Digital monitoring technology and air quality: Evidence from the People's Republic of China

Liang, Pinghan; Liu, Yadi; Tian, Shu
Wo king Pape
Digi al moni o ing echnology and ai quali y: E idence
om he People's Republic o China
ADB Economics Wo king Pape Se ies, No. 788
P o ided in Coope a ion wi h:
Asian De elopmen Bank (ADB), Manila
Sugges ed Ci a ion: Liang, Pinghan; Liu, Yadi; Tian, Shu (2025) : Digi al moni o ing echnology and
ai quali y: E idence om he People's Republic o China, ADB Economics Wo king Pape Se ies, No.
788, Asian De elopmen Bank (ADB), Manila,
h ps://doi.o g/10.22617/WPS250251-2
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/322388
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. 788
July 2025
Digi al Moni o ing Technology and Ai Quali y
E idence om he People’s Republic o China
This pape empi ically in es iga es he impac o digi al en i onmen al moni o ing echnologies on ai
quali y in he People’s Republic o China using public p ocu emen da a. The indings show ha ollowing
he adop ion o digi al en i onmen al moni o ing echnology: (i) ci y-le el PM2.5 concen a ions exhibi
a sizeable educ ion, (ii) egula o y en o cemen imp o es and hea ily pollu ing i ms exi mo e, (iii) g een
inno a ion is enhanced h ough mo e accu a e pollu an iden i ica ion, and (i ) he e ec i eness o hese
echnologies depends on he ex en o in o ma ion disclosu e and public pa icipa ion.
Abou he Asian De elopmen Bank
ADB is a leading mul ila e al de elopmen bank suppo ing inclusi e, esilien , and sus ainable g ow h ac oss
Asia and he Paci ic. Wo king wi h i s membe s and pa ne s o sol e complex challenges oge he , ADB
ha nesses inno a i e inancial ools and s a egic pa ne ships o ans o m li es, build quali y in as uc u e,
and sa egua d ou plane . Founded in 1966, ADB is owned by 69 membe s—50 om he egion. DIGITAL MONITORING
TECHNOLOGY AND AIR QUALITY
EVIDENCE FROM THE PEOPLE’S REPUBLIC OF CHINA
Pinghan Liang, Yadi Liu, and Shu Tian
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
Pinghan Liang, Yadi Liu, and Shu Tian
No. 788 | July 2025
Pinghan Liang ([email p o ec ed])
and Yadi Liu ([email protected]) a e p o esso s
a he School o Go e nmen , Sun Ya -sen Uni e si y.
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.
Digi al Moni o ing Technology and Ai Quali y:
E idence om he People’s Republic o China
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. WPS250251-2
DOI: h p://dx.doi.o g/10.22617/WPS250251-2
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
oano 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
oob 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
heADB logo.
Co igenda o ADB publica ions may be ound a h p://www.adb.o g/publica ions/co igenda.
No es:
In his publica ion, “$” e e s o Uni ed S a es dolla s.
ADB ecognizes “China” as he People’s Republic o China.
ABSTRACT
Since he ea ly 21s cen u y, ai quali y conce ns in he People’s Republic o China (PRC) ha e
ga ne ed signi ican a en ion om bo h he public au ho i ies and socie y. This s udy in es iga es
he e ec s o digi al en i onmen al moni o ing echnology on ai pollu ion. Speci ically, we explo e
he da a om go e nmen p ocu emen o digi al en i onmen al moni o ing echnologies o e he
2014–2019 pe iod. The baseline esul s indica e ha on a e age, each addi ional en i onmen al
con ac pe 100,000 esiden s signed by go e nmen s is associa ed wi h an 8-pe cen age poin
educ ion in ci y PM2.5 le els. This e ec a ises om mo e accu a e pollu an iden i ica ion, which
s eng hens en o cemen o en i onmen al egula ions, acili a ing any necessa y ansi ion and,
whe e applicable, o de ly exi o hea ily pollu ing en e p ises, and os e s g een inno a ion.
Fu he , his e ec exhibi s egional a ia ion in he ex en o en i onmen al conce n and he le el
o in o ma ion disclosu e. The esul s sugges ha echnology-d i en en i onmen al go e nance,
suppo ed by public engagemen and policy amewo ks, plays a c ucial ole in enhancing ai
quali y in he PRC.
Keywo ds: digi al en i onmen al moni o ing, PM2.5, public moni o ing, in o ma ion disclosu e
JEL codes: O18, O33, Q50
_______________________
This s udy was p epa ed as a backg ound ma e ial o Ha nessing Digi al T ans o ma ion o Good: Asian
De elopmen Policy Repo and unded by he Republic o Ko ea e-Asia and Knowledge Pa ne ship Fund
TA 6920 (P omo ing Digi aliza ion o G een and Inclusi e G ow h in Asia).

1. INTRODUCTION
The e is inc easing in e es om he public, ci y adminis a o s, and egula o s in ai quali y, as
sa e, clean ai is essen ial o ci ies and communi ies o h i e (Smi h and Doe 2020). In 2010, he
People’s Republic o China (PRC) became he wo ld’s second-la ges economy; howe e , his
g ow h came a a signi ican en i onmen al cos . Regions such as Beijing, Tianjin, and Hebei
P o ince ha e equen ly su e ed om se e e smog, wi h ai pollu ion le els eaching haza dous
le els. These ala ming indica o s ha e p omp ed he PRC go e nmen o con inuously p io i ize
en i onmen al issues, implemen ing a se ies o policy e o ms and echnological ad ancemen s
aimed a en i onmen al imp o emen .
In 2013, he PRC go e nmen announced an in es men o CNY1.7 illion ($277 billion)
o e he nex 5 yea s o comba he dange ous and apidly wo sening ai pollu ion esul ing om
he PRC’s apid economic g ow h (Wu 2013). This ini ia i e includes ac i ely le e aging digi al
echnologies o enhance en i onmen al go e nance e iciency, as demons a ed by he
widesp ead implemen a ion o digi al en i onmen al moni o ing sys ems in u ban a eas (Schlæge
and Zhou 2019). Digi al echnology o e s unique ad an ages in p edic ing pollu ion le el changes
and acking dispe sion pa hways. This echnology add esses he p incipal-agen in o ma ion
asymme y be ween he go e nmen and pollu e s. By p o iding eal- ime, anspa en da a, i
educes unce ain y and e ec i ely lowe s pollu ion le els. Despi e he nume ous ad an ages o
digi al moni o ing echnology and i s g owing ole in en i onmen al go e nance in he PRC, he
impac o digi al moni o ing echnology on ai pollu ion managemen emains insu icien ly
unde s ood and wa an s u he in es iga ion.
We ob ained en i onmen al moni o ing p ocu emen con ac s om he Chinese
Go e nmen P ocu emen Ne wo k o he yea s 2014–2019, ma ched hem wi h ci ies based on
he add esses o he pu chasing pa ies o examine he impac o go e nmen digi al moni o ing
echnology on ci y PM2.5 le els. The baseline eg ession esul s indica e ha , on a e age, each
addi ional en i onmen al con ac pe 100,000 esiden s co esponds o an es ima ed 8%
2
educ ion in PM2.5 le els. This e ec a ises om mo e accu a e pollu an iden i ica ion, which
s eng hens go e nmen al en o cemen o en i onmen al egula ions, expedi es he exi o hea ily
pollu ing en e p ises, and os e s g een inno a ion.
Fi s , ou esea ch di ec ly b oadens and supplemen s discussions on en i onmen al
moni o ing echnology wi hin he con ex o he PRC’s digi aliza ion (Jin e al. 2020, Schlæge and
Zhou 2019, Hao 2018), emphasizing he signi icance o echnology-aided decision-making.
T adi ional en i onmen al go e nance amewo ks, such as ma ke -based egula ions, ha e
posi i ely impac ed en i onmen al go e nance (Mao and Wang 2016, Zhang e al. 2019).
Howe e , hey o en exhibi empo al delays. Ou indings indica e ha digi al moni o ing
echnology signi ican ly enhances he e iciency o ai pollu ion egula ion, wi h i s impac s ongly
co ela ed o he densi y o moni o ing de ices—a ela ionship con i med wi hin indus ial pa ks
(Huang e al. 2019). We ex end his conclusion o ci y le el, inding ha highe le els o digi al
moni o ing in as uc u e e ec i ely educe o e all ai pollu ion. This inding ca ies subs an ial
implica ions o o he ace s o he PRC’s en i onmen al go e nance, sugges ing ha in ensi ied
moni o ing e o s can suppo b oade en i onmen al imp o emen goals.
Second, some s udies ha e discussed he impac o en i onmen al moni o ing s a ions
unde cen al go e nmen egula ion on ai quali y. On he one hand, moni o ing s a ions ha e
imp o ed ai quali y, bu his e ec appea s o be localized (Yang e al. 2024; He e al. 2020). On
he o he hand, en i onmen al moni o ing s a ions signi y a shi in en i onmen al moni o ing
au ho i y, and he inc ease in digi aliza ion and in o ma ion disclosu e educes he po en ial o
local da a manipula ion (G eens one e al. 2022). Repo ed ai pollu ion concen a ions showed
an immedia e inc ease ollowing he es ablishmen o moni o ing s a ions. This also mo i a ed
local go e nmen o icials o add ess en i onmen al pollu ion. No s udies, o he bes o ou
knowledge, ha e explo ed how local go e nmen s eac o u ban ai quali y when da a egula ion
au ho i y is ans e ed upwa ds. This issue is c i ical because inc eased in o ma ion disclosu e
no only ele a es public and go e nmen awa eness o en i onmen al pollu ion bu also o ces
3
local go e nmen s o adop mo e e icien pollu ion con ol measu es in esponse o public
demands. Consequen ly, he expansion o he ai moni o ing ne wo k h ough he p ocu emen o
ad anced digi al moni o ing echnologies becomes a key s a egy o local go e nmen s. Ou
s udy e eals ha ci ies wi h highe le els o en i onmen al in o ma ion disclosu e expe ience a
signi ican ly g ea e educ ion in ai pollu ion because o he impac o digi al moni o ing.
Las ly, ou esea ch unde sco es ha echnology, while o e ing nume ous go e nance
bene i s, canno independen ly esol e complex en i onmen al issues. Digi al moni o ing
echnology se es p ima ily as an in o ma ion p o ide in ai pollu ion managemen ; howe e , i s
e ec i eness depends on in eg a ion wi hin a b oade sys emic go e nance amewo k. On he
one hand, G eens one e al. (2022) indica e an immedia e and las ing inc ease o epo ed ai
pollu ion le el a e he in oduc ion o au oma ed moni o ing o ai quali y. Yang e al. (2024) show
ha local go e nmen s’ s a egic esponse o he au oma ed moni o s may limi he en o cemen
o en i onmen al egula ion. On he o he hand, Bun aine e al. (2024) show ha ci izen
pa icipa ion ia social media could imp o e use o he eal- ime emission in o ma ion and
s eng hen en i onmen al go e nance. Hence, angible go e nance ou comes a e achie ed only
when en i onmen al issues ga ne subs an ial a en ion om bo h go e nmen s and he public.
This indica es ha , al hough echnology supplies essen ial ools and pla o ms o go e nance,
meaning ul change a ises h ough collabo a i e engagemen be ween echnology and key social
o ces, including go e nmen , indus y, and he public.
The s uc u e o his epo is as ollows: Sec ion 2 e iews he ele an backg ound;
Sec ion 3 de ails he me hodologies employed in his s udy; Sec ion 4 p esen s he indings and
discussion; and Sec ion 5 concludes wi h a summa y o key insigh s and policy ecommenda ions.
4
2. INSTITUTIONAL BACKGROUND AND DEVELOPMENT OF ENVIRONMENTAL
MONITORING TECHNOLOGY
2.1 Ai Pollu ion and Digi al T ans o ma ion
Ai pollu ion, a c i ical global en i onmen al challenge, pa icula ly a ec s de eloping coun ies,
dis up ing he Ea h’s empe a u e balance, diminishing biodi e si y, and exace ba ing espi a o y
diseases (Chak abo y and Lee 2019；Syed, Folz, and Ali 2023)，and po en ially educing labo
p oduc i i y (He, Liu, and Sal o 2019) .In ecen yea s, he PRC has aken ac i e measu es o
educe ai pollu ion and has made no able p og ess (Fig.1), om he pe spec i e o
PM2.5concen a ions, he e has been a sus ained annual educ ion in o e all pollu ion le els
ac oss he CPPRR, indica ing a end o imp o emen in ai quali y o e ime. The go e nmen
ini ia ed his e o in 2013 wi h he “Ai Pollu ion P e en ion and Con ol Ac ion Plan,” which
launched a na ionwide campaign o ackle ai pollu ion and emphasized he use o digi al
moni o ing me hods o s eng hen ai quali y supe ision.
In he e a o sma ci ies, digi al echnology o e s new oppo uni ies o en i onmen al
go e nance. In ields such as a ic managemen and public sa e y, digi al ans o ma ion has
shown posi i e esul s (Wes e man e al. 2014；Hollis 2019, A dabili e al. 2023). Technologies
such as big da a, cloud compu ing, in e ne o hings (IoT), and a i icial in elligence (AI), wi h hei
eal- ime da a collec ion, powe ul da a p ocessing, and p edic i e capabili ies, can analyze da a
pa e ns in conjunc ion wi h a iables such as wea he condi ions, a ic densi y, and indus ial
emissions, enabling accu a e pollu ion o ecas ing and p eemp i e policy measu es, be e
espond o apidly changing en i onmen al condi ions and policy needs (Shen 2018, Agges am
and Mangalagiu 2020). Cu en ly, en i onmen al moni o ing echnology has been widely applied
in soil de ec ion (Wilson 2012), wa e quali y es ing (Chen e al. 2018), and ai quali y moni o ing
(Ma e al. 2014, Van e al. 2015). Empi ical e idence om Wes e n coun ies shows ha
en i onmen al moni o ing and en o cemen ac i i ies p o ide a signi ican de e en e ec ,
educing bo h iola ions and emissions (G ay and Shimshack 2011).
11
Va iable names
N
Mean
S d.De .
Min
Max
In o ma ion Disclosu e
(PITI)
589
3.89
0.30
2.74
4.41
Low-ca bon ci y pilo
(dummy)
1,398
0.42
0.49
0.00
1.00
Ca bon emission ading
pilo (dummy)
1,398
0.15
0.36
0.00
1.00
Join G een Pa en s
1,130
2.67
3.80
1
15
Sou ce: Au ho s’ calcula ions.
4. RESULTS
4.1 Baseline Resul s
Table 2 p esen s he baseline eg ession esul s. In Column (1), we do no con ol o ixed e ec s
o o he in luencing ac o s, and he eg ession coe icien indica es ha an addi ional
en i onmen al moni o ing p ocu emen con ac pe 100,000 people co esponds o a 27%
educ ion in PM2.5 (log) concen a ion. Column (2) displays he eg ession esul s wi h ci y and
yea ixed e ec s included. In Column (3), we u he inco po a e ci y-le el economic and social
cha ac e is ics, while Column (4) shows he inal eg ession esul s wi h addi ional wea he ac o s
included. We ind ha he p ocu emen o digi al en i onmen al moni o ing echnology consis en ly
and signi ican ly educes u ban PM2.5 concen a ions. The eg ession esul s in Column (4)
e eal ha , holding o he ac o s cons an , an addi ional en i onmen al moni o ing con ac pe
100,000 people is associa ed wi h an 8% educ ion in PM2.5 (log) concen a ion, highligh ing he
impo ance o including con ol a iables.
This esul e lec s he e ec i eness o digi al moni o ing echnology in ai pollu ion
managemen , imp o ing ai quali y h ough “moni o ing” and “ o ecas ing.” Fi s , digi al moni o ing
echnology can signi ican ly enhance he accu acy o pollu an sou ce iden i ica ion wi hin he
de ec ion scope. These da a enable go e nmen s and egula o y agencies o quickly de ec
pollu ion sou ces and ake co esponding emedial ac ions. Second, digi al moni o ing, combined
wi h machine lea ning, can analyze and o ecas pollu ion, which g ea ly educes he occu ence
o high emissions and se e e pollu ion e en s.

12
Table 2. Baseline Es ima ion
ln (PM2.5)
ln (PM2.5)
ln (PM2.5)
ln (PM2.5)
(1)
(2)
(3)
(4)
En i onmen
Moni o ing
-0.273***
-0.089***
-0.082***
-0.075**
(0.039)
(0.023)
(0.023)
(0.022)
Economic and social
cha ac e is ic a iables
NO
NO
YES
YES
Wea he cha ac e is ic
a iables
NO
NO
NO
YES
Time FE
NO
YES
YES
YES
Ci y FE
NO
YES
YES
YES
N
922
922
882
882
adj. R-sq
0.0648
0.587
0.668
0.673
No e: This able p esen s he baseline eg ession esul s o he model. En i onmen Moni o ing is ou dependen
a iable, ep esen ing he numbe o con ac s pe 100,000 people, while ln(PM2.5) ep esen s he loga i hms o he
annual a e age PM2.5 in he ci y. The eg ession coe icien s a e es ima ed using he OLS model. * p < 0.1, ** p < 0.05,
*** p < 0.01; S anda d e o s in pa en heses a e clus e ed a he ci y le el.
Sou ce: Au ho s’ calcula ions.
4.2 Robus ness Analyses
We conduc ed a se ies o eg essions o examine he obus ness o ou main esul s. Fi s ly, we
excluded he e ec s o municipali ies (Column 1), as hei highe adminis a i e le els di e in
esou ce alloca ion and en i onmen al go e nance compa ed o egula ci ies, making he esul s
mo e gene alizable. Secondly, pilo en i onmen al policies, such as low-ca bon ci ies o ca bon
emission ading pilo s, could cause biases in pollu ion le els and he e ec i eness o moni o ing
echnology, so we also con olled o his in luence (Column 2). The PRC is also a majo emi e
o sul u dioxide (SO₂), hus we eplaced he dependen a iable wi h ci y-le el SO₂ concen a ion
(Column 3) o examine he impac o digi al moni o ing echnology on o he ai pollu an s. We
ound ha o e e y addi ional digi al moni o ing con ac pe 100,000 people, he SO₂ le el
dec eases by 15%. This inding sugges s ha digi al moni o ing echnology exe s a g ea e
in luence on SO₂ han on PM2.5. This is likely because en i onmen al moni o ing echnologies
p ima ily a ge la ge-scale, high-emission sou ces, such as coal powe plan s and s eel mills,
which a e majo con ibu o s o SO₂ emissions.
13
Addi ionally, since digi al moni o ing equipmen has a ce ain du abili y pe iod, p e iously
ins alled equipmen may con inue o exe in luence. We hus measu ed u ban digi al
en i onmen al moni o ing in ensi y using he cumula i e numbe o en i onmen al moni o ing
con ac s (Column 4). Since he indus ial sec o is a signi ican sou ce o PM2.5 emissions, we
also calcula ed he a io o con ac numbe s o he sha e o indus ial ou pu in GDP, p o iding a
measu e o digi al moni o ing in ensi y wi hin he indus ial sec o (Column 5). Fu he , as
anspo a ion is also a signi ican sou ce o pollu ion, we included a ic moni o ing in he analysis
o ai pollu ion (Column 6). Ou coe icien o in e es emains signi ican and sligh ly la ge ,
sugges ing ha , beyond i s ole in iden i ying a ic iola ions and managing conges ion, a ic
moni o ing has a spillo e e ec in pollu ion con ol (Column 7). We also eplaced he explana o y
a iable wi h u ban digi al moni o ing p ocu emen expendi u e o e-measu e he in ensi y o
digi al moni o ing echnology implemen a ion. While he coe icien o in e es is sligh ly smalle in
magni ude and signi icance han when measu ed by con ac numbe s, he signi icance o all
obus ness es esul s consis en ly indica es he s able supp essi e e ec o digi al en i onmen al
moni o ing echnology on ai pollu ion.
14
Table 3. Robus ness Check
No e: This able p esen s he se en obus ness es s men ioned. We used OLS es ima ion o conduc a se ies o eg essions o examine he obus ness o ou main
esul s. *p < 0.1, ** p < 0.05, *** p < 0.01; S anda d e o s in pa en heses a e clus e ed a he ci y le el.
Sou ce: Au ho s’ calcula ions.
Excluding
municipali ies
Inco po a e
Policy Pilo
ln(SO2)
Cumula i e
Con ac s
En Moni _SecInd
En _ ans_con
ac _100k
Digi al Moni o ing
alues
(1)
(2)
(3)
(4)
(5)
(6)
(7)
En i onmen
Moni o ing
-0.075***
-0.073***
-0.158***
(0.022)
(0.021)
(0.041)
ln(En _cumula i e
Con ac s)
-0.043**
(0.018)
Sec_con ac s
-0.021**
(0.010)
En _ ans_con ac _
100k
-0.081**
(0.029)
Digi al Moni o ing
alues
-0.007*
(0.004)
Low-ca bon ci y pilo
Ca bon emission
ading pilo
NO
YES
NO
NO
NO
NO
NO
Con ol a iables
YES
YES
YES
YES
YES
YES
YES
Time FE
YES
YES
YES
YES
YES
YES
YES
Ci y FE
YES
YES
YES
YES
YES
YES
YES
N
873
882
882
1517
882
1089
882
adj. R-sq
0.673
0.674
0.780
0.408
0.676
0.592
0.530
15
4.3 Endogenei y Analysis
We employ a wo-s age eg ession wi h ins umen al a iables o add ess po en ial endogenei y
issues in his s udy. D awing on he app oach by A kin (2013), we cons uc a supply shock:
,2014
,2014
,2014
i
i
ci y
i
ci y
iI
Con ac
Con ac
ϕ
∈
=∑
(2)
,
, ,2014
ln( )
i
i i ci y
iI
IV Con ac
ϕ
∈
= ×∑
(3)
He e, ep esen s he p opo ion o digi al en i onmen al moni o ing con ac s in a gi en ci y
o he o al digi al en i onmen al moni o ing con ac s in he p o ince a he ini ial sample pe iod,
which measu es he egion’s esponsi eness o ma ke oppo uni ies in digi al moni o ing
de elopmen . The e o e, he in e ac ion be ween and he o al numbe o con ac s wi hin
he p o ince a each ime pe iod e lec s ma ke -d i en shocks in digi al moni o ing de elopmen .
As shown in Table 4, he F-s a is ics a e g ea e han 10. The esul indica es ha , a e con olling
o ma ke shocks, he e ec o digi al moni o ing echnology on PM2.5 supp ession is mo e
p onounced (wi h a coe icien o -0.28). This pa ly sugges s ha ma ke esponsi eness and
ex e nal shocks play a mo e signi ican ole in he p omo ion o echnology han simple policy-
d i en e o s.
2014,i
ϕ
2014,i
ϕ
16
Table 4. Resul s o Endogenei y Tes
IV es ima ion
(1)
Second-s age eg ession
En i onmen Moni o ing
-0.28 *
(0 .160)
Economic and social cha ac e is ic a iables
YES
Wea he cha ac e is ic a iables
YES
Time FE
YES
Ci y FE
YES
Fi s -s age eg ession
Indus y shock_IV
0.143 ***
(0.013)
Con ol a iables
YES
Time FE
YES
Ci y FE
YES
F-s a .
114.35
N
877
No e: This able p esen s he eg ession esul s es ima ed using p o incial public secu i y moni o ing con ac s as he
IV a iable. The model includes a o al o 1,021 samples, and he eg ession coe icien s we e ob ained using 2SLS
es ima ion. In all models, con ol a iables, yea ixed e ec s, and ci y ixed e ec s a e ully con olled. *p < 0.1, ** p <
0.05, *** p < 0.01; S anda d e o s in pa en heses a e clus e ed a he ci y le el.
Sou ce: Au ho s’ calcula ions.
4.4 Mechanism Analysis
Digi al echnology se es p ima ily as a ool o da a analysis and decision suppo , a he han
di ec ly educing pollu an concen a ions. I s impac is mos e iden in i s indi ec ole, whe e i
enhances go e nance e iciency by acili a ing in o ma ion low and s eng hening decision-
making. As shown in Column (1), en i onmen al moni o ing echnology signi ican ly imp o es he
iden i ica ion o pollu ion cases; he coe icien indica es ha each addi ional en i onmen al
moni o ing p ocu emen con ac pe 100,000 people is associa ed wi h a 46% inc ease in
en i onmen adminis a i e penal y cases iden i ied by he go e nmen . This inding indica es ha
p ocu ing digi al moni o ing echnology signi ican ly enhances anspa ency in bo h in o ma ion
disclosu e and egula o y en o cemen . By educing in o ma ion asymme y in he p incipal-agen
ela ionship be ween pollu ing en e p ises and he go e nmen , digi al moni o ing s eng hens
accoun abili y (Liu, Uchida, and Bao 2024).
The inc eased isk o egula o y sc u iny di ec ly compels en e p ises o adop mo e
ad anced en i onmen ally iendly echnologies in hei p oduc ion and managemen p ocesses.
To u he in es iga e his e ec , we examined he impac o digi al en i onmen al echnology on

17
di e en ypes o en e p ises, inding ha hea ily pollu ing i ms a e mos a ec ed. As indica ed
in Column (2), each addi ional en i onmen al moni o ing p ocu emen con ac pe 100,000
people is associa ed wi h a 73.3% educ ion in he numbe o hea ily pollu ing en e p ises in a
ci y. Tian ian e al. (2024) obse ed simila ends, no ing ha inc eased egula o y in ensi y has
d i en he sys ema ic elimina ion o ou da ed p oduc ion capaci ies in he indus ial sec o ,
including go e nmen -led closu es o small, hea ily pollu ing ac o ies. This ou come is la gely
because o he subs an ial compliance cos s associa ed wi h adop ing ad anced clean
echnologies, which pose pa icula challenges o pollu ion-in ensi e indus ies (Zhang and Li
2023).
On he o he hand, he Po e Hypo hesis posi s ha s ic e en i onmen al egula ions
ul ima ely enhance co po a e g een pe o mance and compe i i eness (Zhang e al. 2024, Cohen
and Tubb 2017). To explo e his, we examined he impac o adop ing digi al en i onmen al
moni o ing echnology on he g een pa en s co-in en ed by en e p ises in he ci y. The eg ession
esul s in Column (3) indica e ha digi al moni o ing echnology accele a es was e u iliza ion and
educes ene gy consump ion, leading o “inno a ion compensa ion.”
Table 5. Resul s o Mechanism Analysis
Adminis a i e
penal y(log)
Pollu ing
Fi m(log)
Join G een
Pa en s(log)
(1) (2) (3)
En i onmen Moni o ing
0.463**
-0.733*
0.378**
(0.226)
(0.374)
(0.152)
Economic and social cha ac e is ic
a iables
YES
YES
YES
Wea he cha ac e is ic a iables
YES
YES
YES
Time FE
YES
YES
YES
Ci y FE
YES
YES
YES
N
715
773
704
adj. R-sq
0.041
0.361
0.110
No e: The able eco ds h ee possible mechanisms h ough which digi al en i onmen al moni o ing echnology a ec s
u ban PM2.5 le els. Column (1) ep esen s he impac o en i onmen al moni o ing echnology on he numbe o u ban
en i onmen al p o ec ion cases; Column (2) shows he impac o en i onmen al moni o ing echnology on he numbe
o hea ily pollu ing en e p ises in he ci y; Column (3) ep esen s he impac o en i onmen al moni o ing echnology on
he numbe o g een pa en s in he ci y. In all models, con ol a iables, yea ixed e ec s, and ci y ixed e ec s a e ully
con olled. *p < 0.1, ** p < 0.05, *** p < 0.01; S anda d e o s in pa en heses a e clus e ed a he ci y le el.
Sou ce: Au ho s’ calcula ions.
18
4.5 He e ogenei y Analysis
4.5.1 Public In e es and Go e nmen A en ion
As p e iously ou lined, digi al en i onmen al moni o ing echnology unc ions p ima ily as an
“in o ma ion p o ide ,” while he go e nmen , en e p ises, and indi iduals ac as he ac ual “use s”
o his da a. Inc eased a en ion o ai quali y da a o en e lec s highe expec a ions o he
en i onmen , and he supe iso y p essu e esul ing om his heigh ened awa eness is mo e
likely o ansla e in o ac ual pollu ion con ol decisions, d i ing en i onmen al imp o emen .
Ou analysis ocuses on he e ec s o public and go e nmen al a en ion on en i onmen al
pollu ion. Public in e es is measu ed using he Baidu Sea ch Index, whe eas go e nmen
a en ion is assessed h ough he equency o en i onmen ally ela ed e ms in go e nmen wo k
epo s—a widely adop ed app oach in en i onmen al egula ion s udies (Tu e al. 2024, Chen
and Chen 2018). Based on median alues, bo h a en ion me ics a e ca ego ized in o high and
low g oups. The eg ession esul s a e p esen ed in Table 6.
In egions wi h highe public en i onmen al conce n, an addi ional p ocu emen con ac
pe 100,000 people leads o a 10% educ ion in PM2.5 le els (Column 1). This esul unde sco es
he impo ance o public in luence in en i onmen al go e nance. As public conce n o he
en i onmen inc eases, go e nmen s a e likely o ace g ea e egula o y p essu e, leading o
s ic e pollu ion con ol measu es (Wang and Jia 2021, Wang and Cao 2024).
Simila ly, in egions wi h highe le els o go e nmen en i onmen al conce n, each
addi ional en i onmen al moni o ing p ocu emen con ac pe 100,000 esiden s co esponds o
a 9% educ ion in PM2.5 le els (Column 3). Compa ing hese coe icien s shows ha public
a en ion exe s a s onge en i onmen al impac han go e nmen conce n. This can be a ibu ed
o sus ained public a en ion, which ypically os e s s onge social o e sigh and en i onmen al
eedback, he eby enhancing pollu ion con ol e ec i eness. Howe e , in a eas wi h lowe le els
o public and go e nmen en i onmen al conce n, we do no obse e a signi ican impac o digi al
en i onmen al moni o ing on pollu ion con ol. This esul highligh s he syne gy be ween
19
echnology and human beha io (i.e., social a en ion and go e nmen decision-making),
highligh ing he condi ions and dependencies o echnology o achie e e ec i e ou comes.
Table 6. Impac on Public In e es and Go e nmen A en ion
High_public
in e es
Low_public
in e es
High_go e nmen
a en ion
Low_go e nmen
a en ion
(1)
(2)
(3)
(4)
En i onmen Moni o ing
-0.103**
-0.038
-0.088**
-0.051
(0.033)
(0.031)
(0.027)
(0.030)
Economic and social
cha ac e is ic a iables
YES
YES
YES
YES
Wea he cha ac e is ic
a iables
YES
YES
YES
YES
Time FE
YES
YES
YES
YES
Ci y FE
YES
YES
YES
YES
P- alue o in e -g oup
di e ence coe icien
0.050**
0.049**
N
606
276
483
399
adj. R-sq
0.742
0.655
0.770
0.661
No e: This able shows he he e ogenei y o he impac o digi al en i onmen al moni o ing echnology on PM2.5 unde
public and go e nmen en i onmen al conce n, wi h he median used as he g ouping s anda d. Columns (1)–(2)
ep esen he he e ogenei y o public en i onmen al in e es on u ban PM2.5, while Columns (3)–(4) ep esen he
he e ogenei y o go e nmen en i onmen al conce n. We used a wo-sided - es o examine he signi icance o
di e ences be ween coe icien s ac oss di e en g oups. In all models, con ol a iables, yea ixed e ec s, and ci y
ixed e ec s a e ully con olled. * p < 0.1, ** p < 0.05, *** p < 0.01; S anda d e o s in pa en heses a e clus e ed a he
ci y le el.
Sou ce: Au ho s’ calcula ions.
4.5.2 En i onmen al In o ma ion Disclosu e
Public in e es in en i onmen al in o ma ion is p edica ed on he accessibili y and anspa ency o
en i onmen al da a. We use he Pollu ion In o ma ion T anspa ency Index (PITI) o measu e he
le el o en i onmen al in o ma ion disclosu e in PRC ci ies, wi h co e e alua ion con en including
he disclosu e o en e p ise emissions in o ma ion, en i onmen al egula o y anspa ency, ai
quali y in o ma ion, and public pa icipa ion. Because o da a a ailabili y, we ob ained he
en i onmen al in o ma ion disclosu e index o 120 ci ies om 2014 o 2018.
The eg ession esul s a e shown in Table 7. In egions wi h highe le els o en i onmen al
in o ma ion disclosu e, each addi ional en i onmen al moni o ing p ocu emen con ac pe
20
100,000 esiden s co esponds o a 7% educ ion in PM2.5 le els (Column (1)). This inding
unde sco es he c i ical ole o in o ma ion disclosu e in e ec i e en i onmen al go e nance. A
lack o anspa en en i onmen al da a can encou age lax p oduc ion p ac ices among en e p ises
and os e en -seeking beha io among go e nmen o icials. Enhanced anspa ency, howe e ,
di ec ly es ic s oppo uni ies o o icials o bene i om concealing in o ma ion, he eby
ein o cing public o e sigh and bols e ing go e nmen accoun abili y (Wei and He 2022, Chen e
al. 2022). We u he p esen ou baseline eg ession coe icien s, he e ogenei y by a en ion le el,
and in o ma ion disclosu e h ough coe icien isualiza ion (Fig. 4).
Table 7. Impac on In o ma ion Disclosu e
ln(PM2.5)
(1)
(2)
High in o ma ion disclosu e
Low in o ma ion disclosu e
En i onmen Moni o ing
-0.069**
-0.011
(0.028)
(0.036)
Economic and social
cha ac e is ic a iables
YES YES
Wea he cha ac e is ic a iables
YES
YES
Time FE
YES
YES
Ci y FE
YES
YES
P- alue o in e -g oup di e ence
coe icien
0.093*
N
403
340
adj. R-sq
0.677
0.761
No e: This able shows he impac o digi al en i onmen al moni o ing on PM2.5 unde he he e ogenei y o in o ma ion
disclosu e. We used a wo-sided - es o examine he signi icance o di e ences be ween coe icien s ac oss di e en
g oups. In all models, con ol a iables, yea ixed e ec s, and ci y ixed e ec s a e ully con olled. * p < 0.1, ** p < 0.05,
*** p < 0.01; S anda d e o s in pa en heses a e clus e ed a he ci y le el.
Sou ce: Au ho s’ calcula ions.
27
Jin, Lei, Bin Wang, Guangming Shi, Ba nabas C. Seyle , Xue Qiao, Xun ei Deng, Xia Jiang,
Fumo Yang, and Yu Zhan. 2020. Impac o China’s ecen amendmen s o ai quali y
moni o ing p o ocol on epo ed ends. A mosphe e 11(11):1199.
h ps://doi.o g/10.3390/a mos11111199
Liu, Jianlei, Kona i Uchida, and Chenyu Bao. 2024. En i onmen al egula ion, co po a e
en i onmen al disclosu e, and i m pe o mance: E idence om China. Paci ic-Basin
Finance Jou nal 85, 102367. h ps://doi.o g/10.1016/j.pac in.2024.102367
Luo, Weijie, Di Xiang, and Xiaoming Zhang. 2024. Does he Po e hypo hesis hold in China?
e idence om he “2+26” egional ai pollu ion ea men policy. En i onmen and
De elopmen Economics 29(4): 319–337. h ps://doi.o g/10.1017/S1355770X24000093
Ma, Zongwei, Xue ei Hu, Lei Huang, Jun Bi, and Yang Liu. 2014. Es ima ing g ound-le el PM2.5
in China using sa elli e emo e sensing. En i onmen al Science & Technology 48(13),
7436–7444. h ps://pubs.acs.o g/doi/abs/10.1021/es5009399
Mao, Jie, and Chunhua Wang. 2016. Tax incen i es and en i onmen al p o ec ion: E idence
om China’s axpaye -le el da a. China Finance and Economic Re iew 4(1), 14.
h ps://link.sp inge .com/a icle/10.1186/s40589-016-0040-0
Mao, Huizhen, Wen Wang, Liyong Lu, Wenzhong Huang, Yu Min, Chaohui Wang, Weibin Liao,
Jay Pan, Wei Lin, and Ke Ju. 2024. Die a y pa e ns and he e ec o long- e m PM2.5
exposu e on me abolic synd ome among Chinese adul s: A c oss-sec ional s udy.
En i onmen al Resea ch Le e s 19(11), 114002.
h ps://iopscience.iop.o g/a icle/10.1088/1748-9326/ad1d9c/pd
Pe y, Geo ge LW, Rupe Seidl, And é M. Bell é, and We ne Ramme . 2022. An ou look o
deep lea ning in ecosys em science. Ecosys ems 25(8), 1700–1718.
h ps://link.sp inge .com/con en /pd /10.1007/s10021-022-00789-y.pd
Schlæge , Jespe , and Jingjing Zhou. 2019. Digi al En i onmen al Moni o ing in U ban China. In:
Delman, J., Y. Ren, O. Luo a, M. Bu ell, and O. Almén, eds. G eening China’s U ban
Go e nance. ARI - Sp inge Asia Se ies Vol 7. Singapo e: Sp inge .
h ps://link.sp inge .com/chap e /10.1007/978-981-13-0740-9_7

28
Shahbaz, Muhammad, Hooi Hooi Lean, and Muhammad Shahbaz Shabbi . 2012. En i onmen al
Kuzne s cu e hypo hesis in Pakis an: Coin eg a ion and G ange causali y. Renewable and
Sus ainable Ene gy Re iews 16(5), 2947–2953. h ps://doi.o g/10.1016/j. se .2012.02.015
Shen, Yi. 2018. Da a sha ing p ac ices, in o ma ion exchange beha io s, and knowledge
disco e y dynamics: a s udy o na u al esou ces and en i onmen al scien is s.
En i onmen al Sys ems Resea ch. 6 (1), 9–14.
h ps://link.sp inge .com/a icle/10.1186/s40068-017-0086-5
Smi h, J. and A. Doe. 2020. Ai Quali y Moni o ing Using IoT and Big Da a. In e na ional Jou nal
o En i onmen al S udies 12(3), 45–60.
Syed, Mohi, Rodney J. Folz, and Uzma Ali. 2023. En i onmen al ac o s and hei impac on
ai way diseases: Explo ing ai pollu ion, indoo and ou doo alle gens, and clima e change.
Cu en Pulmonology Repo s 12(3), 162–170.
h ps://link.sp inge .com/a icle/10.1007/s13665-023-00319-8
Wu, Wencong. 2013, July 25. S a e Council gi es app o al o ough ai emissions plan. China
Daily. h p://usa.chinadaily.com.cn/epape /2013-07/25/con en _16829757.h m
Tang, Xiao, Yinglun Wang, and Hong ao Yi. 2023. Da a manipula ion h ough pa onage
ne wo ks: E idence om en i onmen al emissions in China. Jou nal o Public
Adminis a ion Resea ch and Theo y 33(2), 342–356.
h ps://doi.o g/10.1093/jopa /muac019
Tu, Chaoyang, Yuxi Liang, and Yong Fu. 2024. How does he en i onmen al a en ion o local
go e nmen s a ec egional g een de elopmen ? Empi ical e idence om local
go e nmen s in China. Humani ies and Social Sciences Communica ions 11(1), 1–14.
h ps://www.na u e.com/a icles/s41599-024-02887-9
Van den Bossche, Jo is, Jan Pe e s, Jan Ve wae en, Dick Bo eldoo en, Jan Theunis, and
Be na d De Bae s. 2015. Mobile moni o ing o mapping spa ial a ia ion in u ban ai
quali y: De elopmen and alida ion o a me hodology based on an ex ensi e da ase .
A mosphe ic En i onmen 105, 148–161. h ps://doi.o g/10.1016/j.a mosen .2015.01.017
Wang, Chen, and Juanjuan Cao. 2024. Ai pollu ion, heal h s a us and public awa eness o
en i onmen al p oblems in China. Scien i ic Repo s 14(1), 19861.
h ps://www.na u e.com/a icles/s41598-024-69992-2
29
Wang, Junsong, and Yanhui Jia. 2021. Social media’s in luence on ai quali y imp o emen :
E idence om China. Jou nal o Cleane P oduc ion 298, 126769.
h ps://www.sciencedi ec .com/science/a icle/pii/S0959652621009884
Wang, Y., L. Li, C. Chen, and D. Huang. 2022. Spa io empo al a ia ion in PM2.5 concen a ions
and hei ela ionship wi h socioeconomic ac o s in China’s majo ci ies. En i onmen al
Pollu ion 300, 118985.
Wei, Ye, and Wenjian He. 2022. Can an i-co up ion imp o e he quali y o en i onmen al
in o ma ion disclosu e? En i onmen al Science and Pollu ion Resea ch 29, 5345–5359.
h ps://link.sp inge .com/a icle/10.1007/s11356-021-15932-w
Wes e man, Geo ge, Didie Bonne , and And ew McA ee. 2014. The nine elemen s o digi al
ans o ma ion. MIT Sloan Managemen Re iew 55(3), 1–6.
h ps://www.p oques .com/open iew/427bcb94c8d30228 d1e5aa2 945bd0e/1?pq-
o igsi e=gschola &cbl=26142
Wo ld Bank. 2022, Oc obe 12. China’s T ansi ion o a Low-Ca bon Economy and Clima e
Resilience Needs Shi s in Resou ces and Technologies. h ps://www.wo ldbank.o g
Wilson, Alphus Dan. 2012. Re iew o elec onic-nose echnologies and algo i hms o de ec
haza dous chemicals in he en i onmen . P ocedia Technology 1, 453–463.
h ps://doi.o g/10.1016/j.p o cy.2012.02.101
Xu, Shengyu, Jinqiu Yang, and Ruile Li. 2024. How does g een c edi guidelines a ec
en i onmen ally iendly en e p ises’ ESG? A quasi-na u al expe imen om China. Public
Lib a y o Science One, 19(5),e0304384. h ps://doi.o g/10.1371/jou nal.pone.0304384
Yang, Lin, Ya ang Lin, Jin Wang, and Fangyuan Peng. 2024. Achie ing ai pollu ion con ol
a ge s wi h echnology-aided moni o ing: Be e en o cemen o localized e o s?
Ame ican Economic Jou nal: Economic Policy 16(4), 280–315.
h ps://www.aeaweb.o g/a icles?id=10.1257/pol.20220810
Zhang, Qi, Zhi Yu, and Dongmin Kong. 2019. The eal e ec o legal ins i u ions: en i onmen al
cou s and i m en i onmen al p o ec ion expendi u e. Jou nal o En i onmen al
Economics and Managemen 98(3), 102254. h ps://doi.o g/10.1016/j.jeem.2019.102254
Zhang, Wanli, Bin Zhu, Yongling Li, and Dan Yan. 2024. Re isi ing he Po e hypo hesis: A
mul i-coun y me a-analysis o he ela ionship be ween en i onmen al egula ion and
g een inno a ion. Palg a e Communica ions, 11(1), 1–15.
h ps://www.na u e.com/a icles/s41599-024-02671-9
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. 789
July 2025
Digi al Moni o ing Technology and Ai Quali y
E idence om he People’s Republic o China
This pape empi ically in es iga es he impac o digi al en i onmen al moni o ing echnologies on ai
quali y in he People’s Republic o China using public p ocu emen da a. The indings show ha ollowing
he adop ion o digi al en i onmen al moni o ing echnology: (i) ci y-le el PM2.5 concen a ions exhibi
a sizeable educ ion, (ii) egula o y en o cemen imp o es and hea ily pollu ing i ms exi mo e, (iii) g een
inno a ion is enhanced h ough mo e accu a e pollu an iden i ica ion, and (i ) he e ec i eness o hese
echnologies depends on he ex en o in o ma ion disclosu e and public pa icipa ion.
Abou he Asian De elopmen Bank
ADB is a leading mul ila e al de elopmen bank suppo ing inclusi e, esilien , and sus ainable g ow h ac oss
Asia and he Paci ic. Wo king wi h i s membe s and pa ne s o sol e complex challenges oge he , ADB
ha nesses inno a i e inancial ools and s a egic pa ne ships o ans o m li es, build quali y in as uc u e,
and sa egua d ou plane . Founded in 1966, ADB is owned by 69 membe s—50 om he egion. DIGITAL MONITORING
TECHNOLOGY AND AIR QUALITY
EVIDENCE FROM THE PEOPLE’S REPUBLIC OF CHINA
Pinghan Liang, Yadi Liu, and Shu Tian