Ca allo, Edua do A.; Cepeda, Ana; Panizza, Ugo
Wo king Pape
En i onmen al damage news and s ock e u ns: E idence
om La in Ame ica
IDB Wo king Pape Se ies, No. IDB-WP-1593
P o ided in Coope a ion wi h:
In e -Ame ican De elopmen Bank (IDB), Washing on, DC
Sugges ed Ci a ion: Ca allo, Edua do A.; Cepeda, Ana; Panizza, Ugo (2024) : En i onmen al damage
news and s ock e u ns: E idence om La in Ame ica, IDB Wo king Pape Se ies, No. IDB-WP-1593,
In e -Ame ican De elopmen Bank (IDB), Washing on, DC,
h ps://doi.o g/10.18235/0012962
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/299400
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/
En i onmen al Damage News and S ock
Re u ns:
E idence om La in Ame ica
Edua do Ca allo
Ana Cepeda
Ugo Panizza
WORKING PAPER No IDB-WP-1593
In e -
A
me ican De elopmen Bank
Depa men o Resea ch and Chie Economis
May 2024
* In e -
A
me ican De elopmen Bank
** In e na ional Mone a y Fund
*** Gene a G adua e Ins i u e and CEPR
En i onmen al Damage News and S ock
Re u ns:
E idence om La in Ame ica
Edua do Ca allo*
Ana Cepeda**
Ugo Panizza***
In e -
A
me ican De elopmen Bank
Depa men o Resea ch and Chie Economis
May 2024
Ca aloging-in-Publica ion da a p o ided by he
In e -Ame ican De elopmen Bank
Felipe He e a Lib a y
Ca allo, Edua do A.
En i onmen al damage news and s ock e u ns: e idence om La in Ame ica
/ Edua do Ca allo, Ana Cepeda, Ugo Panizza.
p. cm. — (IDB Wo king Pape Se ies ; 1593)
Includes bibliog aphical e e ences.
1. Ca bon axes-La in Ame ica. 2. Clima ic changes-Economic aspec s-La in
Ame ica. 3. En i onmen al impac cha ges-La in Ame ica. 4. Ca bon dioxide
mi iga ion-Economic aspec s-La in Ame ica. I. Cepeda, Ana. II. Panizza, Ugo.
III. In e -Ame ican De elopmen Bank. Depa men o Resea ch and Chie
Economis . IV. Ti le. V. Se ies.
IDB-WP-1593
h p://www.iadb.o g
Copy igh © 2024 In e -Ame ican De elopmen Bank ("IDB"). This wo k is subjec o a C ea i e
Commons license CC BY 3.0 IGO (h ps://c ea i ecommons.o g/licenses/by/3.0/igo/legalcode). The
e ms and condi ions indica ed in he URL link mus be me and he espec i e ecogni ion mus be
g an ed o he IDB.
Fu he o sec ion 8 o he abo e license, any media ion ela ing o dispu es a ising unde such license
shall be conduc ed in acco dance wi h he WIPO Media ion Rules. Any dispu e ela ed o he use o
he wo ks o he IDB ha canno be se led amicably shall be submi ed o a bi a ion pu suan o he
Uni ed Na ions Commission on In e na ional T ade Law (UNCITRAL) ules. The use o he IDB's name
o any pu pose o he han o a ibu ion, and he use o IDB's logo shall be subjec o a sepa a e
w i en license ag eemen be ween he IDB and he use and is no au ho ized as pa o his license.
No e ha he URL link includes e ms and condi ions ha a e an in eg al pa o his license.
The opinions exp essed in his wo k a e hose o he au ho s and do no necessa ily e lec he iews o
he In e -Ame ican De elopmen Bank, i s Boa d o Di ec o s, o he coun ies hey ep esen .
En i onmen al Damage News and S ock Re u ns:
E idence om La in Ame ica
Edua do Ca allo
In e -Ame ican De elopmen Bank
Ana Cepeda
In e na ional Mone a y Fund
Ugo Panizza
Gene a G adua e Ins i u e & CEPR
May 2024∗
Abs ac
This pape s udies he in e play be ween en i onmen al pe o mance and inancial alua ion
o i ms in La in Ame ica and he Ca ibbean. We p o ide insigh s in o how en i onmen al
conside a ions a e in eg a ed in o inancial decision-making and in es o beha io by analyz-
ing he s ock ma ke eac ion o en i onmen al news o i ms wi h di e en le els o ca bon
emission in ensi y. We ind ha high emission in ensi y i ms end o unde pe o m a e
he elease o en i onmen al damage news. Ou baseline es ima es indica e ha , a e he
elease o such news, i ms a he 75 h pe cen ile o he dis ibu ion o emission in ensi y
expe ience s ock e u ns ha a e 17% lowe han hose o i ms a he 25 h pe cen ile o he
dis ibu ion o emission in ensi y. These esul s sugges ha in es o s ca e abou and p ice
ca bon isk, bu only when his isk is salien .
JEL classi ica ions: G12, G14, G18, G32, G38, Q54
Keywo ds: Ca bon emissions, Clima e change, En i onmen al news, S ock e u ns
∗Con ac in o ma ion: ca allo[email p o ec ed]; acepeda [email p o ec ed]; [email p o ec ed]h. We would
like o hank Ma ha Elena Delgado o ou s anding esea ch assis ance and Ma ina Hengge and Richa d Va ghese
o use ul discussions on opics ela ed o his pape . The iews exp essed in his pape a e hose o he au ho s
and do no necessa ily ep esen he iews o he In e -Ame ican De elopmen and he IMF, hei Execu i e
Boa ds, o Managemen .
1 In oduc ion
We in es iga e he impac o en i onmen al damages news on he s ock e u ns o i ms in
La in Ame ica and he Ca ibbean. We ocus on how his impac a ies wi h i m-le el ca bon
emission in ensi y. We ind ha high emission in ensi y i ms unde pe o m a e he elease
o en i onmen al damage news and show ha his esul is p incipally d i en by domes ic
news. Ou indings indica e ha en i onmen al ac o s play an impo an ole in i ms’ inancial
pe o mance and in es o s’ decision-making.
The e is nea unanimi y on he an h opogenic o igin o clima e change and a be e un-
de s anding o he link be ween ca bon emissions and inancial pe o mance is key o de ising
policies aimed a c ea ing incen i es o educing emissions. Fi ms, pa icula ly in ce ain sec-
o s, a e signi ican con ibu o s o eleasing g eenhouse gas (GHG).1These emissions a e key
d i e s o global wa ming and clima e change, phenomena ha a e al e ing he ea h’s na u al
sys ems a an unp eceden ed a e.
We ocus on en i onmen al damages because GHG emissions con ibu e o he in ensi ica ion
o ex eme wea he e en s, such as hu icanes, loods, and d ough s, which ha e de as a ing
ecological and economic consequences (see, o example, Min e al. (2011)) Mo eo e , ising
global empe a u es esul in mel ing pola ice caps, ising sea le els, and des oying habi a s.
This cascade o e ec s dis up s biodi e si y and poses subs an ial isks o wildli e, ma ine li e,
and na u al ecosys ems. The al e a ion in clima e pa e ns also has p o ound implica ions
o ag icul u e, wa e esou ces, and human heal h, c ea ing a domino e ec o en i onmen al
challenges. E en hough en i onmen al damages a e no necessa ily ela ed o clima e change
( hink o an oil spill), media co e age o such e en s can make he clima e change discou se
salien and impac in es men decisions, ei he h ough a p e e ence shi owa ds g eene asse s
o heigh ened public awa eness o en i onmen al isks, a ec ing i ms’ pe cei ed u u e inancial
pe o mance. In ac , he e is e idence ha in es o s pe cei e i ms wi h highe ca bon emissions
as less en i onmen ally esponsible (Dahlmann e al., 2019), and his pe cep ion can lead o a
mo e nega i e eac ion a e he elease o en i onmen al damage news, e en i he speci ic piece
o news is no di ec ly di ec ly ela ed o clima e change (Lee e al., 2015).
To s udy he ela ionship be ween s ock ma ke pe o mance, en i onmen al damage news,
and ca bon emission in ensi y, we combine i m-le el emission in ensi y da a wi h daily s ock
e u n da a and a speci ically cons uc ed high- equency da ase o en i onmen al damage news.
Ou analysis includes a sample o 840 i ms based in 16 coun ies ac oss La in Ame ica and he
Ca ibbean, co e ing a pe iod o 13 yea s (2009-2022). By using an econome ic speci ica ion ha
builds on Hengge e al. (2023), we ind ha he elease o en i onmen al damage news nega i ely
impac s he s ock ma ke pe o mance o i ms wi h high emission in ensi ies. We obse e ha
i ms a he 75 h pe cen ile o emission in ensi y dis ibu ion expe ience s ock e u ns ha a e
1In he es o he pape , we will use ca bon emissions and g een house gas (GHG) in e changeably.
2
17% lowe compa ed o hose a he 25 h pe cen ile. Ou indings a e p ima ily in luenced by
he ex ensi e ma gin (i.e., he p esence o a leas one piece o news) and a e mos ly associa ed
wi h he elease o domes ic a he han egional o global news. Addi ionally, ou esul s a e
no a ibu ed o a speci ic coun y, no do hey cap u e a spu ious ela ionship d i en by i m
cha ac e is ics such as size, p o i abili y, o sec o o ope a ion, which could be co ela ed wi h
bo h ca bon emission in ensi y and s ock e u ns.
Ou pape is ela ed o wo s ands o li e a u e. The i s s and ocuses on on he link
be ween ca bon emissions and s ock e u ns. Two in luen ial pape s by Bol on and Kacpe czyk
(2021,2022) ind ha he le el o ca bon emissions is associa ed wi h highe s ock e u ns
in a c oss-sec ion o i ms. Acco ding o Bol on and Kacpe czyk (2021,2022), his nega i e
associa ion is d i en by he p esence o a ca bon isk p emium: in es o s equi e highe e u ns
o hold s ocks o companies ha a e exposed o ca bon p icing and egula ion isk. This esul
has been challenged by Zhang (2023), who sugges s ha esea che s should ocus on lagged
emissions, and by Aswani e al. (2023), who sugges ha ca bon emission in ensi y (ins ead o
he le el o emissions) should be he app op ia e measu e o ca bon isk ( o a ebu al, see
Bol on and Kacpe czyk, 2023).
Th ee pape s ha , like us, use high- equency s ock p ices a e Bol on e al. (2022), Millische
e al. (2022), and Hengge e al. (2023). These au ho s s udy he impac o ca bon p ice on s ock
e u ns wi hin he Eu opean Union Emission T ading Scheme (EU ETS). Bol on e al. (2022)
and Millische e al. (2022) show ha i m-le el a ailabili y o ca bon allowances is an impo an
d i e o he link be ween ca bon p ice and s ock e u ns. These esul s indica e ha , wi hin
he EU ETS, ca bon p ice a ec s s ock e u ns h ough an inpu cos channel. Hengge e al.
(2023) use ca bon policy su p ises and show ha ansi ion isk also plays a ole because igh
ca bon policy a ec s he pe o mance o i ms ha do no pa icipa e in he EU ETS. The
p esence o a ansi ion isk is also suppo ed by esea ch showing ha ca bon p icing shocks
ha e a nega i e e ec on s ock e u ns o b own i ms (Be hold e al., 2023). This s and o he
li e a u e highligh s he di e en e ec o ca bon emission in equilib ium and in he a e ma h
o shocks. In equilib ium, g een asse s a e expec ed o ha e lowe e u n because hey make i
possible o hedge clima e isk. Howe e , g een asse s a e expec ed o ou pe o m b own asse s
in he a e ma h o shocks ha e eal he isk associa ed wi h he la e ( o a heo e ical model
ha cla i ies his di e ence see Pás o e al., 2021) .
The second s and is ela ed o he e ec o news, especially en i onmen al news, on s ock
e u ns. Resea ch ha analyzes he link be ween news and s ock e u ns goes back o a leas
Cu le e al. (1989), who cas doub s on he idea ha s ock p ice mo emen s a e ully d i en
by news abou u u e cash lows and discoun a es. Mo e ecen wo k uses sen imen analysis
o news o show ha media one has an impac on s ock p ices and ha he impac is gene ally
la ge o nega i e news (see, among o he s, Ga cia, 2013, Zhang e al., 2016, F aibe ge e al.,
2021). Jeon e al. (2022) show ha he sensi i i y o s ock e u ns o news has been inc easing
3
o e ime and i is s onge o i ms wi h high media isibili y.
Mo e closely ela ed o ou wo k a e s udies ha ocus on en i onmen al news. Capelle-
Blanca d and Laguna (2010) s udy s ock ma ke eac ions o indus ial disas e s by using da a
on 64 explosions in chemical plan s and e ine ies. They ind ha pe ochemical i ms expe ience
a sha p d op in ma ke alue in he wo days ha ollow he epo o an explosion. Along
simila lines, Ca pen ie and Su e (2015) use New Yo k Times epo ing o s udy how equi y
ma ke s espond o 161 majo en i onmen al and non-en i onmen al acciden s. Con a y o
p e ious s udies, hey do no ind a long-las ing e ec o en i onmen al acciden s on s ock
e u ns. Flamme (2013) uses co po a e news ela ed o en i onmen al issues o conduc an
e en s udy and inds a posi i e associa ion be ween he s ance o en i onmen al epo ing
and s ock e u ns. She also documen s ha his ela ionship has become asymme ic o e
ime, wi h a dec easing posi i e e ec o eco- iendly beha io and an inc easingly nega i e
e ec o eco-ha m ul beha io . Faccini e al. (2023) use ex ual and na a i e analysis o
clima e change- ela ed news and show ha clima e isk associa ed wi h go e nmen in e en ions
has an e ec on equi y e u ns in he Uni ed S a es. Engle e al. (2020) use ex analysis o
newspape co e age o clima e change o build clima e change isk-hedged po olios. A dia
e al. (2020) build a clima e change conce n index based on news published in US newspape s
and ind ha s ock e u ns o i ms exposed o his index a e associa ed wi h i m-le el ca bon
emission in ensi y. El Ouadghi i e al. (2021) use da a on clima e- ela ed disas e s oge he
wi h US news on clima e change and pollu ion and Google sea ches on hese e ms o show ha
public a en ion is posi i ely associa ed wi h e u ns o sus ainabili y- ocused s ock indices and
nega i ely associa ed wi h e u ns o con en ional s ock indices. Bessec and Fouquau (2022,
2024) also ocus on en i onmen al news co e age in US newspape s and use ex ual analysis o
classi y hese news i ems along di e en measu es o onali y and unce ain y. They ind ha
news eleases abou he en i onmen wi h ei he a nega i e o unce ain one lead o lowe and
mo e ola ile s ock e u ns o ca bon-in ensi e i ms.
To he bes o ou knowledge, we a e he i s o s udy how ca bon emission a ec he
link be ween en i onmen al news and s ock e u ns in he con ex o La in Ame ica and he
Ca ibbean, a egion ha is highly ulne able o he nega i e impac s o clima e change and
en i onmen al deg ada ion.
The es o he pape is o ganized as ollows. Sec ion 2desc ibes he da a, wi h special
ocus on ou pu pose-buil and no el se ies o clima e damage news. Sec ion 3desc ibes ou
empi ical s a egy and p esen s ou baseline es ima ions oge he wi h a ba e y o obus ness
checks. Sec ion 4 ocuses sou ces o he e ogenei y ac oss ypes o news and coun ies. Sec ion 5
discusses an al e na i e app oach (based on he e en s udy me hodology) o es ima ing how
ca bon emission in ensi y a ec s he ela ionship be ween en i onmen al damage news and s ock
e u ns. Sec ion 6concludes.
4
2 Da a
To s udy he link be ween en i onmen al damage news and s ock e u ns we me ge ou ypes
o da a: i) i m-le el yea ly inancial da a om Re ini i Da as eam; ii) i m-le el daily s ock
e u ns, also om Re ini i Da as eam; iii) i m-le el yea ly da a on ca bon emission in ensi y
om U gen em; and i ) coun y-le el daily coun da a on news ela ed o en i onmen al damage
om he Fac i a Snapsho s API.
The U gen em da ase p o ides i m-le el annual da a on emissions ca ego ized unde he
G eenhouse Gas P o ocol’s scopes 1, 2, and 3. Scope 1 emissions measu e di ec emissions om
sou ces con olled by he i m. Scope 2 emissions measu es indi ec emissions associa ed wi h he
pu chase o elec ici y, s eam, hea , o cooling. Scope 3 emissions encompass indi ec emissions
in a i m’s ups eam and downs eam alue chain. Fi m emissions a e epo ed annually. Ou
key a iable o in e es is i m-le el emission in ensi y, which is de ined as he sum o scopes 1
and 2 emissions exp essed in ons o ca bon dioxide equi alen ( CO2e) pe million dolla s o
e enue ( CO2e/$m e enue). Emissions da a span he 2009-2022 pe iod and co e 841 publicly-
lis ed i ms ac oss 18 sec o s in 16 coun ies in La in Ame ica and he Ca ibbean (Figu e 1). As
emissions a e disclosed wi h a one-yea lag, ou empi ical analysis uses lagged emissions.
The Fac i a Snapsho s Applica ion P og amming In e ace (API) e ie es speci ic his o -
ical a icles and news based on use -de ined c i e ia in e aced wi h he Dow Jones p emium
publica ion a chi e o o e 8,500 licensed news sou ces. Que ies a e o mula ed using Fac i a’s
code iden i ie s, which ca ego ize a icles in o 2,012 egions (i.e., coun ies, s a es/p o inces,
municipali ies, ci ies, economic o poli ical unions, e c), 1,182 indus ies, 32 languages, 9,363
news sou ces, and 1,230 subjec s. We use e sion 12.0 (upda ed on No embe 8, 2022) o he
Dow Jones In elligen Iden i ie s (DJID). DJID is a sys em de eloped by Dow Jones ha uses
a p op ie a y classi ica ion and axonomy sys em o classi y and ag he con en o he Fac i a
business in elligence pla o m. Thanks o i s in e nal consis ency and s anda dized classi ica-
ion, DJIS allows us o sea ch and e ie e da a om sou ces ha con ain a as amoun o
in o ma ion. The iden i ie s a e applied o Fac i a using a mix o au oma ed echnologies (in-
cluding A i icial In elligence) and a e hen manually checked by Dow Jones’ coding specialis s,
who ensu e ha he iden i ie s a e accu a e and consis en in 28 languages.
We a ge ed news classi ied unde he subjec s “Co po a e C ime/Legal Ac ion” and “Na -
u al En i onmen .” The i s subjec ocuses on co po a e c imes, legal in es iga ions, lawsui s,
and cou ulings. I includes cases in which a company is ei he he de endan o he plain i .
Na u al en i onmen is de ined as wildli e, clima e, and na u al esou ces. The news s o ies we
examine a e abou ac i i ies a ec ing he en i onmen and heal h haza ds ela ed o en i on-
men al issues. In o de o educe he isk o picking up posi i e en i onmen al news, we exclude
news i ems which a e pa o he wo subjec s men ioned abo e bu a e sub-classi ied unde
5
in es o s, yield esul s which a e essen ially iden ical o hose ob ained when we only included
i m-yea ixed e ec s. I any hing, he absolu e alue o he poin es ima e o βis sligh ly la ge
when we con ol o i m-qua e ixed e ec s (compa e columns 1-3 o Table 2wi h columns 4-6
o he same able and also he wo panels o Figu e 2).
As equa ion 2does no allow us o es ima e he main e ec s o ca bon emission and ED
News, we also es ima e a e sion o he model ha subs i u e he i m-yea (o i m-qua e )
and coun y-day ixed e ec s, wi h i m and ime ixed e ec s. The esul s o he in e ac i e
e m co obo a e he indings o Table 2, while he coe icien s o he main e ec s a e imp ecisely
es ima ed (see Appendix Table A5).
As men ioned, equa ion 2implici ly assumes ha ca bon emission in ensi y is he only i m-
speci ic a iable ha a ec s he ela ionship be ween ED News and s ock e u ns. We now elax
his assump ion by allowing o a iche se o in e ac i e e ec s and es ima e he ollowing
equa ion:
Ri(c),d(y)= (βCEIi(c),y−1+Xi(c),y−1Γ) ×Newsc,d(y)−1+ϕi(c),y +δd(y),c +εi(c),d(y)(2)
whe e Xi(c),y−1is a ma ix o ime- a ying i m cha ac e is ics ha a e po en ially co ela ed
wi h ca bon emission in ensi y ( he main e ec s o hese a iables a e cap u ed by he i m-yea
ixed e ec s ϕi(c),y).
We s a by augmen ing he model wi h he in e ac ion be ween ED news and i m size as
measu ed by he log o o al asse s (in cons an USD). We ind ha he in e ac i e coe icien
is nega i e and s a is ically signi ican (column 1 o Table A6). As la ge i ms a e likely o be
mo e isible, his esul is consis en wi h he indings o Jeon e al. (2022) ha he sensi i i y
o s ock e u ns is s onge o i ms wi h high media isibili y. Mo e impo an o ou pu poses
is he ac ha con olling o he in e ac ion be ween ED news and i m size has no e ec on
ou pa ame e o in e es : β emains nega i e, s a is ically signi ican and wi h a magni ude
which is basically iden ical o ha o ou baseline es ima es. Nex , we measu e i m size wi h
log u no e and again ind ha la ge i ms end o unde pe o m in he a e ma h o ED News
eleases. Ou pa ame e o in e es is unchanged (column 2 o Table A6). Columns 3 and 4
o Table A6 indica e ha mo e p o i able i ms (p o i abili y is measu ed wi h he ope a ing
ma gin in %) pe o m be e han he ma ke a e he elease o ED news, while le e age does
no ma e . Again, ou pa ame e o in e es do no change a e we include hese con ols. In
column 5, we augmen he model wi h he in e ac ion be ween ED News and a ich se o sec o
ixed e ec s. This exe cise se s a pa icula ly high ba because he sec o dummies abso b a
subs an ial amoun o he c oss- i m a iance in ca bon emission. None heless, ou esul s a e
obus o augmen ing he model wi h his la ge se o in e ac ion. Finally, we include all he
in e ac i e e ec s in he same model. As be o e, we ind ha la ge i ms end o ou pe o m
a e he elease o ED News and ha pa ame e o in e es does no change (column 6 o Table
A6). Resul s a e also obus o es ima ing he equa ions o Table A6 by subs i u ing i m-yea
12
ixed e ec s wi h qua e -yea ixed e ec s (see Table A6).
Nex , we explo e whe he he esul ha s ock e u ns o high ca bon in ensi y i m unde -
pe o m a e he elease o ED News is d i en by he ex ensi e o in ensi e ma gin. As a i s
s ep, we e-es ima e he models o Table 2by eplacing he dummy ha akes alue one i he e
is a leas one piece o ED News wi h a a iable ha measu es he ac ual numbe o ED News
in a gi en day. We ind esul s ha quali a i ely simila o he baseline es ima es (see Table
3). The poin es ima e o column 1 indica e ha one addi ional ED News, is associa ed wi h
s ock e u ns o i ms a he 75 h pe cen ile o ca bon emission in ensi y a e 0.12 basis poin s
lowe han s ock e u ns o i ms a he 25 h pe cen ile o he dis ibu ion (9% o a e age daily
e u ns).
As expec ed, he poin es ima es ob ained when using he con inuous measu e o news a e
smalle (because he a e age alue is la ge ), bu he s anda dized coe icien s a e basically
iden ical ( he coe icien o column 1 o Table 2scaled by he s anda d de ia ion o he News
dummy is 0.0025 and ha o column 1 o Table 3scaled by he s anda d de ia ion o numbe
o news is 0.0023).
We also in e ac i m-le el ca bon emission in ensi y wi h bo h he dummy and he con in-
uous measu e o ED News. The i s in e ac ion cap u es he ex ensi e ma gin and he second
he in ensi e ma gin. We ind ha bo h coe icien s a e nega i e and ha he wo coe icien s
a e join ly s a is ically signi ican (see he F es s a he bo om o Table 4). Howe e , only he
ex ensi e ma gin is indi idually s a is ically signi ican .
Taken a ace alue, he poin es ima es o column 1 Table 4indica e ha , on days when
he e is exac ly one ED news i em, s ock e u ns o i ms a he 75 h pe cen ile o ca bon
emission in ensi y a e 22 basis poin s lowe han s ock e u ns o i ms a he 25 h pe cen ile
o he dis ibu ion (16% o a e age daily e u ns); his di e ence inc eases o 29 basis poin s on
days wi h 2 pieces o ED News and o 46 basis poin s on days wi h 5 pieces o ED News.
13
Table 3: Baseline Es ima ions: In ensi e Ma gin
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news. News measu es he numbe o ED News eleased a ime . Columns 1 and 4 use he speci ica ion o
equa ion 2, columns 2-3 and 5-6 ha e a iche lag s uc u e. All eg essions include coun y-day ixed e ec s. The
models o columns 1-3 include i m-yea ixed e ec s and he models o columns 4-6 include i m-qua e ixed
e ec s. The s anda d e o s a e epo ed in pa en hesis and a e clus e ed a he i m le el
(1) (2) (3) (4) (5) (6)
CEIy−1×News −1-0.000442*** -0.000958*** -0.00121*** -0.000424*** -0.00104*** -0.00132***
(0.000110) (0.000165) (0.000215) (0.000107) (0.000181) (0.000224)
CEIy−1×News 0.000336* 0.000435 0.000335** 0.000427
(0.000193) (0.000380) (0.000156) (0.000312)
CEIy−1×News +1 0.000166 -1.55e-06 0.000135 -2.14e-05
(0.000150) (9.96e-05) (0.000115) (7.39e-05)
CEIy−1×News −26.76e-05 -8.59e-06
(0.000456) (0.000500)
CEIy−1×News +2 -8.56e-05 -0.000144
(0.000283) (0.000192)
Cons . -0.0953*** 0.788*** 1.157*** -0.0968*** 0.791*** 1.163***
(0.00886) (0.00980) (0.0120) (0.00864) (0.00657) (0.00750)
N. Obs. 856,232 698,293 645,165 856,232 698,293 645,165
R-squa ed 0.193 0.189 0.187 0.201 0.199 0.199
Time-coun y FE ✓ ✓ ✓ ✓ ✓ ✓
Fi m-yea FE ✓ ✓ ✓ x x x
Fi m-qua e FE x x x ✓ ✓ ✓
*** p<0.01, ** p<0.05, * p<0.1
14
Table 4: In ensi e e sus Ex ensi e Ma gin
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news. We use bo h a dummy ha akes alue one a e he elease o a leas one ED News pieces ( his a iable
cap u es he ex ensi e ma gin) and a a iable ha measu es he numbe o ED News pieces (in ensi e ma gin).
Column 1 includes ime-coun y and i m-yea ixed e ec s, Column 1 includes ime-coun y and i m-qua e
ixed e ec s, and column 3 include i m and ime ixed e ec s. The bo om panel o he able epo s he esul s
o a se ies o F- es s o he join signi icance o he ex ensi e and in ensi e ma gin. The s anda d e o s a e
epo ed in pa en hesis and a e clus e ed a he i m le el.
(1) (2) (3)
CEIy−1×News −1(Dummy) -0.000607*** -0.000774*** -0.000494***
(9.28e-05) (9.00e-05) (9.20e-05)
CEIy−1×News −1(Con inuous) -0.000215 -0.000132 -0.000240
(0.000142) (0.000123) (0.000146)
CEIy−1-5.23e-05***
(1.28e-05)
News −1(Dummy) -1.598
(1.026)
News −1(Con inuous) 1.169***
(0.436)
Cons . -0.0844*** -0.0830*** -0.0963
(0.00902) (0.0101) (0.105)
N. Obs 856,232 856,232 862,050
R-squa ed 0.193 0.201 0.102
Fi m FE x x ✓
Time FE x x ✓
Time-coun y FE ✓ ✓ x
Fi m-yea FE ✓x x
Fi m-qua e FE x ✓x
F es 55.66 40.62 26.86
p alue 0.000 0.000 0.00
*** p<0.01, ** p<0.05, * p<0.1
4 News and Coun y He e ogenei y
Ha ing es ablished ha s ocks o i ms wi h high ca bon emissions end o unde pe o m a e
he elease o ED News, we now explo e he e ogenei y ac oss ypes o news and whe he he
esul s a e d i en by a pa icula coun y.
As a i s s ep, we es ima e he models o columns 1 and 4 o Tables 2and 3sepa a ely o
domes ic, egional, and global news. We ind ha he in e ac i e e ms a e always nega i e, bu
ha hey a e s a is ically signi ican only o domes ic news (see Tables A8,A9, and A10, in he
Appendix; one coe icien is ma ginally signi ican in one o he ou global news eg ession). The
poin es ima es o he domes ic news eg ession a e simila (albei sligh ly la ge in absolu e
alue) o hose ob ained when using all news, while he poin es ima es o egional and global
news a e ma ginally smalle .
Nex , we un a ho se ace ha includes in e ac ions be ween i m-le el ca bon emissions and
15
he h ee ypes o news. We also ind in his case ha he in e ac i e e ms a e always nega i e
bu ha he only in e ac ion which is s a is ically signi ican is ha wi h domes ic news (see
Table 5). In ac he poin es ima es and s anda d e o s a e simila o hose ob ained when we
es ima ed he model wi h one ype o news a a ime (compa e Table 5wi h Tables A8-A10).
Table 5: Ho se ace Reg essions
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news om h ee sou ces: Domes ic newspape s and magazines, Regional newspape s and magazines, and Global
newspape s and magazines. All eg essions include coun y-day ixed e ec s. The models o columns 1 and 4
include i m-yea ixed e ec s and he models o columns 2 and 4 include i m-qua e ixed e ec s. Columns 1
and 2 use a dummy ha akes alue one a e he elease o a leas one ED News. Columns 3 and 4 use a measu e
o he numbe o ED news. The s anda d e o s a e epo ed in pa en hesis and a e clus e ed a he i m le el
(1) (2) (3) (4)
CEIy−1×Domes icNews −1-0.000963*** -0.000978*** -0.000425*** -0.000398**
(0.000307) (0.000315) (0.000159) (0.000156)
CEIy−1×RegionalNews −1-0.000611 -0.000682 -0.000637 -0.000722
(0.000839) (0.000780) (0.000685) (0.000637)
CEIy−1×W o ldNews −1-0.000783 -0.000892 -0.000727 -0.000814
(0.000494) (0.000608) (0.000469) (0.000558)
Cons . -0.0841*** -0.0828*** -0.0947*** -0.0960***
(0.0112) (0.0121) (0.00972) (0.0101)
N. Obs. 856,232 856,232 856,232 856,232
R-squa ed 0.193 0.201 0.193 0.201
Time-coun y FE ✓ ✓ ✓ ✓
Fi m-yea FE ✓x✓x
Fi m-qua e FE x ✓x✓
News is Dummy Dummy Con inuous Con inuous
Robus s anda d e o s clus e ed a i m le el in pa en heses
*** p<0.01, ** p<0.05, * p<0.1
One possible sou ce o conce ns is ha ou sample is domina ed by a small numbe o
la ge coun ies. B azilian i ms accoun o 43% o he obse a ions, and he op h ee coun ies
(B azil, Mexico, and Chile) accoun o 77% o obse a ions. To make su e ha he obse a ions
a e no d i en by an indi idual coun y o by a small se o coun ies, we e-es ima e he baseline
eg essions by weighing each obse a ion by he in e se o he numbe o obse a ions in ha
speci ic coun y. Thus, each obse a ion in ol ing B azilian i ms has a weigh o 1
368,215, and
each obse a ion in ol ing Panamanian i ms has a weigh o 1
5925. Wi h his weigh ing scheme,
each coun y has exac ly he same weigh . The esul s, epo ed in Table 6, a e essen ially
iden ical o he baseline indings o Tables 2and 3. They hus con i m ha ou indings a e no
d i en by a pa icula coun y o g oup o coun ies wi h a la ge numbe o i ms.
As a u he obus ness check we e-es ima e he baseline model (wi h and wi hou in e se
coun y weigh s) by d opping one coun y a a ime. The esul s con i m ha he indings
a e no d i en by jus one in luen ial coun y (see Appendix Table A11). The poin es ima es
become smalle (bu s ill s a is ically signi ican a he 1 pe cen con idence le el) when we
16
exclude B azil and la ge when we exclude Mexico ( his is he coun y wi h he second la ges
numbe o obse a ions, co esponding o 19% o he o al).
We also es ima e coun y-by-coun y eg essions o he 6 coun ies wi h he la ges numbe
o obse a ions (A gen ina, B azil, Chile, Colombia, Mexico, and Pe u). We ind ha he
coe icien o in e es is nega i e and s a is ically signi ican (o e y close o being s a is ically
signi ican in he case o A gen ina) in ou coun ies, nega i e bu a om being s a is ically
signi ican in Colombia and posi i e bu close o ze o in Chile (Appendix Table A12).
Table 6: Weigh ed Reg essions
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news. All eg essions include coun y-day ixed e ec s. The models o columns 1 and 4 include i m-yea ixed
e ec s and he models o columns 2 and 4 include i m-qua e ixed e ec s. Columns 1 and 2 use a dummy ha
akes alue one a e he elease o a leas one ED News piece. Columns 3 and 4 use a measu e o he numbe o
ED News pieces. Each obse a ion is weigh ed by he in e se o he numbe o obse a ions in i s speci ic coun y.
The s anda d e o s a e epo ed in pa en hesis and a e clus e ed a he i m le el
(1) (2) (3) (4)
CEIy−1×News −1-0.000868*** -0.000887*** -0.000372** -0.000352**
(0.000208) (0.000236) (0.000144) (0.000142)
Cons . 1.184*** 1.184*** 1.178*** 1.178***
(0.00417) (0.00471) (0.00463) (0.00455)
N. Obs 856,232 856,232 856,232 856,232
R-squa ed 0.287 0.294 0.287 0.294
Time-coun y FE ✓ ✓ ✓ ✓
Fi m-yea FE ✓x✓x
Fi m-qua e FE x ✓x✓
News is Dummy Dummy Con inuous Con inuous
Robus s anda d e o s in pa en hesis clus e ed a he i m le el
*** p<0.01, ** p<0.05, * p<0.1
5 E en S udy
As an al e na i e me hod o s udy how en i onmen al damage news a ec ca bon in ensi e
i ms, we use a wo-s ep e en -s udy app oach. In he i s s ep, we conduc a classic e en
s udy and compu e abno mal e u ns a ound he elease o en i onmen al damage news o all
i ms in ou sample. In he second s ep, we look a he ela ionship be ween hese abno mal
e u ns and ca bon emission in ensi y.
To es ima e abno mal e u ns associa ed wi h en i onmen al damages news eleased on day
T, we s a by es ima ing he ollowing eg ession o a 57 ading days window ha s a s a
ime T−60 and ends a ime T−3:
i(c),d =α+βmc,d +ui(c),d (3)
17
whe e i(c),d is he daily e u n o i m i(based in coun y c) on day dand mc,d is he ma ke
e u n in coun y con day d.
Unlike mos e en s udies, we ha e se e al e en s which a e close o each o he (as discussed
in he da a sec ion, on a e age, we ha e mo e han wo e en s pe mon h). The p oblem wi h
ha ing e en s close o each o he is ha hey con amina e he es ima ion window (in he ypical
e en s udy he e should no e en s in he es ima ion window). Conside , o ins ance, a si ua ion
in which he e is one e en on Ap il 15 and one e en on May 20. The es ima ion window o he
i s e en s a s in mid Janua y (app oxima ely 12 weeks be o e he e en ) and ends on Ap il
13, and he es ima ion window o he second e en s a s in mid Feb ua y and ends on May
17. The e is hus a subs an ial o e lap be ween he wo es ima ion windows and he e en o
Ap il 15 could a ec he pa ame e es ima es used o compu e he excess e u ns o he May 20
e en . In ou case, he p oblem is e en wo se, as we ha e many cases wi h es ima ions windows
ha include mul iple e en s.
One way o add ess his p oblem is o exclude all cases in which he e a e e en s which
o e lap wi h ano he e en ’s es ima ion window. The p oblem wi h his s a egy is ha we
would end up wi h a e y small numbe o e en s. As an in e media e s a egy, we allow o
some o e lap. Speci ically, we include in ou sample e en s wi h a maximum o e lap o 10 days
and exclude e en s wi h longe o e laps.5. This s a egy yields a sample o 28,461 i m-e en s.
We hen use he pa ame e es ima es o equa ion 3 o ob ain excess (“abno mal”) e u ns as
ou -o -sample o ecas e o o e a 5-day e en window ha s a s a T−2and ends a T+ 2:
a i(c),d = i(c),d −(ˆα+ˆ
βmc,d)(4)
and compu e a e age accumula ed a e age abno mal e u ns o e en E as:
AARi(c),E =1
5
T+2
X
d=T−2
a i(c),d.(5)
The a io AAR
σa √5(whe e σa is he s anda d de ia ion o ˆui(c),d in he es ima ion window) is a
- es on AAR.
As some o ou es ima es yield e y la ge excess e u n, we im ou da a a 5 pe cen o
he abno mal e u n a iable, and we a e le wi h a sample o 25,615 i m-e en s wi h a e age
excess e u ns ha ange be ween -1,500 and 1,500 basis poin s. Nea ly 20,000 (77 pe cen o
he o al) o hese es ima ed excess e u ns a e no s a is ically signi ican , 2,869 (11 pe cen o
he o al) a e posi i e and s a is ically signi ican a he 5 pe cen con idence le el, and 3,077
(12 pe cen o he o al) a e nega i e and s a is ically signi ican .
Ha ing buil ou sample o i m-e en abno mal e u ns, we a e now eady o es whe he
ca bon-in ensi e i ms a e mo e likely o expe ience nega i e abno mal e u ns a ound elease
5We explo e wi h di e en maximum o e laps and ob ain simila esul s
18
o en i onmen al damages news.6We s a by eg essing abno mal e u ns o e he log o
ca bon emission in ensi y and ind a nega i e bu no s a is ically signi ican coe icien (column
1 o Table 7). Lack o signi icance could be due o he ac ha ou dependen a iable is
imp ecisely es ima ed (77 pe cen o obse a ions a e no s a is ically signi ican , o en because
o la ge con idence in e al). One possible solu ion would be o concen a e he analysis on
s a is ically signi ican e u ns. Howe e , i would be a bi a y o include an obse a ion wi h a
- es o , say, 1.97 and exclude an obse a ion wi h a - es o 1.95. As an al e na i e, we escale
ou - es o ange be ween 0 and 1 and weigh each obse a ion by i s own - es . In his
way we gi e mo e weigh o p ecisely es ima ed abno mal e u ns and less weigh o abno mal
e u ns wi h a la ge con idence in e al. The weigh ed eg essions show a much s onge and
s a is ically signi ican nega i e ela ionship be ween ca bon emission in ensi y and abno mal
e u ns. The poin es ima e indica e ha a 1 pe cen inc ease in ca bon emission in ensi y
is associa ed wi h a 242 basis poin s dec ease in abno mal e u ns in he days ha su ound
he elease o en i onmen al damage news (column 2 o Table 7). The esul s a e obus o
subs i u ing he log o ca bon emissions wi h he le el o ca bon emissions (column 3 o Table
7). We ind quali a i ely simila esul s when we augmen he model wi h coun y and e en
ixed e ec s (columns 4 and 5 o Table 7). Howe e , con olling o ixed e ec s educes he
magni ude o he e ec o ca bon emission in ensi y on abno mal e u ns. The poin es ima es
now sugges ha a one pe cen inc ease in ca bon emissions is associa ed wi h a 65 basis poin s
dec ease in abno mal e u ns on he 5 days ha su ound he elease o en i onmen al damages
news.
To p obe u he , we use wo non-pa ame ic app oaches o allow o non-linea i ies in he
ela ionship be ween abno mal e u ns and ca bon emission in ensi y. We i s compu e a e age
abno mal e u ns a di e en poin s in he dis ibu ion o ca bon emissions. Speci ically, we
compu e - es weigh ed a e age abno mal e u ns o all i ms in he bo om 10 pe cen o he
dis ibu ion o ca bon emissions, and we hen mo e o i ms in he bo om 20, 30, 40 and 50
pe cen o he dis ibu ion o ca bon emissions. We also compu e a e age abno mal e u ns o
i ms in he op 50, 60, 70, 80 and 90 pe cen o he dis ibu ion o ca bon emissions.7Figu e 4
plo s he a e age alues wi h hei co esponding 95 pe cen con idence in e als and shows ha
i ms in he bo om pa o he dis ibu ion o ca bon emission end o ha e posi i e abno mal
e u ns, while i ms in he uppe pa o he dis ibu ion (and his is especially he case o i ms
in he op 20 h and 10 h pe cen ile) end o ha e nega i e abno mal e u ns.
6In o de o a oid ocusing on i ms wi h ex eme alues o ca bon emissions, we also im ca bon emissions
a 5 pe cen . We a e le wi h a sample o 23,119 i m-e en s. O hese i m-e en s, 5,327 (23 pe cen o he
o al) ha e s a is ically signi ican abno mal e u ns, 2,560 i m-e en s (11 pe cen o he o al) ha e posi i e
and s a is ically signi ican abno mal e u ns, and 2,767 (12 pe cen o he o al) ha e nega i e and s a is ically
signi ican abno mal e u ns.
7In p ac ice, we un - es -weigh ed eg essions wi h no con ols on di e en sub-samples and epo he
cons an and i s 95 pe cen con idence in e al.
19
Table 7: Abno mal Re u ns and Ca bon Emissions
This able plo s he esul o a se o eg ession in which abno mal e u ns a e eg essed on i m-yea le el ca bon
emissions. The eg essions o columns 2-5 a e weigh ed by he absolu e alue o abno mal e u n -s a is ics
escaled o ange be ween 0 and 1. The eg essions o columns 4-5 include coun y and e en ixed e ec s.
(1) (2) (3) (4) (5)
ln(CEI)-0.209 -2.419** -0.649**
(0.240) (1.057) (0.220)
CEI -0.017*** -0.005**
(0.004) (0.0018)
Cons an -2.478** 12.30*** 6.082*** 5.112*** 3.556***
(1.021) (3.990) (1.621) (0.883) (0.382)
N. Obs 23,119 23,119 23,119 23,058 23,058
R2 0.000 0.018 0.030 0.289 0.289
Weigh s x ✓ ✓ ✓ ✓
Coun y Fixed E ec s x x x ✓ ✓
E en Fixed e ec s x x x ✓ ✓
Robus s anda d e o s in pa en hesis clus e ed a coun y and e en le el
*** p<0.01, ** p<0.05, * p<0.1
Figu e 4: A e age Abno mal Re u ns a Di e en Le els o Ca bon Emissions
This igu e plo s a e age abno mal e u ns wi h 95% con idence in e als o subsamples o i ms a di e en le els
o ca bon emissions. CEI<P10 plo s a e age abno mal e u ns o i ms in he bo om 10% o he dis ibu ion
o ca bon emission in ensi y; CEI<P20 plo s a e age abno mal e u ns o i ms in he bo om 20% o he
dis ibu ion o ca bon emission in ensi y, and so on o CEI<P30, CEI<P40, and CEI<P50. CEI>P50 plo s
a e age abno mal e u ns o i ms in he op 50% o he dis ibu ion o ca bon emission in ensi y, CEI>P60 plo s
a e age abno mal e u ns o i ms in he op 60% o he dis ibu ion o ca bon emission in ensi y, and so on o
CEI>P70, CEI>P80, CEI>P90. All a e ages a e weigh ed by he absolu e alue o abno mal e u n -s a is ics
escaled o ange be ween 0 and 1.
-15
-12.5
-10
-7.5
-5
-2.5
0
2.5
5
7.5
10
CEI<P10
CEI<P20
CEI<P30
CEI<P40
CEI<P50
CEI>P50
CEI>P60
CEI>P70
CEIP80
CEI>P90
We also es ima e a se ies o eg essions using dummies ha ake alue one a di e en poin s
o he dis ibu ion o ca bon emissions. Fo ins ance, he eg ession epo ed in column 1 o Table
20
8includes a dummy ha ake alue one o all i ms in he bo om 10 pe cen o he dis ibu ion
o ca bon emission and alue 0 o all o he i ms. Simila ly, he eg ession o column 2 uses
a dummy ha akes alue one o all he i ms in he bo om 20 pe cen in he dis ibu ions.
Columns 3 and 4 ocus on he di e en ex eme o he dis ibu ion and use dummies ha ake
alue 1 o i ms in he op 20 and 10 pe cen o he dis ibu ion o ca bon emissions. Columns
5 and 10 join ly include he bo om 10 pe cen and op 10 pe cen dummies.
The in e p e a ion o he eg essions o Table 8is di e en om ha o he a e age alues o
Figu e 4. In he la e case, we a e looking a a e age alues in di e en poin s o he dis ibu ion
o ca bon emission in ensi y. In he o me , we a e es ing o he di e ence o a e age alues a
di e en poin s o he dis ibu ion. Fo ins ance, he poin es ima es o column 1 in Table 8 ell
us ha a e age abno mal e u ns a ound he elease o en i onmen al damage news o i ms in
he bo om 10 pe cen o he dis ibu ion o ca bon emission in ensi y a e 75 basis poin s highe
han abno mal e u ns in he emaining 90 pe cen o i ms, bu he di e ence is no s a is ically
signi ican . Columns 3 and 4, ins ead, indica e ha abno mal e u ns a e signi ican ly lowe o
i ms in he op 20 and 10 pe cen o he dis ibu ion o ca bon emission in ensi y ( he di e ence
is 1,200 and 1,500 basis poin s, espec i ely). Column 5 shows ha abno mal e u ns o i ms
in he bo om 10 pe cen o he dis ibu ion o ca bon emission in ensi y a e highe (bu no
signi ican ly highe ) han hose o i ms a he 10 h-90 h pe cen ile o he dis ibu ion o ca bon
emission in ensi y and abno mal e u ns o i ms in he op 10 pe cen a e signi ican ly lowe
han hose o i ms a he 10 h-90 h pe cen ile o he dis ibu ion o ca bon emission in ensi y.
Columns 6-10 o Table 8a e quali a i ely simila when we con ol o coun y and e en ixed
e ec s. Howe e , con olling o hese a iables esul s in lowe poin es ima es. Fo ins ance,
he di e ence in abno mal e u ns be ween i ms in he op 10 h pe cen ile and i ms in he
bo om 90 h pe cen ile goes om 1,500 o 400 basis poin s.
21
Table A5: Es ima ions wi h Main E ec s
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iables a e i m-yea le el ca bon emissions, coun y-day-le els ED News, and he in e ac ion be ween hese
wo a iables. Columns 2 and 3 also include a iche lag s uc u e. All eg essions include i m and ime ixed
e ec s. The s anda d e o s a e epo ed in pa en hesis and a e clus e ed a he i m le el.
(1) (2) (3)
CEIy−1×News −1-0.000866*** -0.000834*** -0.00123***
(0.000199) (0.000143) (0.000196)
CEIy−1-5.20e-05*** -6.00e-05 -5.39e-05
(1.29e-05) (7.58e-05) (5.45e-05)
News −10.434 -0.933 -4.242***
(0.725) (1.112) (1.304)
CEIy−1×News 0.000601 0.000558
(0.000464) (0.000499)
CEIy−1×News +1 0.000382 0.000337
(0.000456) (0.000458)
News 5.553*** 3.833***
(1.150) (1.271)
News +1 4.899*** 3.527***
(1.024) (1.169)
CEIy−1×News −20.000206
(0.000746)
CEIy−1×News +2 8.13e-05
(0.000552)
News −2-6.661***
(1.330)
News +2 -3.560***
(1.373)
Cons . -0.0975 0.223* 1.431***
(0.105) (0.127) (0.158)
N. Obs 862,050 704,073 650,929
R-squa ed 0.102 0.100 0.098
Fi m FE ✓ ✓ ✓
Time FE ✓ ✓ ✓
*** p<0.01, ** p<0.05, * p<0.1
28
Table A6: Mul iple Fi m-Le el In e ac ions ( i m-yea FE)
The sec o s a e: Accommoda ion and ood (S1); Adminis a ion and Suppo (S2); Ag icul u e, o es y and
ishing (S3); Cons uc ion (S4); Educa ion (S5); Elec ici y, gas and s eam (S6); En e ainmen (S7); Financial
and insu ance (S8); Heal h and social wo k (S9); In o ma ion and communica ion (S10); Manu ac u ing (S11);
Mining (S12); P o essional and echnical (S13); Public adminis a ion (S14); Real es a e (S15); T anspo and
s o age (S16); Wa e , sewage, and wa e managemen (S17); Wholesale and e ail (excluded g oup).
(1) (2) (3) (4) (5) (6)
CEIy−1×News −1-0.0009*** -0.0009*** -0.0009*** -0.0009*** -0.0009*** -0.0009***
(0.00017) (0.00017) (0.00017) (0.00018) (0.00018) (0.00020)
ln(Asse sy−1)×News −1-1.414*** -2.474***
(0.352) (0.868)
ln(T u ny−1)×News −1-0.982*** 0.977
(0.350) (0.850)
Op. Ma g.y−1×News −10.00019*** 1.61e-05
(0.00001) (0.0006)
Le e agey−1×News −1-0.00306 -0.00290
(0.0114) (0.0118)
S1×News −11.290 0.166
(4.525) (4.980)
S2×News −13.855 7.802
(4.492) (4.965)
S3×News −1-9.716* -7.952
(5.688) (5.119)
S4×News −12.446 3.546
(3.362) (3.530)
S5×News −15.538 6.296
(5.547) (5.791)
S6×News −1-2.156 -0.658
(2.711) (3.035)
S7×News −1-1.339 -2.835
(5.986) (5.978)
S8×News −1-2.197 2.443
(2.771) (3.398)
S9×News −1-2.837 -4.510
(4.263) (5.075)
S10 ×News −12.253 4.201
(3.511) (3.756)
S11 ×News −1-0.216 -0.233
(2.666) (2.742)
S12 ×News −11.590 2.837
(3.469) (3.884)
S13 ×News −1-4.192 -3.135
(4.568) (5.791)
S14 ×News −1-3.801 -4.936*
(2.678) (2.809)
S15 ×News −14.061 5.797
(3.215) (4.143)
S16 ×News −10.392 2.785
(3.853) (4.200)
S17 ×News −11.763 5.934*
(3.586) (3.355)
N. Obs 794,286 785,416 782,125 761,958 856,232 753,871
R-squa ed 0.205 0.205 0.205 0.210 0.193 0.209
Time-coun y FE ✓ ✓ ✓ ✓ ✓ ✓
Fi m-yea FE ✓ ✓ ✓ ✓ ✓ ✓
Robus s anda d e o s in pa en heses; *** p<0.01, ** p<0.05, * p<0.1
29
Table A7: Mul iple Fi m-Le el In e ac ions ( i m-qua e FE)
The sec o s a e: Accommoda ion and ood (S1); Adminis a ion and Suppo (S2); Ag icul u e, o es y and
ishing (S3); Cons uc ion (S4); Educa ion (S5); Elec ici y, gas and s eam (S6); En e ainmen (S7); Financial
and insu ance (S8); Heal h and social wo k (S9); In o ma ion and communica ion (S10); Manu ac u ing (S11);
Mining (S12); P o essional and echnical (S13); Public adminis a ion (S14); Real es a e (S15); T anspo and
s o age (S16); Wa e , sewage, and wa e managemen (S17); Wholesale and e ail (excluded g oup).
(1) (2) (3) (4) (5) (6)
CEIy−1×News −1-0.0009*** -0.001*** -0.001*** -0.001*** -0.001*** -0.0009***
(0.0002) (0.0002) (0.0002) (0.00019) (0.00021) (0.00022)
ln(Asse sy−1)×News −1-1.243*** -2.372***
(0.363) (0.883)
ln(T u ny−1)×News −1-0.840** 1.048
(0.353) (0.876)
Op. Ma g.y−1×News −10.00014** -0.000165
(0.00001) (0.00067)
Le e agey−1×News −1-0.00614 -0.00572
(0.0111) (0.0115)
S1×News −11.180 0.184
(4.850) (5.306)
S2×News −12.103 6.421
(4.582) (5.136)
S3×News −1-8.903 -6.845
(5.480) (4.802)
S4×News −12.725 4.104
(3.358) (3.483)
S5×News −13.822 4.776
(5.555) (5.833)
S6×News −1-1.748 -0.176
(2.719) (3.024)
S7×News −1-3.334 -4.286
(5.111) (5.178)
S8×News −1-1.789 2.936
(2.751) (3.328)
S9×News −1-2.580 -3.875
(4.760) (5.561)
S10 ×News −13.828 5.928
(3.408) (3.631)
S11 ×News −10.587 0.738
(2.643) (2.707)
S12 ×News −13.877 5.547
(3.525) (3.982)
S13 ×News −1-4.058 -3.553
(4.495) (6.231)
S14 ×News −1-6.109** -7.181**
(2.695) (2.803)
S15 ×News −14.728 6.754
(3.221) (4.126)
S16 ×News −10.544 3.276
(3.935) (4.225)
S17 ×News −11.536 5.610
(3.796) (3.859)
N. Obs 794,286 785,416 782,125 761,958 856,232 753,871
R-squa ed 0.213 0.214 0.214 0.218 0.201 0.217
Time-coun y FE ✓ ✓ ✓ ✓ ✓ ✓
Fi m-qua e FE ✓ ✓ ✓ ✓ ✓ ✓
Robus s anda d e o s in pa en heses; *** p<0.01, ** p<0.05, * p<0.1
30
Table A8: Domes ic News
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news in domes ic newspape s and magazines. All eg essions include coun y-day ixed e ec s. The models o
columns 1 and 4 include i m-yea ixed e ec s and he models o columns 2 and 4 include i m-qua e ixed e ec s.
Columns 1 and 2 use a dummy ha akes alue one a e he elease o a leas one ED News piece. Columns 3
and 4 use a measu e o he numbe o ED News pieces. The s anda d e o s a e epo ed in pa en hesis and a e
clus e ed a he i m le el.
(1) (2) (3) (4)
CEIy−1×News −1-0.000977*** -0.000992*** -0.000427*** -0.000399***
(0.000254) (0.000262) (0.000148) (0.000145)
Cons . -0.0882*** -0.0875*** -0.100*** -0.102***
(0.0111) (0.0115) (0.0108) (0.0106)
N. Obs 856,232 856,232 856,232 856,232
R-squa ed 0.193 0.201 0.193 0.201
Time-coun y FE ✓✓✓✓
Fi m-yea FE ✓x✓x
Fi m-qua e FE x ✓x✓
News is Dummy Dummy Con inuous Con inuous
*** p<0.01, ** p<0.05, * p<0.1
Table A9: Regional News
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news in egional newspape s and magazines. All eg essions include coun y-day ixed e ec s. The models o
columns 1 and 4 include i m-yea ixed e ec s and he models o columns 2 and 4 include i m-qua e ixed
e ec s. Columns 1 and 2 use a dummy ha akes alue one a e he elease o a leas one ED News piece.
Columns 3 and 4 use a measu e o he numbe o ED News pieces. The s anda d e o s a e epo ed in pa en hesis
and a e clus e ed a he i m le el.
(1) (2) (3) (4)
CEIy−1×News −1-0.000771 -0.000832 -0.000675 -0.000756
(0.000863) (0.000818) (0.000724) (0.000676)
Cons . -0.128*** -0.128*** -0.128*** -0.127***
(0.00373) (0.00354) (0.00355) (0.00331)
N. Obs 856,232 856,232 856,232 856,232
R-squa ed 0.193 0.201 0.193 0.201
Time-coun y FE ✓ ✓ ✓ ✓
Fi m-yea FE ✓x✓x
Fi m-qua e FE x ✓x✓
News is Dummy Dummy Con inuous Con inuous
Robus s anda d e o s clus e ed a i m-le el in pa en heses
*** p<0.01, ** p<0.05, * p<0.1
31
Table A10: Global News
This able epo s a se o eg essions whe e he dependen a iable is daily s ock e u ns and he main con ol
a iable is he in e ac ion be ween i m-le el ca bon emission in ensi y and he elease o en i onmen al damage
news in global newspape s and magazines. All eg essions include coun y-day ixed e ec s. The models o
columns 1 and 4 include i m-yea ixed e ec s, and he models o columns 2 and 4 include i m-qua e ixed
e ec s. Columns 1 and 2 use a dummy ha akes alue one a e he elease o a leas one ED News piece.
Columns 3 and 4 use a measu e o he numbe o ED News pieces. The s anda d e o s a e epo ed in pa en hesis
and a e clus e ed a he i m le el.
(1) (2) (3) (4)
CEIy−1×News −1-0.000690* -0.000817 -0.000672 -0.000780
(0.000417) (0.000524) (0.000421) (0.000514)
Cons . -0.129*** -0.129*** -0.129*** -0.129***
(0.00113) (0.00142) (0.00133) (0.00162)
N. Obs. 856,232 856,232 856,232 856,232
R-squa ed 0.193 0.201 0.193 0.201
Time-coun y FE ✓ ✓ ✓ ✓
Fi m-yea FE ✓x✓x
Fi m-qua e FE x ✓x✓
News is Dummy Dummy Con inuous Con inuous
Robus s anda d e o s clus e ed a i m-le el in pa en heses
*** p<0.01, ** p<0.05, * p<0.1
Table A11: Reg ession Resul s wi h Coun y Exclusions
This able epo s he coe icien s and -s a is ics o he in e ac ion be ween i m-le el ca bon emission in ensi y
and he elease o en i onmen al damage news using he same models as column 4 o Table 2and column 2 o
Table 6bu by d opping one coun y a a ime.
Excluded No Weigh ed Weigh ed
Coun y Coe icien s a is ics Coe icien s a is ics N. Obs
A gen ina -0.0010 4.691 *** -0.0008 3.292 *** 4,102,850
Bahamas -0.0010 4.975 *** -0.0009 3.768 *** 4,441,666
B azil -0.0005 3.303 *** -0.0005 2.608 ** 2,869,348
Chile -0.0010 5.262 *** -0.0009 3.900 *** 3,679,330
Colombia -0.0010 4.947 *** -0.0009 3.680 *** 4,251,082
Cos a Rica -0.0010 4.974 *** -0.0009 3.767 *** 4,441,666
Dominican Republic -0.0010 4.973 *** -0.0009 3.766 *** 4,446,960
Ecuado -0.0010 4.983 *** -0.0009 4.105 *** 4,436,372
Hondu as -0.0010 4.973 *** -0.0009 3.766 *** 4,446,960
Jamaica -0.0010 4.973 *** -0.0009 3.766 *** 4,304,022
Mexico -0.0012 18.62 *** -0.0012 10.15 *** 3,626,390
Panama -0.0010 4.976 *** -0.0009 3.787 *** 4,394,020
Pe u -0.0010 4.876 *** -0.0009 3.584 *** 4,092,262
T inidad and Tobago -0.0010 4.971 *** -0.0009 3.741 *** 4,372,844
U uguay -0.0010 4.973 *** -0.0009 3.766 *** 4,446,960
Venezuela -0.0010 4.974 *** -0.0009 3.774 *** 4,436,372
32
Table A12: Coun y by Coun y Reg essions
This able epo s he coe icien s and -s a is ics o he in e ac ion be ween i m-le el ca bon emission in ensi y
and he elease o en i onmen al damage news based on coun y-by-coun y eg essions using he same models as
column 4 o Table 2bu wi h ime ixed e ec s ins ead o coun y- ime ixed e ec s.
Coun y Coe icien p alue
A gen ina -0.0131 0.108
B azil -0.0012 0.000
Chile 0.0001 0.909
Colombia -0.0030 0.479
Mexico -0.0004 0.013
Pe u -0.0025 0.099
33
Appendix B: Fac i a P o essional News Examples
B1: Re iew o he Top 15 News Resul s Ob ained om ou Selec ed Que y
Code in Fac i a P o essional
Figu e A1: Domes ic News o Chile
The selec ed que y opic il e is: ((‘Na u al En i onmen ’ and ‘Co po a e C ime/Legal Ac ion’) and no (‘Ca bon
Seques a ion’ o ‘Ene gy E iciency’ o ‘En i onmen al P o ec ion’)) Fil e c i e ia (including selec ed que y
opic): Sou ces: All domes ic news sou ces om Chile, Rela ed coun y a icle con en : Chile, Da e ange:
01/01/2009 – 12/31/2022, Language: Spanish
34
Figu e A2: Domes ic News o Panama
The selec ed que y opic il e is: ((‘Na u al En i onmen ’ and ‘Co po a e C ime/Legal Ac ion’) and no (‘Ca bon
Seques a ion’ o ‘Ene gy E iciency’ o ‘En i onmen al P o ec ion’)). Fil e c i e ia (including selec ed que y
opic): Sou ces: All domes ic news sou ces om Panama, Rela ed coun y a icle con en : Panama, Da e ange:
01/01/2009 – 12/31/2022, Language: Spanish
35
B2: Re iew o News A icles Iden i ied h ough API News Coun s using Fac-
i a P o essional
Figu e A3: Fil e ed news sea ch based on API coun esul s: domes ic sou ces o Chile
The selec ed que y opic il e is: ((‘Na u al En i onmen ’ and ‘Co po a e C ime/Legal Ac ion’) and no (‘Ca bon
Seques a ion’ o ‘Ene gy E iciency’ o ‘En i onmen al P o ec ion’)). Fil e c i e ia (including selec ed que y
opic): Sou ce: Domes ic news sou ce “La Te ce a,” Rela ed coun y a icle con en : Chile, Da e: 11/07/2022
No e: This news a icle was ansla ed om Spanish (o iginal) o English, powe ed by Google T ansla e.
36
Figu e A4: Fil e ed News Sea ch Based on API Coun Resul s: Domes ic Sou ces o Hondu as
The selec ed que y opic il e is: ((‘Na u al En i onmen ’ and ‘Co po a e C ime/Legal Ac ion’) and no (‘Ca bon
Seques a ion’ o ‘Ene gy E iciency’ o ‘En i onmen al P o ec ion’)) Fil e c i e ia (including selec ed que y
opic): Sou ce: Domes ic news sou ce “C i e io,” Rela ed coun y a icle con en : Hondu as, Da e: 07/28/ 2022
No e: This news a icle was ansla ed om Spanish (o iginal) o English, powe ed by Google T ansla e.
Figu e A5: Fil e ed News Sea ch Based on API Coun Resul s: Regional Sou ces o A gen ina
Fil e c i e ia (including selec ed que y opic): Sou ce: Regional news sou ce “El Me cu io (Chile),” Rela ed
coun y a icle con en : A gen ina, Da e: 10/23/2015
No e: This news a icle was ansla ed om Spanish (o iginal) o English, powe ed by Google T ansla e.
37
