dispa i y be ween he
wo s a es conce ning lea ning ou comes and school a endance. Howe e , ollowing he consecu i e d ough
yea s o 2011–2012, he ends indica e ha he d ough -p one egions o Maha ash a ace lea ning losses
and highe school d opou a es, while he d ough -p one egions o Ka na aka display esilience and p og ess
due o public sec o in e en ions in d ough -p one a eas.
5.5 Resul s om an Al e na i e Re e ence Yea
In ou p ima y empi ical amewo k, we es ablished 2008 as he baseline yea o ou e en s udy, as ou
p ima y da ase s a s om ha yea . Howe e , in his pa icula subsec ion, o he obus ness o ou
es ima es, we shi he baseline yea o 2010, he yea immedia ely p eceding a se ies o consecu i e d ough
e en s. Like he yea 2008, he yea 2010 was also no ma ked by se e e d ough condi ions. Consequen ly,
Equa ion 1, as elabo a ed in he me hodology sec ion (subsec ion 4.2), emains unchanged excep o he
modi ica ion o he T ime dummy a iable.
In his analysis, he T ime a iable se es as a dummy a iable wi h eigh e sions, each designed o
cons uc an e en s udy amewo k. Ac oss hese e sions, i is assigned a alue o 0 o he yea 2010,
which se es as he baseline yea o all e sions. Subsequen ly, o each e sion, he dummy a iable akes
on a alue o 1 o one yea a a ime, namely 2008, 2009, 2011, 2012, 2013, 2014, 2016, and 2018.
The ou comes o his analysis u ilizing he al e na i e base yea o he s a es o Maha ash a and
Ka na aka a e depic ed in Figu es 10–15, whe e Figu es 10, 11, and 12 pe ain o ma h sco es, eading sco es,
and d opou a es o Maha ash a, and Figu es 13, 14, and 15 co espond o he same ou come a iables o
Ka na aka. No ably, hese indings o bo h s a es align closely wi h he p ima y esul s ob ained wi h 2008
as he baseline yea . Thus, i is e iden ha ou conclusions emain consis en i espec i e o he choice o
he baseline yea .
6 Discussion
In ou s udy, we examine he e ec s o d ough and i s eco e y wi h and wi hou p e en a i e measu es,
using an e en s udy app oach. Since d ough s a e empo a y and hei impac a eas may a y wi h each
occu ence, we ocus on me eo ological subdi isions p one o d ough . Speci ically, we concen a e on he
19
semi-a id egions o he Deccan Pla eau in wo neighbo ing s a es, Maha ash a and Ka na aka, which expe-
ience equen se e e d ough s in India. We p o ide empi ical e idence suppo ing he ele ance assump ion
o ou s a egy ha he d ough -p one subdi isions expe ienced se e e d ough compa ed o o he a eas
wi hin hei espec i e s a es du ing he s udy pe iod.
We compa e he occu ence o d ough in hese d ough -p one subdi isions be ween he wo s a es. In
Maha ash a, we obse e lea ning losses in i s d ough -p one subdi isions ollowing d ough e en s, while
in Ka na aka, such losses a e no e iden . We hypo hesize ha his disc epancy is due o he p esence o
d ough -p e en a i e measu es in one s a e and hei absence in he o he .
Howe e , one conce n may a ise ega ding he in e p e a ion o ou indings i Ka na aka simply p o-
ides be e educa ion. Ou analysis, employing a iple di e ence se up o compa e bo h s a es and hei
d ough -p one egions wi h a baseline yea , disp o es his conce n. In ac , ou summa y s a is ics indica e
ha , on a e age, Maha ash a pe o med sligh ly be e in educa ional ou comes compa ed o Ka na aka.
Addi ionally, d ough -p one egions wi hin Maha ash a also demons a ed supe io academic pe o mance
compa ed o hose in Ka na aka. Ano he conce n could ela e o he d ough in ensi y and he imeline
o d ough occu ence in he d ough -p one egions o bo h s a es. Howe e , despi e being adminis a i ely
dis inc , he me eo ological subdi isions in bo h s a es sha e bo de s and belong o he same a id zone,
expe iencing simila clima ic condi ions and encoun e ing compa able d ough inciden s. Mo eo e , we also
demons a e using ou da ase and h ough seconda y sou ces ha bo h egions expe ienced simila d ough
inciden s in e ms o in ensi y and iming du ing he s udy pe iod.
Thus, ou main indings ega ding es sco es and school pa icipa ion suppo he hypo hesis ha p e en-
i e measu es in one s a e con ibu e o educa ion esilience in i s d ough -p one egions. These p e en i e
measu es we e p ima ily a ge ing ag icul u e, a c ucial pa hway h ough which me eo ological d ough a -
ec s human capi al in es men in ag icul u e-dependen economies. The e o e, ou s udy sugges s ha wi h
such p e en i e measu es in place, he d ough -p one egions in ag icul u e-dependen economies may s ill
ace me eo ological d ough s, bu he se e i y o hei impac may be mi iga ed.
Fu he mo e, we acknowledge he ole o co po a e ini ia i es in comba ing wa e c ises in d ough -
p one egions. While p i a e sec o e o s complemen hose o he public sec o , hey o en ope a e on a
smalle scale and wi h limi ed geog aphic co e age. Recognizing he po en ial limi a ions o dis ic -le el
obse a ion and he ime equi ed o es sco es o mani es , we shi ou ocus o expendi u e pa e ns.
Since he main ocus o ou pape is on educa ional ou comes, we place his analysis o p i a e sec o d ough
p e en ion e ec i eness and he associa ed esul s in he online appendix, highligh ing inc eased educa ion-
ela ed expenses in egions wi h a highe densi y o p i a e sec o d ough p e en ion p ojec s. Howe e , we
cau ion agains d awing causal conclusions due o limi a ions in da a g anula i y and he po en ial in luence
20
o o he p i a e sec o p ojec s on ou compa ison g oup.
Mo ing o wa d, p ima y esea ch ocusing on collec ing da a a lowe adminis a i e le els could p o ide
mo e a ge ed insigh s in o he e ec i eness o p i a e sec o d ough p e en i e measu es. Ou s udy
highligh s he impo ance o collec i e e o s in ol ing all s akeholde s in add essing wa e sca ci y. While
ou indings demons a e he posi i e impac s o go e nmen measu es on educa ion esilience, hey also
indica e he need o u he esea ch o assess he causal impac o p i a e sec o ini ia i es. Such e o s
would con ibu e o a comp ehensi e unde s anding o he collec i e ac ions needed o add ess wa e sca ci y
e ec i ely.
21
Re e ences
Acciden al Dea hs & Suicides in India (ADSI). (2022). NCRB. h ps://nc b.go .in/en/acciden al-dea hs-
suicides-india-adsi
Agamile, P., Dimo a, R., & Golan, J. (2021). C op choice, d ough and gende : New insigh s om smallholde s’
esponse o wea he shocks in u al Uganda. Jou nal o Ag icul u al Economics, 72(3), 829–856. h ps://doi.o g/10.1111/1477-
9552.12427
Amja h-Babu, T. S., K upnik, T. J., A a indakshan, S., A shad, M., & Kaechele, H. (2016). Clima e
change and indica o s o p obable shi s in he consump ion po olios o d yland a me s in Sub-Saha an A ica:
Implica ions o policy. Ecological Indica o s, 67, 830–838.
Am i , K., Pandey, R. P., & Mish a, S. K. (2018). Assessmen o me eo ological d ough cha ac e is ics o e
Cen al India. Sus ainable Wa e Resou ces Managemen , 4(4), 999–1010. h ps://doi.o g/10.1007/s40899-017-0205-
5
Angassa, A., & Oba, G. (2008). He de pe cep ions on impac s o ange enclosu es, c op a ming, i e ban and
bush enc oachmen on he angelands o Bo ana, Sou he n E hiopia. Human Ecology, 36, 201–215.
A shad, M., K¨achele, H., K upnik, T. J., Amja h-Babu, T. S., A a indakshan, S., Abbas, A., ... &
M¨ulle , K. (2017). Clima e a iabili y, a mland alue, and a me s’ pe cep ions o clima e change: implica ions o
adap a ion in u al Pakis an. In e na ional Jou nal o Sus ainable De elopmen & Wo ld Ecology, 24(6), 532-544.
A shad, M., Amja h-Babu, T. S., A a indakshan, S., K upnik, T. J., Toussain , V., K¨achele, H.,
& M¨ulle , K. (2018). Clima ic a iabili y and he mal s ess in Pakis an’s ice and whea sys ems: A s ochas ic
on ie and quan ile eg ession analysis o economic e iciency. Ecological Indica o s, 89, 496–506.
Ba ios, S., Oua a a, B., & S obl, E. (2008). The impac o clima ic change on ag icul u al p oduc ion: Is i
di e en o A ica? Food Policy, 33(4), 287–298. h ps://doi.o g/10.1016/j. oodpol.2008.01.003
Benson, C., & Clay, E. J. (2004). Unde s anding he economic and inancial impac s o na u al disas e s (No. 4).
Wo ld Bank Publica ions.
Bhanda i, D. C., Meghwal, P. R., & Lodha, S. (2014). Ho icul u e based p oduc ion sys ems in Indian a id
egions. In Sus ainable De elopmen and Biodi e si y (pp. 19–49). Sp inge In e na ional Publishing.
Bj¨o kman-Nyq is , M. (2013). Income shocks and gende gaps in educa ion: E idence om Uganda. Jou nal o
De elopmen Economics, 105, 237–253. h ps://doi.o g/10.1016/j.jde eco.2013.07.013
Cha e jee, S., Has awala, S., & Kamal, J. (2023). (En-)‘ligh ening’child en: Assessing he impac s o access
o elec ici y on lea ning achie emen le els. Re iew o De elopmen Economics.
DAC (2009). Manual o D ough Managemen , Depa men o Ag icul u e & Coope a ion, Go e nmen o India
Dai, A. (2013). Inc easing d ough unde global wa ming in obse a ions and models. Na u e Clima e Change, 3(1),
52–58. h ps://doi.o g/10.1038/nclima e1633
Depa men o Ag icul u e and Fa me s Wel a e, Go e nmen o India. (2016, Decembe ). Manual o
D ough Managemen . h ps://ag iwel a e.go .in/Documen s/Upda ed%20D ough %20Manual 0.pd
22
Figu e 2: Resul s o eading sco es om baseline scena io
No e: The igu e is based on Equa ion 1, compa ing Ma a hwada and he es o Maha ash a in he e en s udy amewo k
wi h 2008 as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o eading es sco es and include
demog aphic con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
Figu e 3: Resul s o d opou a e om baseline scena io
No e: The igu e is based on Equa ion 1, compa ing Ma a hwada and he es o Maha ash a in he e en s udy amewo k
wi h 2008 as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o he d opou a e and include
demog aphic con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
29
Figu e 4: Resul s o ma h sco es wi h public sec o d ough p e en i e measu es
No e: The igu e is based on Equa ion 2, compa ing NIK and he es o Ka na aka in he e en s udy amewo k wi h 2008
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o ma h es sco es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
30
Figu e 5: Resul s o eading sco es wi h public sec o d ough p e en i e measu es
No e: The igu e is based on Equa ion 2, compa ing NIK and he es o Ka na aka in he e en s udy amewo k wi h 2008 as
he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o eading es sco es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
Figu e 6: Resul s o d opou a e wi h public sec o d ough p e en i e measu es
No e: The igu e is based on Equa ion 2, compa ing NIK and he es o Ka na aka in he e en s udy amewo k wi h 2008
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o he d opou a e and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
31
Figu e 7: Resul s o ma h sco es using al e na i e iden i ica ion s a egy
No e: The igu e is based on Equa ion 2, compa ing d ough -p one egions agains hose no p one o d ough in wo s a es,
namely Maha ash a and Ka na aka, o e ime. This analysis is conduc ed wi hin an e en s udy amewo k, wi h he yea 2008
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o ma h es sco es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
32
Figu e 8: Resul s o eading sco es using al e na i e iden i ica ion s a egy
No e: The igu e is based on Equa ion 2, compa ing d ough -p one egions agains hose no p one o d ough in wo s a es,
namely Maha ash a and Ka na aka, o e ime. This analysis is conduc ed wi hin an e en s udy amewo k, wi h he yea 2008
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o eading es sco es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
Figu e 9: Resul s o d opou a es using al e na i e iden i ica ion s a egy
No e: The igu e is based on Equa ion 2, compa ing d ough -p one egions agains hose no p one o d ough in wo s a es,
namely Maha ash a and Ka na aka, o e ime. This analysis is conduc ed wi hin an e en s udy amewo k, wi h he yea 2008
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o d opou a es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
33
Figu e 10: Resul s using al e na i e e e ence yea o ma h sco es om baseline scena io
No e: The igu e is based on Equa ion 1, compa ing Ma a hwada and he es o Maha ash a in he e en s udy amewo k
wi h 2010 as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o ma h es sco es and include
demog aphic con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
Figu e 11: Resul s using al e na i e e e ence yea o eading sco es om baseline scena io
No e: The igu e is based on Equa ion 1, compa ing Ma a hwada and he es o Maha ash a in he e en s udy amewo k
wi h 2010 as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o eading es sco es and include
demog aphic con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
34
Figu e 12: Resul s using al e na i e e e ence yea o d opou a e om baseline scena io
No e: The igu e is based on Equa ion 1, compa ing Ma a hwada and he es o Maha ash a in he e en s udy amewo k
wi h 2010 as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o he d opou a e and include
demog aphic con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
35
Figu e 13: Resul s using al e na i e e e ence yea o ma h sco es wi h public sec o d ough
p e en i e measu es
No e: The igu e is based on Equa ion 1, compa ing NIK and he es o Ka na aka in he e en s udy amewo k wi h 2010
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o ma h es sco es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
36
Figu e 14: Resul s using al e na i e e e ence yea o eading sco es wi h public sec o d ough
p e en i e measu es
No e: The igu e is based on Equa ion 1, compa ing NIK and he es o Ka na aka in he e en s udy amewo k wi h 2010 as
he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o eading es sco es and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
Figu e 15: Resul s using al e na i e e e ence yea o d opou a e wi h public sec o d ough
p e en i e measu es
No e: The igu e is based on Equa ion 1, compa ing NIK and he es o Ka na aka in he e en s udy amewo k wi h 2010
as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o he d opou a e and include demog aphic
con ols and dis ic ixed e ec s. Ve ical lines ep esen 90% con idence in e als.
37
Appendix
Figu e A1: Resul s o ain all using p ima y empi ical amewo k
No e: The igu e is based on Equa ion 1, compa ing ain all (in millime e s) in Ma a hwada and he es o Maha ash a in he
e en s udy amewo k wi h 2008 as he e e ence pe iod. All poin s ep esen coe icien s om di e en eg essions o a e age
annual ain all. Ve ical lines ep esen 90% con idence in e als.
38
