Tian, Shu; Wu, Yu; Zhou, Wenwen
Wo king Pape
Digi aliza ion and income inequali y: E idence om
households
ADB Economics Wo king Pape Se ies, No. 764
P o ided in Coope a ion wi h:
Asian De elopmen Bank (ADB), Manila
Sugges ed Ci a ion: Tian, Shu; Wu, Yu; Zhou, Wenwen (2025) : Digi aliza ion and income inequali y:
E idence om households, ADB Economics Wo king Pape Se ies, No. 764, Asian De elopmen
Bank (ADB), Manila,
h ps://doi.o g/10.22617/WPS250008-2
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/310439
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/3.0/igo/
ASIAN DEVELOPMENT BANK
ASIAN DEVELOPMENT BANK
6 ADB A enue, Mandaluyong Ci y
1550 Me o Manila, Philippines
www.adb.o g
ADB ECONOMICS
WORKING PAPER SERIES
NO. 764
Janua y 2025
Digi aliza ion and Income Inequali y
E idence om Households
This pape examines how digi aliza ion a ec s income inequali y, using household da a om he People’s
Republic o China. The indings indica e ha digi aliza ion signi ican ly educes income inequali y,
pa icula ly in less-de eloped a eas and among lowe -educa ed households. The e ec emains signi ican
a e add essing obus ness and endogenei y conce ns. Digi aliza ion na ows he income gap by inc easing
ea nings mo e o lowe -income households h ough imp o ed employmen and in es men oppo uni ies.
I also boos s business income o en ep eneu ial households. These esul s sugges ha p omo ing
digi aliza ion can help educe income inequali y in de eloping economies.
Abou he Asian De elopmen Bank
ADB is commi ed o achie ing a p ospe ous, inclusi e, esilien , and sus ainable Asia and he Paci ic,
while sus aining i s e o s o e adica e ex eme po e y. Es ablished in 1966, i is owned by 69 membe s
—49 om he egion. I s main ins umen s o helping i s de eloping membe coun ies a e policy dialogue,
loans, equi y in es men s, gua an ees, g an s, and echnical assis ance.
DIGITALIZATION AND
INCOME INEQUALITY
EVIDENCE FROM HOUSEHOLDS
Shu Tian, Yu Wu, and Wenwen Zhou
ASIAN DEVELOPMENT BANK
The ADB Economics Wo king Pape Se ies
p esen s esea ch in p og ess o elici commen s
and encou age deba e on de elopmen issues
in Asia and he Paci ic. The iews exp essed
a e hose o he au ho s and do no necessa ily
e lec he iews and policies o ADB o
i s Boa d o Go e no s o he go e nmen s
hey ep esen .
ADB Economics Wo king Pape Se ies
Shu Tian, Yu Wu, and Wenwen Zhou
No. 764 | Janua y 2025
Shu Tian ([email p o ec ed]g) is a p incipal economis
a he Economic Resea ch and De elopmen
Impac Depa men , Asian De elopmen Bank.
Yu Wu ([email protected]) is a p o esso a
Nanjing Ag icul u al Uni e si y. Wenwen Zhou
([email p o ec ed]) is a PhD candida e a
Sou hwes e n Uni e si y o Finance and Economics,
Chengdu.
Digi aliza ion and Income Inequali y:
E idence om Households
C ea i e Commons A ibu ion 3.0 IGO license (CC BY 3.0 IGO)
© 2025 Asian De elopmen Bank
6 ADB A enue, Mandaluyong Ci y, 1550 Me o Manila, Philippines
Tel +63 2 8632 4444; Fax +63 2 8636 2444
www.adb.o g
Some igh s ese ed. Published in 2025.
ISSN 2313-6537 (p in ), 2313-6545 (PDF)
Publica ion S ock No. WPS250008-2
DOI: h p://dx.doi.o g/10.22617/WPS250008-2
The iews exp essed in his publica ion a e hose o he au ho s and do no necessa ily e lec he iews and policies
o he Asian De elopmen Bank (ADB) o i s Boa d o Go e no s o he go e nmen s hey ep esen .
ADB does no gua an ee he accu acy o he da a included in his publica ion and accep s no esponsibili y o any
consequence o hei use. The men ion o speci ic companies o p oduc s o manu ac u e s does no imply ha hey
a e endo sed o ecommended by ADB in p e e ence o o he s o a simila na u e ha a e no men ioned.
By making any designa ion o o e e ence o a pa icula e i o y o geog aphic a ea in his documen , ADB does no
in end o make any judgmen s as o he legal o o he s a us o any e i o y o a ea.
This publica ion is a ailable unde he C ea i e Commons A ibu ion 3.0 IGO license (CC BY 3.0 IGO)
h ps://c ea i ecommons.o g/licenses/by/3.0/igo/. By using he con en o his publica ion, you ag ee o be bound
by he e ms o his license. Fo a ibu ion, ansla ions, adap a ions, and pe missions, please ead he p o isions
and e ms o use a h ps://www.adb.o g/ e ms-use#openaccess.
This CC license does no apply o non-ADB copy igh ma e ials in his publica ion. I he ma e ial is a ibu ed
oano he sou ce, please con ac he copy igh owne o publishe o ha sou ce o pe mission o ep oduce i .
ADB canno be held liable o any claims ha a ise as a esul o you use o he ma e ial.
Please con ac pubsma ke [email protected] g i you ha e ques ions o commen s wi h espec o con en , o i you wish
oob ain copy igh pe mission o you in ended use ha does no all wi hin hese e ms, o o pe mission o use
heADB logo.
Co igenda o ADB publica ions may be ound a h p://www.adb.o g/publica ions/co igenda.
No e:
ADB ecognizes “China” as he People’s Republic o China.
ABSTRACT
By cap u ing he adop ion o digi al se ices and applica ions as well as digi al indus y
de elopmen , his pape cons uc s a comp ehensi e index o measu e digi aliza ion a
he ci y le el in he People’s Republic o China (PRC) and in es iga es he impac o
digi aliza ion on income inequali y using household-le el da a in he PRC. The indings
e eal ha a one-uni ad ancemen in he digi aliza ion index signi ican ly educes he
income gap by 1.83% o an a e age household. This mi iga ing e ec emains
s a is ically and economically signi ican a e add essing endogenei y and obus ness
conce ns. The impac is mo e p onounced in less-de eloped a eas and among
households wi h lowe educa ion le els. Fu he analysis shows ha digi aliza ion na ows
he income gap by inc easing ea nings mo e o lowe -income households, p ima ily
h ough enhanced employmen and in es men oppo uni ies. Addi ionally, digi aliza ion
boos s business income o en ep eneu ial households. These indings p o ide aluable
policy insigh s, sugges ing ha de eloping economies can educe income inequali y by
p omo ing digi aliza ion, suppo ing digi al- ela ed job c ea ion, and enhancing inancial
li e acy.
Keywo ds: digi aliza ion, inclusi eness, income inequali y
JEL codes: D30, O10, O30
1. In oduc ion
Income inequali y, which is p e alen in bo h de eloping and de eloped coun ies, is a
pe sis en phenomenon and a undamen al issue o conce n. Acco ding o he 2022 Wo ld
Inequali y Repo eleased by he Wo ld Inequali y Lab, he iches 10% o he global
popula ion cu en ly accoun s o 52.0% o global income, whe eas he poo es hal o he
popula ion ea ns only 8.5% (Chancel e al. 2023). Rising income inequali y is also seen
in Asia and he Paci ic, hough income le els ha e inc eased along wi h apid economic
g ow h and po e y educ ion (Zhuang 2023).
Wi h he inc easing pene a ion o digi al echnology in he li es o esiden s, digi aliza ion
has been emb aced by go e nmen s because o i s po en ial o p omo ing economic
g ow h and educing inequali y. Digi aliza ion can anscend geog aphical limi s; o e s
oppo uni ies o new businesses and easy access o inance; and enhances accessibili y
o in o ma ion, goods, and se ices o each less-de eloped egions and ulne able
g oups wi h limi ed connec i i y. Howe e , he lack o digi al connec i i y and digi al
li e acy can limi people om enjoying he bene i s o he digi al di idend, which can also
widen he income gap be ween g oups. Thus, he in luence o digi aliza ion on income
dis ibu ion is s ill con o e sial. On he b igh side—digi al echnologies such as big da a,
cloud compu ing, in ech, and online pla o ms—can p omo e economic g ow h and o e
oppo uni ies o ulne able g oups, hus inc easing household income and na owing
income gaps (Ahmed and Al-Roubaie 2013; Asongu and Odhiambo 2019; Faizah,
Yamada, and P a amo 2021; Demi e al. 2022). Meanwhile, on he nega i e side,
oppo uni ies b ough by digi al echnologies may no be equally accessible o all g oups,
leading o a signi ican digi al di ide ha de e io a es income equali y (Guellec and
Pauno 2017; Daud, Ahmad, and Ngah 2021).
Howe e , mos o he exis ing s udies a e based on mac o da a a he coun y le el, and
i is di icul o deeply discuss he e ogenei y and he impac mechanism o digi aliza ion
on income inequali y. Using household mic o su ey da a, his pape empi ically
examines he impac o digi aliza ion on income inequali y a he household le el in he
People’s Republic o China (PRC). The e idence om he PRC o e s use ul policy
implica ions o o he de eloping Asian economies in se e al ways. Fi s , he PRC has
expe ienced apid bu unequal p og ess in digi aliza ion.1 The PRC has a a ie y o
digi aliza ion de elopmen le els ac oss ci ies and egions, which o e s a good sample
o assess how he de elopmen o digi aliza ion can a ec income gaps. Second, despi e
an inc ease in income le els in ecen decades, he e emains signi ican inequali y in
income dis ibu ion ac oss he PRC. Acco ding o da a eleased by he Na ional Bu eau
1 CAICT. 2022. Whi e Pape s Resea ch. caic .ac.cn.com/whi e-pape s- esea ch/202220.
2
o S a is ics o China, he PRC’s Gini coe icien ose om 0.28 in 1981 o 0.47 in 2022,
c ossing he in e na ional wa ning line o 0.40 and exceeding he global a e age. Thus,
he PRC o e s a good sample o es how digi aliza ion can in luence income gaps.
U ilizing mic o household-le el su ey da a om he China Household Finance Su ey
(CHFS) co e ing abou 10,000 andomly selec ed households ac oss app oxima ely 544
ci ies om 2013 o 2019, his s udy measu es income inequali y using he Kakwani index
a he household le el, ollowing Bosse and D’Amb osio (2006) and Beck, Demi güç-
Kun , and R. Le ine (2007). Following Bukh and Heeks (2018), his s udy cons uc s a
digi aliza ion index a he ci y le el o cap u e a ious aspec s o digi al echnology
de elopmen , including digi al indus y de elopmen , and adop ion o digi al de ices and
applica ions. The analysis esul s show ha a ci y’s digi aliza ion de elopmen can
signi ican ly educe income gaps a household le el. In pa icula , a one uni inc ease in
he digi aliza ion index measu e is associa ed wi h a 0.0183 educ ion in Kakwani index
o an a e age household, which is 1.94% o he sample mean. This means, o example,
when he a e age digi aliza ion de elopmen du ing he sample pe iod inc eases om
1.06 in Hohho o 1.68 in Chengdu, he household-le el income inequali y na ows by
0.011, o 2.04% o an a e age household in he sample. Mo eo e , his impac is mo e
p onounced in less-de eloped egions in he coun y and households wi h ela i ely lowe
educa ion le els.
To add ess he possible endogenei y issues, we employed he ins umen a iable
app oach. This ins umen a iable is de i ed by g ouping ci ies wi h simila economic
p o iles and calcula ing he a e age digi aliza ion index o o he ci ies wi hin he same
g oup. Such an ins umen is ela ed o digi aliza ion de elopmen as i cap u es economic
de elopmen o a ypical ci y in he g oups bu does no necessa ily link o he income
inequali y o local households o he ci y ha is no pa o he ins umen . This a iable is
a s ong ins umen , and he ins umen ed digi aliza ion o a ci y is s ill signi ican ly and
nega i ely ela ed o income gaps in local households. We also conduc ed a ious
obus ness checks and ound obus esul s on he nega i e impac o digi aliza ion o a
ci y on i s households’ income inequali y.
To be e unde s and he unde lying d i e o his e ec , his s udy explo es po en ial
mechanisms. The e is e idence o show ha digi aliza ion is able o inc ease income
le els, especially in less-de eloped egions and o lowe -educa ed popula ions. The
inc ease in income is ela ed o employmen , in es men , and business income. We ind
e idence ha digi aliza ion is ela ed o mo e job oppo uni ies, especially o less-
educa ed popula ions and in less-de eloped egions, bu we did no ind signi ican
inc eases in wages associa ed wi h digi aliza ion. Meanwhile, digi aliza ion also os e s
pa icipa ion in inancial ma ke s, which o e s mo e in es men oppo uni ies o
3
households. And he in es men oppo uni ies ia access o inancial ma ke s apply o all
popula ions, ega dless o local de elopmen s a us o educa ion le els.
This s udy con ibu es o he exis ing li e a u e in he ollowing way. Fi s , his pape adds
o ex an li e a u e wi h a measu e ha cap u es mo e comp ehensi e in o ma ion on he
le el o digi aliza ion de elopmen . Exis ing s udies end o measu e digi aliza ion
ocusing on ce ain digi al echnologies, such as digi al inance (Daud, Ahmad, and Ngah
2021; Yang and Zhang 2022) and a i icial in elligence (S e enson 2019), o accessibili y
o digi al in as uc u e and de ices, such as mobile phone and in e ne pene a ion a es
(Asongu and Nwachukwu 2018, Wang and Xu 2023). Such measu es may no ha e
comp ehensi ely gauged he digi al ans o ma ion in ecen yea s. The measu e adop ed
in his s udy adds o exis ing measu es wi h in o ma ion on de elopmen o he digi al
indus y, which p esen s a mo e comp ehensi e pic u e o digi aliza ion in economic
ac i i ies.
Second, cu en li e a u e ends o show e idence a he coun y le el, while his pape
p o ides consis en and obus empi ical e idence a he household le el. Mo eo e , he
empi ical es ima es ob ained om he ixed-e ec s model a e con incing by con olling
o he e ogenei y ac oss he panels.
Las ly, his s udy lends suppo o he inclusi eness o digi aliza ion om he pe spec i e
o income inequali y in less-de eloped egions and o less-educa ed popula ions, and i
un eils he possible wo king channels o his e ec . Such e idence has policy implica ions
o pee de eloping economies in Asia and he Paci ic by demons a ing ha digi aliza ion
may wo k well as a ool o boos income le els and close income gaps in less-de eloped
a eas by os e ing job c ea ion.
This pape is s uc u ed as ollows. Sec ion 2 e iews he cu en li e a u e and ou lines
he es able hypo hesis. Sample cons uc ion and esea ch me hods a e ou lined in
sec ion 3. Sec ion 4 discusses he empi ical esul s, and sec ion 5 discusses he indings
and policy implica ions.
2. Li e a u e Re iew
The li e a u e on he impac o echnological ad ances on income dis ibu ion, as well as
income inequali y, is inconclusi e and o e s many di e en explana ions and conclusions.
One s eam o li e a u e a gues ha digi aliza ion helps na ow income inequali y ia he
ollowing h ee channels. Fi s , digi aliza ion enhances connec i i y by b eaking
geog aphical cons ain s and making i ual ma ke s and emo e wo kplaces possible
4
(Shaikh and Ka jaluo o 2015). Enhanced connec i i y opens new business oppo uni ies
and enables new jobs o popula ions acing cons ain s such as physical loca ion and
educa ion le els, and i helps o op imize esou ce alloca ion ac oss di e en egions. Fo
example, eme ging digi al business models—such as e-comme ce, sho ideos, and
online li e s eaming—p o ide new job and income oppo uni ies o people om emo e
a eas and low-income g oups (A asoy 2013, Dau h e al. 2017). Second, digi aliza ion
enables g ea e access o se ices such as inance o ulne able g oups and popula ions
who a e domiciling in a eas wi h limi ed inancial esou ces and se ices (Jack and Su i
2011; Tchamyou, E eyge s, and D. Cassimon 2019; Demi e al. 2022). By enhancing
access o inancial se ices, digi aliza ion enables mo e in es men oppo uni ies and also
lowe s inancing cos s, hus na owing he income gap (Hong, Lu, and Pan 2020; Asongu
and Odhiambo 2019). Thi d, digi aliza ion educes in o ma ion ba ie s o ulne able
g oups and acili a es in o ma ion acquisi ion and analysis o in es men , business, and
job oppo uni ies, which could inc ease hei income and na ow he income gap. Fo
example, Hong, Lu, and Pan (2020) ind ha digi aliza ion allows esiden s o be exposed
o a ious digi al applica ion scena ios, which can e ec i ely alle ia e in o ma ion
asymme y and educe isk a e sion among esiden s, he eby p omo ing hei
pa icipa ion in inancial ma ke in es men ; his e ec is mo e isible among ulne able
g oups in alle ia ing income inequali y h ough income gains.
Meanwhile, o he schola s a gue ha digi aliza ion can widen income inequali y (Yeo,
Hwang, and Lee 2023). This is because he Schumpe e ian ype o c ea i e des uc ion
b ough by echnological ad ances does no b ing he same oppo uni ies o di e en
g oups. Fo example, high-income and be e -educa ed people ha e be e chances o
apply he la es digi al echnology and enjoy digi al di idends, hus widening he gap wi h
low-income and less-educa ed g oups (Acemoglu and Res epo 2020; Daud, Ahmad, and
Ngah 2021). Mo eo e , skilled wo ke s may adap o new echnology and lea n new
knowledge mo e quickly compa ed o unskilled wo ke s, leading o a widening o he
income gap (Guellec and Pauno 2017; Acemoglu and Res epo 2018; Tica, Globan, and
A čabić 2022). In addi ion, algo i hmic mechanisms, based on digi al echnologies such
as big da a and machine lea ning, may s eng hen he supply disc imina ion in c edi
(Philippon 2016)— ha is, highe -income and be e -educa ed people ha e g ea e
access o low-cos inancial p oduc s and digi al echnologies, he eby con ibu ing o a
widening o he income gap.
The e is also e idence ha a nonlinea ela ionship may exis be ween income inequali y
and echnological p og ess. On he one hand, acco ding o he echnological Kuzne s
cu e hypo hesis p oposed by Kim (2012), income inequali y i s ises and hen declines
wi h echnological p og ess (i.e., an in e se U-shaped ela ionship). In he ini ial s age o
echnological inno a ion, he cos o adop ing a new echnology is ela i ely la ge and
11
as a pe cen age o GDP. Wi hin each g oup, we u he di ide ci ies in o h ee g oups by
anking hei GDP. This exe cise esul s in nine g oups o ci ies. Wi hin each g oup, we
exclude he ci y whe e a household is loca ed and compu e he mean alue o he
digi aliza ion index o o he ci ies in he g oup o be he ins umen a iable. To ensu e he
obus ness o he esul s, we also epea his exe cise using a 5 by 5 g ouping and ob ain
an ins umen a iable based on he 25 g oups (5 x 5) o ci ies. To u he exclude possible
simila i y o ci ies om he same p o ince, we econs uc ed he ins umen al a iable by
excluding ci ies om he same p o ince in so ed ci y g oups as an ex a obus ness es .
The ins umen ha cap u es digi aliza ion o simila ci ies may be ela ed o he
digi aliza ion le el o a ci y bu may no a ec he household income dis ibu ion o his
ci y di ec ly.
Table 4: Digi aliza ion and Income Inequali y
(1)
(2)
KKWN Index
KKWN Index
Digi aliza ion index
–0.0183***
–0.0220***
(0.0056)
(0.0067)
Digi aliza ion index sq.
0.0029
(0.0022)
Age
–0.0075***
–0.0075***
(0.0013)
(0.0013)
Age sq.
0.0078***
0.0077***
(0.0012)
(0.0012)
Household size
0.0087***
0.0087***
(0.0018)
(0.0018)
Child en a io
–0.0437***
–0.0432***
(0.0148)
(0.0148)
Old a io
–0.0195**
–0.0196**
(0.0096)
(0.0096)
Unheal hy a io
0.0416***
0.0414***
(0.0100)
(0.0100)
Ln (household o al asse )
–0.0251***
–0.0251***
(0.0010)
(0.0010)
Ln (pe capi a GDP)
0.0333*
0.0364*
(0.0196)
(0.0198)
Yea FE
Y
Y
Household FE
Y
Y
N
107671
107671
R
2
0.014
0.014
Adj. R2
0.014
0.014
Con inued on he nex page
12
FE = ixed e ec s, GDP = g oss domes ic p oduc , N = numbe o obse ed alues.
No es: *, **, and *** deno e signi icance a he 10%, 5%, and 1% le els, espec i ely. The s anda d e o s
a e clus e ed a he ci y-yea le el o ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
The esul s using he ins umen al a iable a e epo ed in Table 5. The esul s o he i s -
s age eg ession in columns (1) and (3) sugges ha he ins umen al a iable is
signi ican ly associa ed wi h digi aliza ion o he ci y. And he second-s age esul s shown
in columns (2) and (8) demons a e ha he ins umen ed digi aliza ion is s ill signi ican ly
and nega i ely associa ed wi h household income inequali ies, con i ming he indings
om he baseline eg ession in Table 4. To mi iga e he in luence o linea end ac o s,
we also use i s -o de di e ences o a iables. The esul s om column (9) o Table 5
indica e ha , e en a e i s -o de di e encing, he digi aliza ion index emains signi ican
and nega i ely co ela ed wi h income inequali y.
To ensu e obus ness o he abo e esul s, we conduc a ious addi ional es s. Column
(1) o Table 6 epo s he i s se o obus ness check whe e we eplace he digi aliza ion
index om p e ious pe iod in he baseline wi h he digi aliza ion in he cu en pe iod. As
shown, he coe icien emains signi ican and nega i e o he Kakwani index, which is
consis en wi h he inding s a ed ea lie . O he obus ness checks include: 1) compu ing
he Kakwani index using o al household income a he han pe capi a income o he
households in Column (2); 2) emo ing he op and bo om 5% o households by pe
capi a income, hen ecalcula ed he Kakwani index o he emaining households in
column (3); 3) excluding households ha su e ed om majo e en s in he ecen 5 yea s
om ou sample in column (4); and 4) excluding u al samples ha may ha e a la ge
gap wi h u ban households o a oid he esul s being d i en by u al households in column
(5). O e all, hese obus ness checks gene a e consis en indings ha con i m ha
digi aliza ion is signi ican ly and nega i ely co ela ed wi h income inequali y.
To u he unde s and he sou ce o he impac o digi aliza ion on income inequali y, we
un sepa a e eg ession o indi idual digi aliza ion indica o s on income inequali y and
epo he esul s in Table 7. As shown, he esul s a e la gely d i en by de elopmen in
digi al indus y, such as he pa en licenses in he 5G indus y, he numbe o digi al
en e p ises, and he numbe o digi al en e p ises.
The li e a u e shows ha digi aliza ion may ha e a he e ogeneous impac on households
in di e en economic de elopmen egions and wi h di e en educa ions le els. (Knigh
2014, Zhang 2021, and Li, Wu, and Xiao 2020) In empi ical es s, we di ide he sample
in o di e en subsamples and explo e how digi aliza ion can a ec income inequali y
di e en ly. Columns (1) and (2) o Table 8 epo he impac s o digi aliza ion on income
inequali y in he mo e ad anced egions o he eas e n p o inces and he ela i ely less-
13
de eloped cen al and wes e n a eas p o inces. As shown, while he coe icien s o
digi aliza ion a e signi ican ly and nega i ely associa ed wi h he Kakwani index in bo h
subsamples, he impac is mo e p onounced o less de eloped cen al–wes e n a ea. In
columns (3) and (4) o Table 8, we examine how digi aliza ion may a ec household
income inequali y wi h di e en educa ion le els. In doing his, we di ide he sample in o
wo g oups based on he median o a e age schooling yea s o he household labo o ce
in 2017. The esul s show ha he e ec o digi aliza ion on he Kakwani index is mo e
p onounced in households wi h less educa ion han in households wi h mo e educa ion.
O e all, Table 8 shows ha he impac o digi aliza ion on income inequali y is mo e
p onounced in less-de eloped a eas and o households wi h lowe educa ion le els,
u he shedding ligh on he inclusi eness bene i o digi aliza ion.
4.2 How Does Digi aliza ion Na ow Income Inequali y?
The p e ious session shows ha digi aliza ion can na ow income inequali y, especially
in less-de eloped a eas and among less-educa ed households. I is na u al o ask how
digi aliza ion can educe income inequali y. This sec ion answe s he ques ion by looking
a how digi aliza ion a ec s income le els o households. Table 9 epo s he es ima ed
esul s o digi aliza ion on household income, by eplacing income inequali y wi h
household income in he baseline model o equa ion (2). I shows ha ci y-le el
digi aliza ion is signi ican ly and posi i ely associa ed wi h household income ac oss
di e en quan ile household income le els. The e ec is bo h s a is ically and
economically signi ican . Mo eo e , he magni ude o coe icien s inc eases as income
le el dec eases. O e all, Table 9 shows ha digi aliza ion posi i ely a ec s household
income, and he impac is mo e p onounced o lowe income households, hus na owing
income inequali y.
To check u he whe he he inding is consis en wi h he p e ious session, we epea
he es in Table 9 by examining he esul s in less-de eloped a eas and less-educa ed
households. The esul s a e epo ed in Table 10. Consis en ly, we ind ha a s onge
posi i e impac o digi aliza ion on household income in less de eloped cen al–wes e n
egions in columns (1) and (2), and o less-educa ed households in columns (3) and (4).
These addi ional esul s con i m ha he impac o digi aliza ion on household income is
mo e p onounced o disad an aged g oups, hus educing income inequali y.
4.3 Wo king Mechanisms
In his sec ion, we u he explo e he possible wo king mechanism in which digi aliza ion
may a ec income dis ibu ion. Fi s , we b eakdown household income in o h ee majo
sou ces: (i) wage income om employmen , (ii) business income om comme cial
ac i i ies, and (iii) in es men income om in es ing in inancial ma ke s. Table 11
p esen s he esul s on how digi aliza ion is ela ed o di e en ypes o household income,
14
showing ha digi aliza ion is signi ican ly and posi i ely ela ed o all h ee ypes o
household incomes.
To u he in es iga e how digi aliza ion could boos household wage income, we examine
he impac o digi aliza ion on employmen oppo uni ies. Table 12 epo s he es ima ed
ela ionship be ween he digi aliza ion le el o a ci y and household labo o ce
pa icipa ion using a P obi model. Resul s show ha digi aliza ion is posi i ely and
signi ican ly associa ed wi h household-le el employmen s a us, and he impac is mo e
p onounced o households in he less-de eloped egions and wi h lowe -educa ion le els.
Besides employmen s a us, i is also in e es ing o know wha ypes o jobs a e ela ed
o digi aliza ion. We analyze he ela ionship be ween digi aliza ion and employmen ypes
and epo he esul s in Table 13. As shown, digi aliza ion is signi ican ly ela ed o jobs
in p i a e companies and MSME business. This implies ha he ole o digi aliza ion
enables lexible and new jobs in he p i a e sec o as well as MSMEs.
To unde s and how digi aliza ion boos s household in es men income, we examine he
associa ion be ween digi aliza ion and new inancial ma ke pa icipa ion o households.
A household is conside ed a pa icipan in he inancial ma ke s i ha household holds
any inancial asse s in he o m o s ocks, unds, bonds, de i a i es, o gold. I a household
did no pa icipa e in he inancial ma ke in he p e ious yea bu pa icipa ed in he
cu en yea , i is de ined as a new pa icipan o he inancial ma ke . As shown in Table
14, digi aliza ion is no signi ican ly posi i ely co ela ed wi h o e all inancial ma ke
pa icipa ion in he ull sample. Howe e , among households in he less de eloped
cen al–wes e n egions and hose wi h lowe educa ion le els, digi aliza ion is
signi ican ly and posi i ely co ela ed wi h new inancial ma ke pa icipa ion. This
indica es ha digi aliza ion o e s new income oppo uni ies by enabling accessibili y o
inancial ma ke s and in es men oppo uni ies o less-p i ileged households. This
p o ides addi ional e idence o he inclusi e cha ac e is ics o digi aliza ion.
5. Conclusion
This s udy examined how digi aliza ion is associa ed wi h household income inequali y.
By cons uc ing a comp ehensi e index o cap u e digi aliza ion and u ilizing household
le el da a om he CHFS in 2013, 2015, 2017, and 2019, his pape documen s a posi i e
ole o digi aliza ion in educing income inequali y a he household le el. This income
inequali y educ ion impac is mo e p onounced in he less-de eloped cen al–wes e n
egions and among less-educa ed households. This impac on income inequali y emains
s a is ically and economically signi ican a e add essing endogenei y conce ns and a
ew obus ness checks.
15
We u he demons a e how digi aliza ion educes income inequali y. E idence shows
ha digi aliza ion has a signi ican and posi i e impac on household income. The income
p omo ion impac is la ge o lowe -income households and households in less
de eloped a eas and wi h lowe educa ion le els, he eby educing income inequali y.
This u he con i ms he inclusi eness bene i o digi aliza ion. Fu he mo e, e idence
shows ha digi aliza ion boos s household incomes by boos ing wage income, business
income, and in es men income.
This s udy p o ides mic o-le el e idence o policymake s o speed up digi aliza ion and
p omo e income oppo uni ies o disad an aged g oups, na owing income inequali y.
Ou indings imply ha digi aliza ion can p omo e inclusi e de elopmen , wi h a la ge
ma ginal e ec o less-de eloped a eas and households wi h lowe le els o educa ion.
P omo ing digi aliza ion in less-de eloped a eas will help boos inclusi eness.
Policymake s may s eng hen digi aliza ion, pa icula ly in emo e and unde de eloped
a eas, o deli e inclusi e bene i s.
16
Table 5: Add essing Endogenei y in he Es ima ion
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Fi s s age
Second
s age
Fi s s age
Second
s age
Fi s s age
Second
s age
Fi s s age
Second
s age
Fi s -o de
di e ence
es ima ion
Digi aliza ion index
KKWN
index
Digi aliza ion
index
KKWN
index
Digi aliza ion
index
KKWN
index
Digi aliza ion
index
KKWN
index
Di e ence
o KKWN
index
Digi aliza ion index
in 9 subg oups
0.6213***
(0.0040)
Digi aliza ion index
in 25 subg oups
0.7932***
(0.0028)
Digi aliza ion in 9
subg oups
excluding ci ies
om he same
p o ince
0.5433***
(0.0039)
Digi aliza ion index
in 25 subg oups
excluding ci ies
om he same
p o ince
0.6787***
(0.0029)
Ins umen ed
Digi aliza ion index
–0.0190***
–0.0191***
–0.0202**
–0.0323***
(0.0038)
(0.0039)
(0.0088)
(0.0088)
Di e ence o
Digi aliza ion index
–0.0155*
(0.0081)
Yea FE
Y
Y
Y
Y
Y
Y
Y
Y
Y
Household FE
Y
Y
Y
Y
Y
Y
Y
Y
Y
Con ols
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
107,671
107,671
107,671
107,671
107,671
107,671
107,671
107,671
48,957
R2
0.703
0.792
0.692
0.757
0.011
Adj. R2
0.703
0.792
0.692
0.757
0.010
FE = ixed e ec s, N = numbe o obse ed alues.
No es: *, **, and *** deno e signi icance a he 10%, 5%, and 1% le els, espec i ely. The s anda d e o s a e clus e ed a he ci y-yea le el o ensu e he e oscedas ici y
obus ness.
Sou ce: Au ho s’ calcula ions.
17
Table 6: Robus ness Check—Digi aliza ion and Income Inequali y
(1) (2) (3) (4) (5)
KKWN Index Based on
Income pe
Capi a
Based on
To al
Income
Remo ed 5% Top
and Bo om
Income o
Households
Excluded
Households ha
Su e om Majo
E en s
Exclude
Ru al
Sample
Digi aliza ion index –0.0119** –0.0189*** –0.0091*** –0.0234***
(0.0052) (0.0056) (0.0033) (0.0062)
Digi aliza ion in
con empo a y
pe iod
–0.0165**
(0.0088)
Yea FE Y Y Y Y Y
Household FE Y Y Y Y Y
Con ols Y Y Y Y Y
N 82,034 107,541 97,894 98,995 72,417
R2 0.008 0.015 0.004 0.005 0.015
Adj. R2 0.008 0.015 0.004 0.004 0.015
FE = ixed e ec s, N = numbe o obse ed alues.
No es: ** and *** deno e signi icance a he 5% and 1% le els, espec i ely. The s anda d e o s a e clus e ed a he
ci y-yea le el o ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
18
Table 7: Digi aliza ion indica o s and Household Income Inequali y
(1) (2) (3) (4) (5) (6)
KKWN Index Mobile
Telephone
Subsc ibe s
Telecommunica io
n se ice e enue
Digi al
Paymen
Co e age
Numbe o
digi al
en e p ises
Digi al indus y
employmen
Pa en
licenses
in he 5G
indus y
Subindex o
Digi aliza ion
index
–0.0156 0.0106 –0.0494 –0.0802*** –0.0742*** –0.1427**
(0.0297) (0.0276) (0.0696) (0.0239) (0.0230) (0.0615)
Yea FE Y Y Y Y Y Y
Household FE Y Y Y Y Y Y
Con ols Y Y Y Y Y Y
N 107,671 107,671 107,671 107,671 107,671 107,671
R2 0.013 0.013 0.013 0.014 0.014 0.013
Adj. R2 0.013 0.013 0.013 0.014 0.014 0.013
FE = ixed e ec s, N = numbe o obse ed alues.
No es: *, **, and *** deno e signi icance a he 10%, 5%, and 1% le els, espec i ely. The s anda d e o s a e clus e ed a he ci y-
yea le el o ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
Table 8: Digi aliza ion and Income Inequali y—He e ogenei y Tes s
(1) (2) (3) (4)
KKWN Index Eas e n
Region
Cen al and Wes e n
Regions
High Educa ion
Households
Low Educa ion
Households
Digi aliza ion index –0.0108* –0.0335*** –0.0178*** –0.0221***
(0.0064) (0.0116) (0.0062) (0.0077)
Yea FE Y Y Y Y
Household FE Y Y Y Y
Con ols Y Y Y Y
N 53,562 54,109 43,961 33,411
R2 0.015 0.014 0.003 0.002
Adj. R2 0.015 0.013 0.003 0.002
FE = ixed e ec s, N = numbe o obse ed alues.
No es: * and *** deno e signi icance a he 10% and 1% le els, espec i ely. The s anda d e o s a e clus e ed a he ci y-yea le el o
ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
19
Table 9: Digi aliza ion and Household Income
(1) (2) (2) (3) (4) (5)
Ln (household
income pe
capi a)
Full sample Q10 Q25 Q50 Q75 Q90
Digi aliza ion
index
0.0622*** 0.2550*** 0.1304*** 0.0621*** 0.0623*** 0.0567***
(0.0142) (0.0268) (0.0154) (0.0107) (0.0100) (0.0140)
Yea FE Y Y Y Y Y Y
Household FE Y Y Y Y Y Y
Con ols Y Y Y Y Y Y
N 106,226 106,098 106,098 106,098 106,098 106,098
R2 0.032
Adj. R2 0.032
FE = ixed e ec s, N = numbe o obse ed alues.
No es: *** deno es he signi icance a he 1% le el. The s anda d e o s a e clus e ed a he ci y-yea le el o ensu e
he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
Table 10: Digi aliza ion and Household Income—He e ogenei y Tes s
(1) (2) (3) (4)
Ln (household income
pe capi a)
Eas e n
Region
Cen al and
Wes e n Regions
High Educa ion
Households
Low Educa ion
Households
Digi aliza ion index 0.0399** 0.0994*** 0.0371*** 0.0899***
(0.0161) (0.0333) (0.0128) (0.0201)
Yea FE Y Y Y Y
Household FE Y Y Y Y
Con ols Y Y Y Y
N 52,888 53,338 43,577 32,764
R2 0.064 0.060 0.064 0.065
Adj. R2 0.064 0.060 0.064 0.064
FE = ixed e ec s, N = numbe o obse ed alues.
No es: *, **, and *** deno e signi icance a he 10%, 5%, and 1% le els, espec i ely. The s anda d e o s a e clus e ed
a he ci y-yea le el o ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
20
Table 11: Digi aliza ion and he Ca ego ies o Household Income
(1) (2) (3)
Ln (wage income pe
capi a)
Ln (business income pe
capi a)
Ln (in es men income pe
capi a)
Digi aliza ion index 0.0698*** 0.1052** 0.3197***
(0.0179) (0.0484) (0.0727)
Yea FE Y Y Y
Household FE Y Y Y
Con ols Y Y Y
N 106,098 15,662 106,098
R2 0.041 0.059 0.296
Adj. R2 0.041 0.058 0.296
FE = ixed e ec s, N = numbe o obse ed alues.
No es: *, **, and *** deno e signi icance a he 10%, 5%, and 1% le els, espec i ely. The s anda d e o s a e clus e ed
a he ci y-yea le el o ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
Table 12: Digi aliza ion De elopmen and Employmen S a us
(1) (2) (3) (4) (5)
Employmen S a us Whole Sample Eas e n
Region
Cen al and Wes e n
Regions
High Educa ion
G oup
Low Educa ion
G oup
Digi aliza ion index 0.0184*** 0.0142* 0.0357** –0.0023 0.0277***
(0.0066) (0.0074) (0.0144) (0.0070) (0.0092)
Yea FE Y Y Y Y Y
Indi idual FE Y Y Y Y Y
Con ols Y Y Y Y Y
N 149,231 75,917 73,314 74,586 75,848
R2 0.008 0.008 0.008 0.016 0.005
Adj. R2 0.008 0.008 0.008 0.016 0.005
FE = ixed e ec s, N = numbe o obse ed alues.
No es: *, **, and *** deno e signi icance a he 10%, 5%, and 1% le els, espec i ely. The s anda d e o s a e clus e ed a he ci y-
yea le el o ensu e he e oscedas ici y obus ness.
Sou ce: Au ho s’ calcula ions.
