B ückne , Ma kus; Ciminelli, Gab iele; Loayza, No man
Wo king Pape
Ex e nal shocks and labo ma ke e o ms in au oc acies
and democ acies: E idence om oil p ice wind alls
ADB Economics Wo king Pape Se ies, No. 752
P o ided in Coope a ion wi h:
Asian De elopmen Bank (ADB), Manila
Sugges ed Ci a ion: B ückne , Ma kus; Ciminelli, Gab iele; Loayza, No man (2024) : Ex e nal shocks
and labo ma ke e o ms in au oc acies and democ acies: E idence om oil p ice wind alls, ADB
Economics Wo king Pape Se ies, No. 752, Asian De elopmen Bank (ADB), Manila,
h ps://doi.o g/10.22617/WPS240528-2
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/310364
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. 752
No embe 2024
Ex e nal Shocks and Labo Ma ke Re o ms in Au oc acies and Democ acies
E idence om Oil P ice Wind alls
This pape explo es how oil p ice wind alls impac labo ma ke egula ion ac oss 83 coun ies om 1970
o 2014. Findings e eal ha wind all gains lead o de egula ion in au oc acies bu no democ acies, while
wind all losses igge de egula ion in democ acies. The pape examines how democ acies use wind all gains
o boos spending, whe eas au oc acies may ocus on en ex ac ion and inc easing economic e iciency.
De egula ion in democ acies du ing losses aligns wi h he c isis-induced e o m hypo hesis, linked o win
de ici s and heigh ened banking c isis isks.
Abou he Asian De elopmen Bank
ADB is commi ed o achie ing a p ospe ous, inclusi e, esilien , and sus ainable Asia and he Paci ic,
while sus aining i s e o s o e adica e ex eme po e y. Es ablished in 1966, i is owned by 69 membe s
—49 om he egion. I s main ins umen s o helping i s de eloping membe coun ies a e policy dialogue,
loans, equi y in es men s, gua an ees, g an s, and echnical assis ance.
EXTERNAL SHOCKS
AND LABOR MARKET
REFORMS IN AUTOCRACIES
AND DEMOCRACIES
EVIDENCE FROM OIL PRICE WINDFALLS
Ma kus B ueckne , Gab iele Ciminelli, and No man Loayza
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
Ma kus B ueckne , Gab iele Ciminelli,
and No man Loayza
No. 752 | No embe 2024
Ma kus B ueckne ([email p o ec ed])
is a p o esso and he head o he Economics
P og am in he Resea ch School o Economics,
Aus alian Na ional Uni e si y. Gab iele Ciminelli
([email protected] g) is an economis a he Economic
Resea ch and De elopmen Impac Depa men ,
Asian De elopmen Bank. No man Loayza
(nloayza@wo ldbank.o g) is he di ec o o he Global
Indica o s G oup o he Wo ld Bank.
Ex e nal Shocks and Labo Ma ke Re o ms in Au oc acies
and Democ acies: E idence om Oil P ice Wind alls
C ea i e Commons A ibu ion 3.0 IGO license (CC BY 3.0 IGO)
© 2024 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 2024.
ISSN 2313-6537 (p in ), 2313-6545 (PDF)
Publica ion S ock No. WPS240528-2
DOI: h p://dx.doi.o g/10.22617/WPS240528-2
The iews exp essed in his publica ion a e hose o he au ho s and do no necessa ily e lec he iews and policies
o he Asian De elopmen Bank (ADB) o i s Boa d o Go e no s o he go e nmen s hey ep esen .
ADB does no gua an ee he accu acy o he da a included in his publica ion and accep s no esponsibili y o any
consequence o hei use. The men ion o speci ic companies o p oduc s o manu ac u e s does no imply ha hey
a e endo sed o ecommended by ADB in p e e ence o o he s o a simila na u e ha a e no men ioned.
By making any designa ion o o e e ence o a pa icula e i o y o geog aphic a ea in his documen , ADB does no
in end o make any judgmen s as o he legal o o he s a us o any e i o y o a ea.
This publica ion is a ailable unde he C ea i e Commons A ibu ion 3.0 IGO license (CC BY 3.0 IGO)
h ps://c ea i ecommons.o g/licenses/by/3.0/igo/. By using he con en o his publica ion, you ag ee o be bound
by he e ms o his license. Fo a ibu ion, ansla ions, adap a ions, and pe missions, please ead he p o isions
and e ms o use a h ps://www.adb.o g/ e ms-use#openaccess.
This CC license does no apply o non-ADB copy igh ma e ials in his publica ion. I he ma e ial is a ibu ed
oano he sou ce, please con ac he copy igh owne o publishe o ha sou ce o pe mission o ep oduce i .
ADB canno be held liable o any claims ha a ise as a esul o you use o he ma e ial.
Please con ac pubsma ke [email protected] g i you ha e ques ions o commen s wi h espec o con en , o i you wish
oob ain copy igh pe mission o you in ended use ha does no all wi hin hese e ms, o o pe mission o use
heADB logo.
Co igenda o ADB publica ions may be ound a h p://www.adb.o g/publica ions/co igenda.
ABSTRACT
We examine he ela ionship be ween oil p ice wind alls and labo ma ke egula ion
empi ically h ough panel eg essions in a sample o 83 coun ies spanning 1970–2014.
We ind ha oil p ice wind all gains lead o a de egula ion o he labo ma ke in
au oc acies bu ha e no e ec s in democ acies. Wind all losses ins ead cause a sizeable
de egula ion in democ acies bu ha e limi ed e ec s in au oc acies. We hen conside
possible ansmission channels. Democ acies appea o edis ibu e he en s s emming
om a posi i e wind all by inc easing go e nmen expendi u e. Ren ex ac ion and
economic e iciency conside a ions a e ins ead bo h plausible de egula ion d i e s
ollowing wind all gains in au oc acies, as expendi u es a e no aised, while g oss
domes ic p oduc and employmen g adually inc ease a e posi i e wind alls. Finally, he
de egula ion ollowing wind all losses in democ acies is consis en wi h he c isis-induced-
e o m hypo hesis, as wind all losses de e io a e he cu en accoun and budge balances
and inc ease he p obabili y o a sys emic banking c isis.
Keywo ds: oil p ice, wind alls, labo ma ke , de egula ion, poli ical ins i u ions
JEL codes: F16, J41, O13, P11, P16, Q02
__________________________
This pape p esen s he indings and conclusions o he au ho s and does no necessa ily ep esen he
iews o he o ganiza ions o which hey a e a ilia ed. We a e g a e ul o he commen s and sugges ions
ecei ed a he IMF-RESDM and he ADB ERDI Semina Se ies.
I. In oduc ion
The e ec s o na u al esou ces and ex e nal shocks on economic g ow h and ins i u ions
is an impo an opic in de elopmen economics. In his pape , we con ibu e o he
li e a u e by examining speci ically he e ec s o oil p ice wind alls on labo ma ke
ins i u ions. To he bes o ou knowledge, we a e he i s o do so.
A majo hu dle ha p e en ed p og ess in he pas on answe ing he ques ion o
how na u al esou ces a ec labo ma ke ins i u ions is he a ailabili y o da a on he la e
o a la ge se o coun ies and yea s. To his end, we use a no el da ase on employmen
p o ec ion legisla ion om Ciminelli and Fu ce i ( o hcoming). This da ase measu es he
ex en o he es ic ions aced by employe s when hey wan o e mina e an inde ini e
employmen con ac . Ou measu e o oil p ice wind alls is an oil expo p ice index, which
we cons uc ollowing con en ional p ac ice as he in e ac ion be ween a coun y’s
a e age g oss domes ic p oduc (GDP) sha e o pe oleum expo s and he annual g ow h
a e o he in e na ional oil p ice (e.g., A ezki and B ueckne 2012). Excluding om he
sample majo oil expo e s such as he Russian Fede a ion and Saudi A abia, we a e able
o es ima e plausibly causal e ec s o oil p ice wind alls on employmen p o ec ion
egula ion.
We s a conside ing educed o m es ima es. Ou main inding is ha he e ec o
oil p ice wind alls on employmen p o ec ion legisla ion depends on poli ical ins i u ions.
Oil p ice wind all gains lead o a signi ican de egula ion o he labo ma ke in au oc acies
while hey do no ha e any signi ican e ec in democ acies. On he o he hand, oil p ice
wind all losses lead o a sizeable de egula ion o he labo ma ke in democ acies bu
ha e only limi ed e ec s in au oc acies. These empi ical esul s a e pe sis en o e he
medium e m and a e obus o a a ie y o model speci ica ions and es ima ion
echniques ha a e s anda d in he li e a u e. Mo eo e , he di e ence be ween he
e ec s o posi i e and nega i e oil p ice wind alls is s a is ically signi ican . Quan i a i ely,
he e ec s o oil p ice wind alls on labo ma ke ins i u ions a e impo an , pa icula ly in
democ acies. A 1-s anda d-de ia ion posi i e wind all induces a de egula ion e o m
wo h abou a six h o he median e o m in au oc acies, while a 1-s anda d-de ia ion
nega i e wind all induces a de egula ion e o m oughly equal o he median e o m in
democ acies.
2
In he second s ep, we explo e po en ial ansmission channels ha migh d i e
ou esul s. Oil is a capi al-in ensi e commodi y. Acco ding o s anda d in e na ional ade
heo y (i.e., he S olpe -Samuelson e ec ), when ma ke s a e pe ec ly compe i i e an
inc ease in he in e na ional oil p ice should inc ease he en al a e ela i e o wages,
and mo e so, he la ge he GDP sha e o oil expo s. Howe e , when labo ma ke s a e
no pe ec ly compe i i e, e.g., due o labo ma ke egula ion, he p edic ions om his
heo y a e no s aigh o wa d. One possible explana ion o ou esul s is ha , in
au oc acies, well-connec ed capi alis inside s lobby he egime ollowing oil p ice
wind alls o libe alize he labo ma ke o educe wo ke s’ ba gaining powe , he eby
keeping wage p essu es in check and app op ia ing a la ge sha e o he en .
A he same ime, one o he main goals o de egula ing he labo ma ke is o gi e
employe s mo e lexibili y in deciding he op imal alloca ion o labo , which should
inc ease e iciency and may a ac highe in es men s. Hence, an al e na i e
in e p e a ion is ha e iciency gains ma e ialize when he e is a de egula ion o he labo
ma ke . Such an in e p e a ion would be in line wi h he iew ha compe i i e ma ke s
achie e e icien alloca ion. Howe e , de egula ing he labo ma ke can cause
employmen losses in he sho e m, as employe s can ake ad an age o he new ules
o dismiss he mos unp oduc i e wo ke s (Caccia o e and Fio i 2016). Oil p ice wind alls
may hen o e a good window o oppo uni y o go e nmen s o implemen a labo ma ke
de egula ion, as implemen ing such e o ms du ing a boom migh educe i s sho - un
cos s.
De egula ing he labo ma ke can also lead o a lowe ba gained wage (Ciminelli,
Du al, and Fu ce i 2022). Hence, some go e nmen s, pa icula ly hose in coun ies in
which democ a ic ins i u ions a e s onge , may be eluc an o implemen a de egula ion
e o m, e en a e pe iods o oil p ice wind alls, o no aliena e o e s. De egula ion
e o ms a e hen delayed un il a c isis hi s, when hey a e pe cei ed o be s ic ly
necessa y. Acco ding o his c isis-induced- e o m hypo hesis, democ a ic coun ies
implemen e o ms ha imp o e e iciency o he socie y o e all bu ha may ha e (high)
sho - e m cos s du ing c ises, when o e s unde s and ha he e may be ew al e na i es
o e o m (Tommasi and Velasco 1996, D azen and G illi 1993). Inso a as oil p ice
wind all losses induce a c isis, hey may be conduci e o a de egula ion e o m.
3
We es hese po en ial ansmission channels by expanding he analysis on he
e ec s o oil p ice wind alls o di e en dependen a iables. These addi ional es ima ions
sugges ha poli ical economy and e iciency conside a ions migh bo h be impo an in
explaining he esul ha au oc a ic egimes de egula e he labo ma ke ollowing posi i e
wind alls. We ind ha posi i e wind alls gene a e impo an en s, and ha he magni ude
o he labo ma ke de egula ion is la ge in mo e au oc a ic egimes. Taken oge he ,
hese esul s ein o ce he hypo hesis ha well-connec ed capi alis s lobby o
de egula ion o app op ia e a la ge sha e o he en s gene a ed by he wind all (o a
leas p e en ha wo ke s success ully ba gain o a highe wage). This poli ical economy
in e p e a ion is b oadly in line wi h iew ha he e a e syne gies be ween economic and
poli ical ins i u ions (e.g., Acemoglu and Robinson 2012); and ha he e is signi ican en
ex ac ion by capi alis s in au oc acies bu no in democ acies (B ueckne 2017).
A he same ime, we also ind ha ax e enues and he p ima y balance inc ease
on impac . GDP and he employmen a e inc ease g adually a e a posi i e wind all in
au oc acies while hey do no inc ease in democ acies. This g adual inc ease o
employmen and GDP in au oc acies migh be he esul o he labo ma ke de egula ion
implemen ed in esponse o he wind all. Gi en hese dynamics, we canno ule ou ha
go e nmen s in au oc a ic egimes unde s and ha posi i e wind alls p o ide a good
oppo uni y o libe alize he labo ma ke , hus s imula ing employmen and imp o ing
o e all economic e iciency in he medium e m, while minimizing sho - e m cos s. While
go e nmen s in democ a ic go e nmen s migh also be awa e ha passing a s uc u al
e o m in good imes can minimize i s sho - un cos s, hey migh s ill be unwilling o pass
such e o m i i is os acized by a majo i y o o e s.
Indeed, poli ical economy conside a ions appea o be e y impo an in explaining
he esponse o he labo ma ke egula ion index o oil p ice wind alls in democ acies. As
in au oc acies, in democ acies, bo h oil en s and ax e enues inc ease on impac
ollowing posi i e wind alls. Bu , di e en ly om au oc acies, go e nmen expendi u es
g adually inc ease a e posi i e wind alls, sugges ing ha democ a ic go e nmen s use
he highe ax e enues s emming om he wind all o edis ibu e pa s o he oil en s o
he wide popula ion. Tu ning o nega i e wind alls, we ind ha hese sha ply de e io a e
he cu en accoun and budge balances. The ac ha democ a ic coun ies implemen
4
labo ma ke de egula ion a e an oil p ice wind all loss ha signi ican ly de e io a es
public inances, as well as he coun y’s ex e nal posi ion suppo s he c isis-induced-
e o m hypo hesis, acco ding o which he poli ical economy o s uc u al e o ms in
democ a ic coun ies is such ha e o ms a e delayed un il a c isis hi s. We u he es
he alidi y o he c isis-induced- e o m hypo hesis by analyzing whe he oil p ice wind alls
a ec he p obabili y o expe iencing a c isis. We ind ha nega i e wind alls subs an ially
inc ease he p obabili y o a sys emic banking c isis in democ acies, while he e a e no
e ec s in o he cases.
Ou pape mainly con ibu es o he li e a u e on he mac oeconomic e ec s o
na u al esou ces and o ha examining he d i e s o economic e o ms. In he 1990s,
he mains eam iew was ha na u al esou ces, in pa icula oil weal h, is no bene icial
o a coun y’s economic de elopmen . Ea ly empi ical s udies, mos ly based on c oss-
coun y eg essions and case s udies, unco e ed a nega i e ela ionship be ween na u al
esou ce dependence and economic g ow h (e.g., Sachs and Wa ne 1995 and 1999).
The e we e also a gumen s made and empi ical e idence p o ided, mos ly in he o m o
c oss-coun y eg essions and case s udies, ha na u al esou ce dependence is
posi i ely co ela ed wi h co up ion, weak s a e capaci y, and he likelihood ha coun ies
a e uled by au oc a s (e.g., Gelb 1990, Ka l 1997, Ross 2001).
In he 2000s and 2010s he mains eam iew ha na u al esou ces a e a cu se
changed somewha . Alexee and Con ad (2009) showed ha when he dependen
a iable is he le el o GDP pe capi a, as a measu e o long- un li ing s anda ds, as
opposed o ansi ional g ow h, he e is no e idence o a esou ce cu se. B unnschweile
and Bul e (2008) showed ha na u al esou ce abundance, as opposed o na u al
esou ce dependence, is no signi ican ly associa ed wi h a highe isk o ci il con lic and
slowe economic g ow h. B ueckne and Ciccone (2010) documen ed ha commodi y
p ice wind alls a e, on a e age, associa ed wi h as e GDP g ow h in sub-Saha an A ica.
A numbe o subsequen empi ical s udies, using ec o au o eg essions (VARs) and
dynamic panel eg essions, unco e ed ha in he sho un, commodi y p ice wind alls
ha e posi i e e ec s on GDP g ow h (e.g., B ueckne , Tesei, and Ciccone 2012; Collie
and Gode is 2012; A aujo e al. 2014; Fe nandez, Schmid -G ohe, and U ibe 2020)
11
B. Local P ojec ions
The speci ica ions in Equa ions (1) and (2) ocus on he sho - e m impac o oil p ice
wind alls on EPL e o ms. To shed mo e ligh on he dynamics and pe sis ence o he
e ec s, we employ he local p ojec ion me hod. This me hod was pionee ed by Jo dà
(2005) and has been widely used he ea e (Aue bach and Go odnichenko 2012, Jo dà
and Taylo 2016, Ramey and Zubai y 2018, among many o he s). Plagbo g-Mølle and
Wol (2021) show ha local p ojec ions and VARs es ima e he same impulse esponses.
Olea and Plagbo g-Mølle (2021) show ha lag-augmen ed local p ojec ions yield
s anda d e o s ha a e asymp o ically alid.
In p ac ice, he local p ojec ion me hod en ails es ima ing he esponse o he
dependen a iable a pe iod +k o he shock a ime di ec ly. We conside a 5-yea
pe iod, including he yea o he shock and he ou ollowing ones. The e o e, o each
𝑘𝑘= 0, … ,4, we es ima e he ollowing speci ica ion:
𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖,𝑡𝑡+𝑘𝑘 −𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖,𝑡𝑡−1 =𝛾𝛾𝑖𝑖+𝜏𝜏𝑡𝑡+𝛽𝛽𝑘𝑘
𝐴𝐴∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴+𝛽𝛽𝑘𝑘
𝐷𝐷∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷+��𝛿𝛿𝑘𝑘,𝑗𝑗
𝐴𝐴∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡+𝑗𝑗
𝐴𝐴+𝛿𝛿𝑘𝑘,𝑗𝑗
𝐷𝐷∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡+𝑗𝑗
𝐷𝐷�
𝑘𝑘
𝑗𝑗=1
+∑�𝜌𝜌𝑘𝑘,𝑙𝑙
𝐴𝐴∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡−𝑙𝑙
𝐴𝐴+𝜌𝜌𝑘𝑘,𝑙𝑙
𝐷𝐷∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡−𝑙𝑙
𝐷𝐷�
2
𝑙𝑙=1 +ϑ𝑘𝑘𝑌𝑌𝑡𝑡−1 +𝜀𝜀𝑖𝑖,𝑡𝑡 (3)
whe e ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴 (∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷) deno e posi i e oil p ice wind alls in au oc acies (democ acies),
cons uc ed as ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡∗𝐴𝐴𝑖𝑖,𝑡𝑡 (∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡∗(1 −𝐴𝐴𝑖𝑖,𝑡𝑡)) wi h 𝐴𝐴𝑖𝑖,𝑡𝑡 being a dummy a iable aking
alue equal o 1 when he poli y2 sco e is below 0; 𝑌𝑌𝑡𝑡 is he log o pe capi a GDP in PPP,
included o con ol o he s age o de elopmen , and he es o he no a ion is as in
Equa ion (2).
The a iables in he summa ion e m a e o wa d shocks á la Teulings and
Zubano (2014), included o ake in o accoun o oil p ice wind alls ha occu wi hin he
𝑡𝑡+𝑘𝑘 ho izon, bu ha a e no cap u ed by he main explana o y a iable (∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡), and
whose omission may bias he esul s. The equa ion is made dynamically comple e by
12
including wo lags o he wind all a iables. The model includes coun y and ime ixed
e ec s, and he es ima ion is ca ied ou h ough OLS.3
We es ima e he model on he ull sample o au oc acies and democ acies,
allowing o di e en e ec s o oil p ice wind alls in hese wo g oups o egimes, a he
han spli ing he sample in wo. This is o maximize he numbe o obse a ions in each
coun y. The esul s would be e y simila i we we e o spli he sample in wo. The
coe icien s o in e es a e he 𝛽𝛽𝑘𝑘
𝐴𝐴s and 𝛽𝛽𝑘𝑘
𝐷𝐷s. They cap u e he e ec o oil p ice wind alls
a ime 𝑡𝑡 on he cumula i e change in EPL du ing he 𝑡𝑡+𝑘𝑘 ho izon in, espec i ely,
au oc acies and democ acies. The esul s a e p esen ed in he o m o impulse esponse
unc ions (IRFs), i.e., we plo he es ima ed coe icien s 𝛽𝛽𝑘𝑘
𝐴𝐴 and 𝛽𝛽𝑘𝑘
𝐷𝐷 and hei espec i e
90% con idence bands.
In he nex s ep, we le e age on he lexibili y o he local p ojec ion me hod o
analyze he e ec s o posi i e and nega i e oil p ice wind alls sepa a ely, gi en ha hese
may ha e di e en e ec s on EPL. We es ima e he ollowing eg ession speci ica ion:
𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖,𝑡𝑡+𝑘𝑘 −𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖,𝑡𝑡−1 =𝛾𝛾𝑖𝑖+𝜏𝜏𝑡𝑡+𝛽𝛽𝑘𝑘
𝐴𝐴,𝑃𝑃∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴,𝑃𝑃+𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴,𝑁𝑁+𝛽𝛽𝑘𝑘
𝐷𝐷,𝑃𝑃∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷,𝑃𝑃+𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷,𝑁𝑁
+��𝛿𝛿𝑘𝑘,𝑗𝑗
𝐴𝐴,𝑃𝑃∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡+𝑗𝑗
𝐴𝐴,𝑃𝑃+𝛿𝛿𝑘𝑘,𝑗𝑗
𝐴𝐴,𝑁𝑁∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡+𝑗𝑗
𝐴𝐴,𝑁𝑁+𝛿𝛿𝑘𝑘,𝑗𝑗
𝐷𝐷,𝑃𝑃∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡+𝑗𝑗
𝐷𝐷,𝑃𝑃+𝛿𝛿𝑘𝑘,𝑗𝑗
𝐷𝐷,𝑁𝑁∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡+𝑗𝑗
𝐷𝐷,𝑁𝑁�
𝑘𝑘
𝑗𝑗=1
+∑�𝜌𝜌𝑘𝑘,𝑙𝑙
𝐴𝐴,𝑃𝑃∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡−𝑙𝑙
𝐴𝐴,𝑃𝑃+𝜌𝜌𝑘𝑘,𝑙𝑙
𝐴𝐴,𝑁𝑁∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡−𝑙𝑙
𝐴𝐴,𝑁𝑁+𝜌𝜌𝑘𝑘,𝑙𝑙
𝐷𝐷,𝑃𝑃∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡−𝑙𝑙
𝐷𝐷,𝑃𝑃+𝜌𝜌𝑘𝑘,𝑙𝑙
𝐷𝐷,𝑁𝑁∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡−𝑙𝑙
𝐷𝐷,𝑁𝑁�
2
𝑙𝑙=1 +ϑ𝑘𝑘𝑌𝑌𝑡𝑡−1 +𝜀𝜀𝑖𝑖,𝑡𝑡 (4)
whe e ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴,𝑃𝑃 and ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷,𝑃𝑃deno e posi i e wind alls (i.e., gains) in au oc acies and
democ acies, espec i ely, cons uc ed by in e ac ing he ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴 and ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷 a iables
om Equa ion (3) wi h a dummy a iable aking alue equal o 1 o posi i e changes in
he oil p ice index. ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴,𝑁𝑁 and ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷,𝑁𝑁deno e nega i e wind alls (i.e., losses) in
au oc acies and democ acies, espec i ely, cons uc ed by in e ac ing he ∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐴𝐴 and
∆𝑂𝑂𝑂𝑂𝑂𝑂𝑖𝑖,𝑡𝑡
𝐷𝐷 a iables om Equa ion (3) wi h a dummy a iable aking alue equal o 0 o
nega i e changes in he oil p ice index. The coe icien s 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑃𝑃 and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑃𝑃 measu e he
3 We also conside ed a andom e ec s local p ojec ion model h ough GLS. The esul s using he wo
app oaches a e e y simila .
13
cumula i e e ec o posi i e oil p ice wind alls (i.e. gains) a ime 𝑡𝑡 on EPL e o ms o e
ho izon 𝑡𝑡+𝑘𝑘 in au oc acies and democ acies, espec i ely. The coe icien s 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁 and
𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁 measu e he cumula i e e ec o nega i e oil p ice wind alls (i.e. losses) a ime 𝑡𝑡 on
EPL e o ms o e ho izon 𝑡𝑡+𝑘𝑘 in au oc acies and democ acies, espec i ely.
IV. Baseline Resul s
A. Sho - un E ec s
Table 3 p esen s he baseline esul s, ob ained es ima ing Equa ions (1) and (2). The
esul s a e epo ed in Panels A and B, espec i ely o he sample o au oc acies and
democ acies. Columns (1) and (2) show esul s om a speci ica ion including only he
con empo aneous alue o he oil p ice wind all a iable. The es ima es in columns (1)
and (2) a e ob ained by es ima ing andom and ixed e ec s models, espec i ely.
Columns (3) and (4) epo esul s when we model oil p ice wind alls o a ec EPL wi h a
1-yea lag. Columns (5) and (6) show es ima es when including bo h he
con empo aneous and he lagged alue o he oil p ice wind all a iable in he model. The
epo ed coe icien s a e no malized o show he e ec o a 1-s anda d-de ia ion oil p ice
wind all (calcula ed on he un es ic ed sample o au oc acies and democ acies, equal o
abou 1.5). Equa ions (1) and (2) assume ha posi i e and nega i e oil p ice wind alls
ha e symme ic e ec s. Fo simplici y, we discuss he esul s e e ing o posi i e
wind alls. Fu he below, we allow o asymme ic e ec s depending on he sign o he
wind all.
We es ima e nega i e coe icien s o oil p ice wind alls in au oc acies, while he
coe icien s es ima ed o democ acies a e no s a is ically signi ican . The coe icien s
es ima ed o he con empo aneous oil p ice wind all a iable and i s 1-yea lag in
au oc acies a e abou -0.2 and -0.1, espec i ely. Bo h hese coe icien s a e s a is ically
signi ican a ei he he 95% o 99% con idence le el and a e b oadly consis en ac oss
he di e en speci ica ions and eg ession models conside ed.
The ac ha he wo coe icien s ha e he same sign and a e bo h signi ican
sugges s ha ei he he cumula i e e ec o oil p ice wind alls on labo ma ke egula ion
inc eases o e ime and/o ha some o he e o m ac ion happens wi h a lag.
14
Quan i a i ely, he es ima ed coe icien o -0.2 on yea oil p ice wind alls can be
in e p e ed as ollows: a 1-s anda d-de ia ion oil p ice wind all leads o a educ ion in he
EPL index o a ound 0.2 uni s, which is equi alen o a ound 0.1 s anda d de ia ions o
he one-pe iod change o he EPL index.
B. Medium- un E ec s
In he baseline analysis, we unco e ed a nega i e, sho - un, e ec o oil p ice wind alls
on he EPL index in au oc acies. Nex , o analyze he pe sis ence and dynamics o his
e ec , we use he local p ojec ion me hod. Th ough he local p ojec ion me hod, we ace
ou he esponse o EPL e o ms o oil p ice wind alls o e a 5-yea ho izon (including he
yea o he shock plus he ou ollowing ones). Figu e 3 depic s he cumula i e impulse
esponse unc ion (IRF) o EPL e o ms o a 1-s anda d-de ia ion inc ease in he oil p ice
index. The IRF is de i ed by es ima ing Equa ion (3). Panel A epo s esul s o
au oc acies (𝛽𝛽𝑘𝑘
𝐴𝐴 coe icien s), while Panel B shows esul s o democ acies (𝛽𝛽𝑘𝑘
𝐷𝐷
coe icien s). Blue solid lines a e poin es ima es, while dashed ed lines deno e 90%
con idence bands.
The IRFs in Figu e 3 indica e ha he e ec s o oil p ice wind alls on EPL e o ms
a e pe sis en , and e en inc ease, o e ime. In au oc acies, he e ec s a e concen a ed
a impac and 1 yea ollowing he shock. The cumula ed e ec 1 yea a e he shock is
abou -0.3 (signi ican ly di e en om ze o a he 99% con idence le el). The e ec sligh ly
dec eases o he es o he ho izon and becomes s a is ically insigni ican in he medium
e m. Tu ning o democ acies, we con i m ha oil p ice wind alls do no ha e any e ec s
on EPL in he sho un, bu we ins ead es ima e a posi i e coe icien in he long un. The
coe icien ha we es ima e o he 5-yea ho izon is equal o abou 0.4 and s a is ically
signi ican a he 90% con idence le el.
We d aw wo conclusions om his analysis. Fi s , he e ec s o oil p ice wind alls
on labo ma ke egula ion ake ime o ully ma e ialize. Second, he cumula i e e ec s
o e 5 yea s ha e he opposi e sign in au oc acies and democ acies, sugges ing ha
di e en channels migh be a play.
15
C. E ec s o Posi i e and Nega i e Wind alls
The analysis ca ied ou so a assumes ha posi i e and nega i e oil p ice changes
(wind all gains and losses) ha e symme ic e ec s on he EPL index. In wha ollows, we
es ima e Equa ion (4) o explo e whe he posi i e and nega i e changes in he oil p ice
index ha e di e en ial e ec s. Figu e 4 shows he new IRFs. Panels A1 and A2
espec i ely epo he e ec s o posi i e and nega i e wind alls in au oc acies (𝛽𝛽𝑘𝑘
𝐴𝐴,𝑃𝑃 and
𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁 coe icien s), while Panels B1 and B2 e e o democ acies (𝛽𝛽𝑘𝑘
𝐷𝐷,𝑃𝑃 and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁
coe icien s). To ease in e p e a ion, he IRFs epo ed in Panels A2 and B2 a e
cons uc ed using he nega i e alue o he 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁 and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁 coe icien s.
The new esul s indica e ha he e ec s o oil p ice wind alls es ima ed om he
es ic ed model in Equa ion (3) a e en i ely d i en by posi i e changes in he oil p ice
index in au oc acies. Tha is, oil p ice wind all gains cause au oc a ic egimes o
implemen employmen p o ec ion de egula ion e o ms (Panel A1). This e ec o posi i e
wind alls is la ge, s a is ically signi ican , and inc eases h oughou he ho izon
conside ed. A 1-s anda d-de ia ion posi i e shock induces a educ ion o he EPL index
wo h abou 0.6 o e a i e-yea ho izon (s a is ically signi ican a he 99% con idence
le el). To pu his in con ex , among au oc acies, he median easing e o m o e a 5-yea
pe iod is equal o 3.6. A educ ion in oil e enues due o a nega i e p ice g ow h also has
nega i e e ec s on he EPL index— ha is, i induces a de egula ion—bu his e ec is no
s a is ically signi ican (Panel A2).
The pic u e is di e en o democ acies. Posi i e wind alls do no ha e any e ec s,
while nega i e ones lead o an easing o labo ma ke egula ion. This e o ming ac ion
ma e ializes only g adually. A 1-s anda d-de ia ion nega i e wind all is es ima ed o
induce a educ ion o he EPL index o abou 0.2 on impac . This e ec g adually inc eases
o e all he ho izons conside ed, o each abou 1.3 4 yea s a e he shock (s a is ically
signi ican a he 95% con idence le el). This amoun s o an impo an e o m in he
con ex o democ a ic coun ies, jus below he median easing e o m implemen ed o e
a 5-yea pe iod.
This analysis sugges s ha posi i e and nega i e wind alls ha e di e en e ec s
on EPL and ha hese di e en ial e ec s di e among au oc acies and democ acies. We
16
co obo a e his asymme y esul by plo ing in Figu e 5 he di e ence be ween he e ec
o a posi i e wind all and ha o a nega i e wind all o e each yea o he ho izon
conside ed, using he coe icien s es ima ed om Equa ion (4). We ind ha , excep a
impac , he coe icien s es ima ed o posi i e and nega i e wind alls a e s a is ically
di e en om each o he a leas a he 90% con idence le el o bo h au oc acies and
democ acies.
In he nex sec ion, we explo e po en ial channels ha could explain he di e en ial
e ec s ha we es ima e. Be o e ha , howe e , we pe o m some addi ional es ima ions
o e i y he obus ness o ou esul s.
Robus ness Checks and Al e na i e Speci ica ions
We assess he sensi i i y o ou esul s o se e al di e en speci ica ions. Fi s , we check
ha ou esul s a e no d i en by indi idual coun ies and es ima e he medium- un e ec
ob ained om he local p ojec ion speci ica ion allowing o asymme ic e ec s be ween
posi i e and nega i e wind alls (Equa ion [4], wi h k=4) d opping one coun y a a ime.
Figu e 6 shows a sca e plo depic ing he new es ima es, wi h he y- and x-axis
espec i ely epo ing poin es ima es and p- alues. These a e close o he ull sample
baseline and con i m ha ou esul s a e no d i en by indi idual coun ies. Mos
O ganiza ion o he Pe oleum Expo ing Coun ies (OPEC) membe s a e excluded om
he analysis because hey a e no co e ed by he EPL da abase (Angola, Republic o he
Congo, Equa o ial Guinea, I an, I aq, Lybia, Qa a , Saudi A abia, Uni ed A ab Emi a es).
We u he e i y ha he esul s a e obus o excluding all emaining OPEC membe s in
Figu e 7.
As ano he exe cise, we check how he es ima es a e a ec ed unde di e en lag
speci ica ions. Speci ically, we es ima e a model including wo lags ( -1 and -2) o he i s
di e ence o he EPL index as well as a model including he i s lag ( -1) o he le el o
he EPL index. The i s model accoun s o po en ial se ial co ela ion in he da a—i he
dependen a iable is co ela ed o e ime, omi ing i s lags can lead o se ially co ela ed
eg ession e o s and biased and ine icien es ima es. The second model in oduces an
au o eg essi e componen o labo ma ke egula ion o cap u e he possibili y ha
changes in labo ma ke egula ion migh depend on he pas le el o egula ion.
17
Es ima es om hese addi ional speci ica ions a e p esen ed in Figu e 8. Those ob ained
con olling o pas changes in EPL a e i ually iden ical o he baseline es ima es, while
hose ob ained con olling o he ini ial le el o egula ion a e quali a i ely simila bu
display some quan i a i e di e ences. The medium- e m e ec o posi i e wind alls in
au oc acies and o nega i e wind alls in democ acies a e sligh ly a enua ed, bu he
newly es ima ed coe icien s a e s ill signi ican ly di e en om ze o a he 10% in e al.
The e ec o nega i e wind alls in au oc acies becomes la ge and no s a is ically
di e en om ze o, a sign ha nega i e wind alls may be an impo an d i e o labo
ma ke egula ion also in au oc acies. We also check ha ou esul s a e obus o (i)
using a andom e ec s a he han a ixed e ec s model, and (ii) es ima ing he e ec s o
au oc acies and democ acies by spli ing he sample in o wo g oups, au oc acies and
democ acies. The esul s om hese al e na i e speci ica ions a e e y close o ou
baseline (a ailable upon eques ).
Nex , we es whe he we ob ain simila esul s i we used a a iable measu ing he
ex en o oil disco e ies, de eloped by Co e and Tsui (2013), a he han he oil p ice
wind all a iable as an explana o y a iable. The esul s o oil disco e ies a e ai ly simila
o hose ha we ob ain o posi i e oil p ice wind alls (Figu e 9). Oil disco e ies induce a
la ge de egula ion o EPL in au oc acies whe eas hey do no ha e any e ec s in
democ acies. Somewha di e en ly om posi i e oil p ice wind alls, he e ec o oil
disco e ies in au oc acies is immedia e and ends o ade ou o e ime.
We also es whe he di iding he sample be ween (i) ad anced coun ies and (ii)
eme ging ma ke s and de eloping coun ies, a he han be ween au oc acies and
democ acies, gi e simila esul s o ou baseline, gi en ha he g oup o democ acies
con ains almos all ad anced coun ies and he e o e he e is some o e lap be ween he
wo. We es ima e a nega i e e ec o bo h posi i e and nega i e wind alls in eme ging
and de eloping coun ies (Figu e 10). Howe e , hese e ec s a e a he small and no
s a is ically signi ican in he case o nega i e wind alls. Fo he sample o ad anced
coun ies, he esponse o nega i e wind alls is simila o he case o democ acies, bu i
is less p ecisely es ima ed and weake . These esul s sugges ha poli ical ins i u ions
ma e mo e han economic de elopmen in de e mining he e ec o oil p ice wind alls on
labo ma ke egula ion.
18
We nex explo e whe he he e ec s o oil p ice wind alls change disc e ely ac oss
au oc acies and democ acies o whe he he e is a mo e linea ela ionship be ween he
deg ee o au oc acy/democ acy and wind alls. We a e able o do so by le e aging on he
poli y2 a iable. The poli y2 a iable a ies be ween –10 and 10. Mo e nega i e (posi i e)
poli y2 sco es deno e s onge au oc a ic (democ a ic) ins i u ions.
Figu e 11 epo s medium- un es ima es o he asymme ic e ec s o oil p ice
wind alls in he sample o au oc a ic and democ a ic coun ies as a unc ion o he poli y2
a iable. These es ima es a e ob ained by es ima ing an al e na i e speci ica ion o
Equa ion (4), o k=4, which includes on he igh -hand side o he es ima ing equa ion:
(i) he posi i e and nega i e oil p ice wind all a iables, (ii) he lagged poli y2 sco e, and
(iii) he in e ac ion be ween he oil p ice wind all a iables and he lagged poli y2 sco e.
Fo wa d and lagged alues o oil p ice wind alls a e also in e ac ed wi h he poli y2 sco e.
The main esul is ha he nega i e e ec o posi i e oil p ice wind alls in au oc acies a e
la ge he mo e he poli y2 a iable akes nega i e alues, while he nega i e esponse o
nega i e wind alls in democ acies is la ge he mo e he poli y2 a iable akes posi i e
alues.
In so-called he edi a y mona chies (poli y2=10), posi i e wind alls lead o a
signi ican easing o EPL, wo h abou -1 o he EPL index. The e ec o a poli y2 sco e
close o 0 is less han hal as la ge and no s a is ically signi ican . In egimes ha a e
democ a ic jus on he su ace (poli y2 sco e close o 0) he esponse o nega i e wind alls
is no s a is ically signi ican , while in he s onges democ a ic egimes (poli y2=10) he
esponse o a 1-s anda d-de ia ion nega i e wind all is abou -1.4 index poin s 4 yea s
a e he wind all and highly s a is ically signi ican . We also ind ha , among au oc a ic
egimes, he less he egime is au oc a ic he mo e i esponds o a nega i e wind all by
de egula ing he labo ma ke (simila o he esponse o democ a ic coun ies). Howe e ,
he es ima ed esponses o nega i e wind alls o au oc a ic coun ies a e no s a is ically
signi ican o any poli y2 sco e.
These esul s sugges ha poli ical economy conside a ions may be impo an in
d i ing he esponses o oil p ice wind alls, in bo h au oc acies and democ acies. The
mo e a egime is au oc a ic, he mo e he capi alis eli e is likely o be closely associa ed
wi h he go e nmen and lobby o libe alize he labo ma ke ollowing posi i e oil p ice
19
wind alls o app op ia e he oil en s. A he same ime, he mo e a go e nmen is
democ a ic, he mo e i will ind i poli ically unappealing o de egula e he labo ma ke s
in no mal imes, hus pos poning de egula ion in imes o c isis. We explo e poli ical
economy conside a ions as po en ial d i e s o ou esul s mo e ho oughly in Sec ion 5.
We close his analysis by conside ing he esponse o employmen p o ec ion
legisla ion o o he ypes o wind alls. In pa icula , we ocus on ag icul u al commodi ies,
which ypically gene a e less en s. Using he da ase assembled by Bazzi and Bla man
(2014), who collec p ice da a o 65 commodi ies and expo da a o majo coun ies in
A ica, he Middle Eas , Asia, and La in Ame ica, we cons uc wo a iables, measu ing
wind alls o he p oduc ion o annual and pe ennial ag icul u al commodi ies.4 Nex , we
es ima e al e na i e speci ica ions o Equa ion (4), in which we eplace ou baseline oil
p ice wind all a iables wi h hese al e na i e a iables. Figu e 10 p esen s new es ima es
o he e ec o annual ag icul u al commodi y wind alls on employmen p o ec ion
legisla ion. The e ec s a e null. E ec s o pe ennial ag icul u al wind alls a e also null
(a ailable upon eques ).
V. Channels
The analysis conduc ed so a has shown ha au oc a ic egimes educe he le el o labo
ma ke egula ion ollowing oil p ice wind all gains, while democ a ic go e nmen s educe
labo egula ion ollowing oil p ice wind all losses. We ind nega i e bu no signi ican
e ec s om wind all losses in au oc acies and no e ec s om wind all gains in
democ acies. In his sec ion, we explo e some po en ial ansmission channels d i ing
hese esul s. We i s discuss he po en ial ansmission channels, hen explo e how
wind alls a ec o he mac oeconomic a iables and he p obabili y o expe iencing a
sys emic banking c isis, and conclude wi h a b ie case s udy analysis.
4 These a iables exclude p ice make s.
20
A. Concep ual F amewo k
We ocus on wo se s o non-mu ually exclusi e explana ions, ela ed o poli ical economy
and e iciency conside a ions. One o he main goals o de egula ing he labo ma ke is
o gi e employe s mo e lexibili y in deciding he op imal alloca ion o labo . In his sense,
de egula ion imp o es labo ma ke e iciency and may a ac highe in es men s,
including om o eign in es o s. Howe e , he e a e some ca ea s. Fi s , de egula ing he
labo ma ke can cause employmen losses in he sho e m, as employe s can ake
ad an age o he new ules o dismiss he mos unp oduc i e wo ke s (Caccia o e and
Fio i 2016). Second, by educing wo ke s’ ba gaining powe , labo ma ke de egula ion
educes wages (Ciminelli, Du al, and Fu ce i 2022).5 Thi d, and possibly in pa due o
he e ec s jus desc ibed, de egula ing he labo ma ke is unpopula (Alesina e al. 2023).
Because o hese easons, some go e nmen s, pa icula ly hose in coun ies in which
democ a ic ins i u ions a e s onge , may be eluc an o implemen de egula ion e o ms
and delay hem un il when hey a e pe cei ed o be s ic ly necessa y.
Analyzing he e ec o oil p ice wind alls on labo ma ke e o ms adds ano he
laye o complexi y. I oil p ice wind all gains induce an expansion o economic ac i i y,
hey may o e a good window o oppo uni y o implemen a labo ma ke e o m, as such
e o ms end o ha e less sho - un cos s i done du ing good economic imes.6 On op o
his, poli ical economy conside a ions may also be ele an . Wo ke s in he oil and ela ed
5 The e exis se e al mechanisms h ough which a educ ion in EPL dec eases he su plus o he wo ke s.
The ea lie mac oeconomics li e a u e assumes EPL o ha e a di ec e ec on ba gaining powe (Blancha d
e al. 1997, Blancha d and Gia azzi 2003). Hence, easing EPL lowe s wo ke s’ ba gaining powe and
esul s in wo ke s cap u ing a lowe sha e o he ma ch su plus. Al e na i ely, EPL can be hough as a
was e ul i ing cos in a s anda d Diamond-Mo ensen-Pissa ides se ing (Mo ensen and Pissa ides 1994).
When wages a e e(nego ia ed), he i m gi es wo ke s a wage p emium o sa ing on i ing cos s oday
bu penalizes hem o ha ing o pay i ing cos s in he u u e (ins ance.g., Caccia o e and Fio i 2016). Tha
penal y is he discoun ed alue o u u e i ing cos s, which is less han he cu en alue due o ime
p e e ences. Hence, EPL inc eases he wo ke ’s sha e o he ma ch su plus and, in u n, a e o m easing
EPL dec eases i . EPL is also well-known o educe job u no e (Ben olila and Be ola 1990, Pissa ides
2000), which in u n weakens he abili y o wo ke s o ex ac he su plus om a ma ch (Cahuc, Pos el-
Vinay, and Robin 2006). Ciminelli e al. (2022) analyze he e ec o EPL e o ms on he wage le el and he
labo sha e and ind ha EPL de egula ion unambiguously lowe s he wage and lowe s he labo sha e
when labo and capi al a e ela i e complemen .
6 A labo ma ke de egula ion e o m implemen ed du ing good economic imes may lead employe s o
expand he labo inpu knowing ha hey will ha e mo e lexibili y o educe i in he u u e shall business
needs change (Du al e al. 2020).
27
The 1999 labo code e o m allowed o mo e lexible employmen con ac s, easie
dismissal p ocedu es, and sho - e m labo a angemen s. P essu e om domes ic eli es
and o eign oil companies played a signi ican ole in shaping he e o ms. Oil majo s
demanded labo lexibili y o manage hei ope a ions e icien ly, pa icula ly in capi al-
in ensi e p ojec s like he Tengiz oil ield. Thei in luence led o a labo ma ke model ha
p io i ized cos e iciency, lexibili y in hi ing and i ing, and he widesp ead use o
manpowe agencies (So bello 2023). The cen aliza ion o poli ical powe mean ha he
go e nmen could implemen labo de egula ion wi h minimal opposi ion om wo ke s o
unions, which we e o en weakened by op-down con ol. The EPL indica o o Ciminelli
and Fu ce i ( o hcoming) cap u es he 1999 e o m as a dec ease o 27 index poin s (ou
o a 0–100 scale). This is a la ge e o m, which alls wi hin he uppe decile o all
de egula ion e o ms in ou sample.
The C isis-Induced 2002 De egula ion Re o m in Colombia
The case o he labo ma ke de egula ion e o m decided by he Go e nmen o Colombia
in Decembe 2002 is e y di e en om ha o Kazakhs an jus e iewed, as i was
implemen ed in by a democ a ic go e nmen in esponse o a majo economic c isis ha
g ea ly inc eased he unemploymen a e. In he yea s leading up o he e o m, he oil
p ice c ashed by o e 60% o e 1997–1998. I la e slowly eco e ed h oughou 1999–
2002, eaching i s end-1996 le el in Decembe 2002, he mon h o he e o m. Gi en he
delay wi h which e o ms a e unde aken in democ acies in esponse o nega i e wind alls
(we es ima e he peak e ec o be 4 yea s a e he wind all), we see he 1997–1998 oil
p ice c ash o be he d i e o he libe alizing e o m.
Colombia has been p oducing oil since he ea ly 20 h cen u y and a he end o he
1990s was one o he leading oil p oduce s in La in Ame ica, eaching p oduc ion o
a ound 700,000 ba els pe day o oil in 1999 (Pe y and Oli e a 2009). A signi ican
po ion o Colombia’s oil p oduc ion was expo ed, making oil one o he coun y’s main
expo commodi ies and con ibu ing o a subs an ial pa o Colombia’s o eign exchange
ea nings. As a esul o he Asian inancial c isis o 1997 and o he Russian deb de aul
o 1998, global demand o oil and o he commodi ies dec eased. F om i s end-1996 peak
o i s end-1998 ough, he oil p ice c ashed by o e 60%. In 1999, Colombia expe ienced
28
one o i s wo s -e e economic c ises, wi h he GDP con ac ing by o e 4% and he
unemploymen a e ising o o e 20%, om jus o e 12% in 1997. The c isis also led o
oubles in he inancial sec o , wi h local banks acing signi ican ins abili y and sol ency
p oblems.
The economic c isis p ecipi a ed a sense o u gency o implemen economic
e o ms. By 2002, he unemploymen a e was s ill a ound 16%, wice as much he le el
in he mid-1990s. The 2002 labo ma ke e o m in Colombia, embodied in Law 789, was
mo i a ed by he need o add ess he coun y’s high unemploymen and labo ma ke
in o mali y. Gi en he na ow ocus o he indica o o Ciminelli and Fu ce i ( o hcoming)
on employmen p o ec ion legisla ion and he ac ha only se e ance paymen s we e
e o med in his a ea, he e o m is wo h jus a 2-index poin dec ease index. Bu he
e o m is conside ably mo e signi ican i o he a eas o labo ma ke egula ion a e
conside ed, as i in oduced g ea e labo ma ke lexibili y in hi ing and i ing by making
i easie o employe s o lay o wo ke s in empo a y con ac s, educed labo cos s by
modi ying o e ime pay ules, and incen i ized hi ing h ough app en iceship p og ams.
The success o he e o m is deba ed (Ga i ia 2005), bu Colombia did see a g adual
labo ma ke eco e y o e he subsequen yea s, wi h he unemploymen a e slowing o
11% by 2007.
VI. Conclusion
The ole o na u al esou ces in shaping ins i u ions is an impo an opic in de elopmen
economics. In his pape , we ocused on oil and ook a i s s ab a analyzing he e ec s
o oil p ice wind alls on labo ma ke egula ion. Basic in e na ional ade heo y sugges s
ha o an oil-expo ing coun y wi h pe ec ly compe i i e ac o ma ke s, an inc ease in
he in e na ional oil p ice leads o an inc ease in he en al a e ela i e o wages.
Howe e , i ma ke s a e no compe i i e, in pa icula , i he e is signi ican employmen
p o ec ion legisla ion, wo ke s ha e he ba gaining powe o bid up hei wages.
Combining a no el da abase o employmen p o ec ion legisla ion wi h in o ma ion
on plausibly exogenous oil p ice wind alls, we ound ha he e ec s o oil p ice wind alls
on labo ma ke ins i u ions a e di e en be ween au oc acies and democ acies. Among
29
he o me , oil p ice wind all gains lead o a weakening o employmen p o ec ion. This
e ec is pe sis en o e he medium e m. On he o he hand, oil p ice wind all gains ha e
no e ec on labo ma ke egula ion in democ a ic coun ies. A he same ime, oil p ice
wind all losses lead o subs an ial job p o ec ion de egula ion in democ acies, while hey
ha e limi ed e ec s in au oc acies.
We conside ed wo se s o explana ions, one based on poli ical economy and he
o he on e iciency conside a ions, as po en ial d i e s o ou esul s. One explana ion is
ha go e nmen s in au oc a ic egimes ecognize ha he e a e e iciency gains ha
ma e ialize when he e is a de egula ion o he labo ma ke . Such an in e p e a ion would
be in line wi h he iew ha compe i i e ma ke s achie e e icien alloca ion. Gi en ha
de egula ing he labo ma ke can cause employmen losses in he sho e m, oil p ice
wind alls may o e a good window o oppo uni y o go e nmen s o implemen labo
ma ke de egula ion, as implemen ing such e o ms du ing good economic imes migh
educe i s sho - un cos s.
The o he se o explana ions is based on poli ical economy conside a ions. On he
one hand, a e an oil wind all, capi al owne s may lobby he go e nmen o loosen up
labo ma ke egula ion o weaken he ba gaining powe o wo ke s and app op ia e a
la ge sha e o he en s s emming om he highe oil p ice. Gi en ha in au oc acies
elec ions ei he do no ake place o do no ca y a eal h ea o he uling eli e, such
lobbying ac i i ies may be success ul. In democ a ic coun ies, ins ead, go e ning pa ies
ace eelec ion and he e o e need o conside how much a weakening o labo ma ke
ins i u ions would dampen hei eelec ion p ospec s. Re o ms a e hen delayed un il hey
a e pe cei ed as a las - eso op ion, which is when a c isis hi s.
To shed ligh on hese po en ial ansmission channels, we wen beyond
employmen p o ec ion and explo ed he e ec o oil p ice wind alls on a ange o o he
ou come a iables. Posi i e oil p ice wind alls gene a e impo an en s. In democ acies,
ax e enues go up a e he wind all and go e nmen s inc ease expendi u es, possibly
o edis ibu e pa s o hese en s o he wide popula ion. In au oc acies, ins ead, he e
is no en edis ibu ion h ough highe expendi u es, bu we ind ha GDP and he
employmen a e bo h inc ease in he medium e m a e a posi i e wind all, which is no
he case in democ acies. These esul s sugges ha , while he esponse o democ acies
30
o a posi i e wind all seems o be en edis ibu ion, e iciency, and en ex ac ion can
bo h explain he de egula ion o he labo ma ke ollowing posi i e wind alls in
au oc acies. On he o he hand, he c isis-induced- e o m hypo hesis appea s o be a
alid explana ion o he de egula ion o he labo ma ke ollowing wind all losses in
democ acies, as we ind nega i e oil p ice wind alls o sha ply de e io a e he cu en
accoun and he budge balance and subs an ially inc ease he p obabili y o a sys emic
banking c isis in hese coun ies.
Employmen p o ec ion legisla ion is jus one aspec o labo ma ke ins i u ions.
We ocused on i o he sake o iden i ica ion pu poses, bu he poin is mo e gene al.
Fu u e esea ch should go beyond employmen p o ec ion and assess he ole o na u al
esou ces on o he ins i u ions a ec ing he dis ibu ion o income be ween labo and
capi al.
31
FIGURES AND TABLES
Figu e 1: E olu ion o Employmen P o ec ion Legisla ion o e Time
No es: Median and in e qua ile ange o he employmen p o ec ion legisla ion index o Ciminelli and Fu ce i
( o hcoming) o he samples o au oc a ic and democ a ic coun ies o e he ime sample conside ed in his pape .
The index anges om 0 o 1, wi h highe alues indica ing highe employmen p o ec ion.
Sou ce: Au ho s’ calcula ions based on da a om Ciminelli and Fu ce i ( o hcoming).
32
Figu e 2: F equency o Re o ms o e Time
(as % o o al obse a ions)
No es: F equency (in %) o nega i e changes (easing), no changes (s a us quo) and posi i e changes ( igh ening) o
he employmen p o ec ion legisla ion index o Ciminelli and Fu ce i ( o hcoming) o he sample o au oc a ic and
democ a ic coun ies and o each yea o he analysis. Highe alues o he index indica e highe employmen
p o ec ion.
Sou ce: Au ho s’ calcula ions based on da a om Ciminelli and Fu ce i ( o hcoming).
33
Figu e 3: Dynamic E ec s o Oil P ice Wind alls on he Employmen P o ec ion
Legisla ion Index
No es: E ec o a 1-s anda d-de ia ion oil p ice wind all on he employmen p o ec ion legisla ion index o e a 5-yea
ho izon in au oc acies (Panel A) and democ acies (Panel B). X-axes deno e he ho izon o he e ec , wi h 0 indica ing
he yea o he wind all, while Y-axes deno e he magni ude o he e ec . Blue solid lines epo he poin es ima es,
while ed dashed lines a e 90% con idence bands. Es ima es a e ob ained plo ing he 𝛽𝛽𝑘𝑘
𝐴𝐴 (Panel A) and 𝛽𝛽𝑘𝑘
𝐷𝐷 (Panel B)
coe icien s and hei s anda d e o s es ima ed om Equa ion (3).
Sou ce: Au ho s’ calcula ions.
34
Figu e 4: Asymme ic E ec s o Oil P ice Wind alls on he Employmen P o ec ion
Legisla ion Index
No es: E ec o a 1-s anda d-de ia ion nega i e and posi i e oil p ice wind alls on he employmen p o ec ion legisla ion
index o e a 5-yea ho izon in au oc acies and democ acies. Es ima es a e ob ained plo ing he 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑃𝑃 and 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁 (Panel
A) and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑃𝑃and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁 (Panel B) coe icien s and hei s anda d e o s es ima ed om Equa ion (4). See no es in Figu e
3.
Sou ce: Au ho s’ calcula ions.
35
Figu e 5: Di e en ial E ec s o Posi i e and Nega i e Oil P ice Wind alls
No es: Di e ence be ween he coe icien s es ima ed o he e ec o posi i e and nega i e wind alls on he employmen
p o ec ion legisla ion index o e a 5-yea ho izon in au oc acies and democ acies. Es ima es a e ob ained by plo ing
he 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑃𝑃-𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁 (Panel A) and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑃𝑃-𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁 (Panel B) combina ions o he coe icien s es ima ed om Equa ion (4). See no es
in Figu e 3.
Sou ce: Au ho s’ calcula ions.
36
Figu e 6: Robus ness Check on Coun y Sample
No es: Poin es ima es (y-axes) and p- alues (x-axes) o he medium- e m e ec o posi i e and nega i e oil p ice
wind alls on he employmen p o ec ion legisla ion index ob ained om a obus ness check exe cise in which each
coun y is excluded om he sample one a a ime. The es ima ing equa ion is Equa ion (4) wi h k=4. Blue ( ed) do s
epo es ima es o he 𝛽𝛽𝑘𝑘
𝐴𝐴,𝑃𝑃 (𝛽𝛽𝑘𝑘
𝐴𝐴,𝑁𝑁) (Panel A) and 𝛽𝛽𝑘𝑘
𝐷𝐷,𝑃𝑃(𝛽𝛽𝑘𝑘
𝐷𝐷,𝑁𝑁) (Panel B) coe icien s.
Sou ce: Au ho s’ calcula ions.
43
Table 1: Desc ip i e S a is ics o Employmen P o ec ion Legisla ion Index
A. AUTOCRACIES B. DEMOCRACIES
Obs.
Media
n
S.d. Min Max Obs.
Media
n
S.d. Min Max
Index 974 21.82 10.70 2.17 58.17 2327 23.25 11.48 2.00 58.15
Tigh ening 46 3.97 7.80 0.01 30.61 171 2.04 6.19 0.01 39.13
Easing 19 -3.70 6.73 -22.31 -0.20 65 -1.48 3.84 -23.33 -0.02
No es: The able epo s desc ip i e s a is ics o he employmen p o ec ion legisla ion index o Ciminelli and Fu ce i
( o hcoming) o he wo samples o au oc acies (poli y2 sco e below 0, Panel A) and democ acies (poli y2 sco e
abo e 0, Panel B). The ow deno ed by Index shows s a is ics o he le el o he index. Rows deno ed by easing
and igh ening epo s a is ics o , espec i ely, nega i e and posi i e changes o he index. The index anges on
a scale om 0 o 100, wi h highe alues indica ing mo e employmen p o ec ion.
Sou ce: Au ho s’ calcula ions based on da a om Ciminelli and Fu ce i ( o hcoming).
Table 2: Desc ip i e S a is ics o Oil P ice Wind alls
A. AUTOCRACIES B. DEMOCRACIES
Obs. Median S.d. Min Max Obs. Median S.d. Min Max
Posi i e 594 0.12 1.41 0.00 13.74 1378 0.10 0.83 0.00 12.77
Nega i e 410 -0.06 1.00 -6.99 0.00 952 0.04 0.55 -7.62 0.00
No es: The able epo s desc ip i e s a is ics o oil p ice wind alls o he wo samples o au oc acies (poli y2
sco e below 0, Panel A) and democ acies (poli y2 sco e abo e 0, Panel B).
Sou ce: Au ho s’ calcula ions.
44
Table 3: E ec o Oil P ice Wind alls on Changes in Employmen P o ec ion Legisla ion
A. AUTOCRACIES
(1) (2) (3) (4) (5) (6)
Oil wind all
-0.20***
-0.23**
-0.17**
-0.20***
(0.08)
(0.09)
(0.07)
(0.07)
Oil wind all,
-0.10**
-0.10***
-0.10***
-0.11***
lag 1
(0.04)
(0.04)
(0.04)
(0.04)
Obse a ions
915
915
884
884
878
878
R-squa ed
0.08
0.08
0.06
0.06
0.07
0.08
Model RE FE RE FE RE FE
B. DEMOCRACIES
(1) (2) (3) (4) (5) (6)
Oil wind all
0.08
0.05
0.08
0.04
(0.06)
(0.06)
(0.06)
(0.07)
Oil wind all,
-0.01
-0.04
-0.01
-0.04
lag 1
(0.04)
(0.04)
(0.04)
(0.04)
Obse a ions
2,278
2,278
2,243
2,243
2,243
2,243
R-squa ed
0.05
0.05
0.05
0.05
0.05
0.05
Model RE FE RE FE RE FE
No es: Es ima es o he con empo aneous and lagged e ec s o oil p ice wind alls on
changes o he employmen p o ec ion legisla ion index, ob ained es ima ing a andom
e ec s (Equa ion 1) and a ixed e ec s model (Equa ion 2) o he sample o au oc a ic
coun ies (poli y2 sco e below 0, Panel A) and democ a ic coun ies (poli y2 sco e abo e
0, Panel B).
All speci ica ions include ime ixed e ec s. *, **, *** deno e s a is ical
signi icance a he, espec i ely, 90%, 95%, and 99% con idence le el. S anda d e o s
(shown in pa en heses) a e Hube obus and clus e ed a he coun y le el.
Sou ce: Au ho s’ calcula ions.
45
Table 4: E ec s o Oil P ice Wind alls on Va ious Dependen Va iables
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
EPL
Ren
CA
GDP
Employmen
Re enues
Expendi u e
Budge
Impac Responses
Au +
-0.19**
0.79**
0.62**
0.01
0.09
0.37***
-0.16
0.55**
(0.09)
(0.31)
(0.30)
(0.23)
(0.09)
(0.12)
(0.20)
(0.22)
Au -
-0.04
-0.42**
-0.49**
-0.42*
0.05
-0.52***
0.43***
-0.94***
(0.11)
(0.17)
(0.22)
(0.22)
(0.07)
(0.13)
(0.11)
(0.18)
Dem +
0.02
0.74***
0.46***
-0.05
0.00
0.34**
0.16
0.21*
(0.17)
(0.26)
(0.16)
(0.13)
(0.11)
(0.14)
(0.12)
(0.11)
Dem -
-0.24
-0.87***
-1.51***
-0.09
0.07
-0.58**
0.43**
-1.01***
(0.18)
(0.14)
(0.51)
(0.38)
(0.12)
(0.26)
(0.21)
(0.25)
Obs
2,886
2,724
2,912
2,989
1,755
2,361
2,319
2,319
Adj. R2
0.025
0.366
0.086
0.136
0.052
0.066
0.071
0.149
3-yea esponses
Au +
-0.47***
0.34**
0.32
0.63
0.33***
0.45**
0.19
0.32**
(0.15)
(0.13)
(0.35)
(0.53)
(0.12)
(0.21)
(0.26)
(0.13)
Au -
-0.30
0.03
-0.08
-0.27
0.18
0.97
1.24**
-0.26
(0.40)
(0.28)
(0.21)
(0.34)
(0.17)
(0.70)
(0.60)
(0.33)
Dem +
-0.01
0.42***
0.08
0.36
0.09
0.41***
0.48**
-0.06
(0.27)
(0.14)
(0.19)
(0.33)
(0.11)
(0.14)
(0.22)
(0.14)
Dem -
-0.65**
-0.63***
-1.22***
0.02
0.14
-0.34
0.58
-0.93***
(0.32)
(0.17)
(0.33)
(0.53)
(0.20)
(0.33)
(0.37)
(0.25)
Obs.
2,804
2,645
2,829
2,906
1,674
2,285
2,240
2,240
Adj. R2
0.044
0.350
0.089
0.173
0.055
0.045
0.106
0.207
5-yea esponses
Au +
-0.60***
0.36**
-0.05
1.38**
0.54***
0.03
-0.12
0.16
(0.22)
(0.16)
(0.39)
(0.66)
(0.14)
(0.33)
(0.29)
(0.40)
Au -
-0.53
0.43
-0.77**
-0.21
-0.51*
0.85
0.95**
-0.12
(0.48)
(0.29)
(0.35)
(0.76)
(0.28)
(0.60)
(0.45)
(0.43)
Dem +
0.11
0.45***
-0.34
0.72
0.23
0.36*
0.56***
-0.18
(0.21)
(0.12)
(0.38)
(0.67)
(0.25)
(0.19)
(0.21)
(0.23)
Dem -
-1.27**
-0.03
-1.44**
0.41
-0.08
-0.24
0.61
-0.89
(0.50)
(0.16)
(0.58)
(0.65)
(0.33)
(0.38)
(0.73)
(0.62)
Obs.
2,640
2,487
2,665
2,740
1,512
2,135
2,092
2,092
Adj. R2
0.056
0.378
0.082
0.218
0.058
0.032
0.099
0.201
CA = cu en accoun balance, EPL = employmen p o ec ion legisla ion, GDP = g oss domes ic p oduc .
No es: E ec s o posi i e (+) and nega i e (-) oil p ice wind alls on a ange o dependen a iables in au oc acies (Au )
and democ acies (Dem), a impac as well as wo and ou yea s a e he wind alls. Es ima es a e ob ained es ima ing
Equa ion (4) o k=0,2,4, eplacing 𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖,𝑡𝑡 wi h (i) oil en s as a sha e o GDP (Column 2), (ii) he cu en accoun balance
as a sha e o GDP (Colum 3), (iii) 100 imes he log o eal pe capi a GDP in local cu ency (Column 4), (i ) he
employmen - o-popula ion a io (Column 5), ( ) ax e enues as a sha e o GDP (Column 6), ( i) go e nmen
expendi u es as a sha e o GDP (Column 7), and ( ii) he p ima y balance as a sha e o GDP (Column 8). *, **, ***
deno e s a is ical signi icance a he, espec i ely, 90%, 95%, and 99% con idence le el. S anda d e o s (shown in
pa en heses) a e Hube obus and clus e ed a he coun y le el.
Sou ce: Au ho s’ calcula ions.
46
Table 5: E ec s o Posi i e and Nega i e Oil P ice Wind alls on a Sys emic Banking
C isis P obabili y
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
P ob
P ob
P ob
Log
Log
Log
Con log
Con log
Con log
Au +
-0.00
-0.02
-0.00
0.02
-0.02
-0.02
0.02
-0.02
-0.02
(0.04)
(0.07)
(0.09)
(0.07)
(0.14)
(0.20)
(0.17)
(0.18)
(0.23)
Au -
0.04
-0.02
-0.06
0.09
0.02
-0.08
0.09
0.02
-0.07
(0.13)
(0.14)
(0.19)
(0.29)
(0.30)
(0.37)
(0.27)
(0.29)
(0.35)
Dem +
-0.19
-0.21
-0.37
-0.34
-0.40
-0.87
-0.34
-0.39
-0.84*
(0.14)
(0.14)
(0.28)
(0.26)
(0.28)
(0.60)
(0.30)
(0.32)
(0.50)
Dem -
0.33***
0.31**
0.38**
0.65***
0.66**
0.71**
0.65**
0.64**
0.67**
(0.12)
(0.14)
(0.16)
(0.23)
(0.26)
(0.30)
(0.27)
(0.29)
(0.32)
Obs.
2,882
2,620
2,108
2,882
2,620
2,108
2,882
2,620
2,620
Pseudo R
2
0.03
0.06
0.17
0.03
0.06
0.17
Time FE
NO
NO
YES
NO
NO
YES
NO
NO
YES
Coun y FE
NO
YES
YES
NO
YES
YES
NO
YES
YES
GDP = g oss domes ic p oduc , USD = Uni ed S a es dolla .
No es: E ec s o posi i e (+) and nega i e (-) oil p ice wind alls on he p obabili y o a sys emic banking c isis occu ing
in au oc acies (Au ) and democ acies (Dem). Es ima es a e ob ained by es ima ing p obi (P ob), logi (Log), and
condi ional logi (Con log) models. The dependen a iable akes alue 1 i a sys emic banking c isis occu s ei he a
yea o +1. Explana o y a iables include wind alls a yea , as well as o wa d wind all a iables (yea +1), lagged
wind all a iables (yea -1 and -2), and lagged GDP pe capi a in USD (yea -1). Time and coun y ixed e ec s a e
included depending on he speci ica ion. Coe icien s epo he e ec o wind all a iables a yea . *, **, *** deno e
s a is ical signi icance a he, espec i ely, 90%, 95%, and 99% con idence le el. S anda d e o s (shown in
pa en heses) a e clus e ed a he coun y le el.
Sou ce: Au ho s’ calcula ions.
47
Table 6: Ma ginal E ec s o Posi i e and Nega i e Oil P ice Wind alls on a Sys emic
Banking C isis P obabili y
(1)
(2)
(3)
(4)
(5)
(6)
P ob
P ob
P ob
Log
Log
Log
Au +
-0.00
-0.00
-0.00
0.00
-0.00
-0.00
(0.00)
(0.01)
(0.01)
(0.00)
(0.01)
(0.01)
Au -
0.00
-0.00
-0.01
0.00
0.00
-0.00
(0.01)
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
Dem +
-0.02
-0.02
-0.04
-0.02
-0.02
-0.04
(0.02)
(0.02)
(0.03)
(0.01)
(0.01)
(0.03)
Dem -
0.04***
0.04**
0.04**
0.04***
0.04***
0.03**
(0.01)
(0.02)
(0.02)
(0.01)
(0.01)
(0.01)
Obs.
2,882
2,620
2,108
2,882
2,620
2,108
Pseudo R
2
0.03
0.06
0.17
0.03
0.06
0.17
Time FE
NO
NO
YES
NO
NO
YES
Coun y FE
NO
YES
YES
NO
YES
YES
No es: Ma ginal e ec s o posi i e (+) and nega i e (-) oil p ice wind alls on he p obabili y o a sys emic banking c isis
occu ing in au oc acies (Au ) and democ acies (Dem). Ma ginal e ec s a e calcula ed a he means o he dependen
a iables. See no es in Table 5.
Sou ce: Au ho s’ calcula ions.
48
REFERENCES
Abiad, Abdul, and Ashoka Mody. 2005. “Financial Re o m: Wha Shakes I ? Wha Shapes
I ?” Ame ican Economic Re iew, 95 (1): 66–88.
Acemoglu Da on, and James A. Robinson. 2012. Why Na ions Fail: The O igins o Powe ,
P ospe i y, and Po e y.
Agnello, Luca, Vi o Cas o, João To a Jalles, and Rica do M. Sousa. 2015. “Wha
De e mines he Likelihood o S uc u al Re o ms?” Eu opean Jou nal o Poli ical
Economy, 37: 129–45.
Alesina, Albe o, Da ide Fu ce i, Jona han D. Os y, Ch is Papageo giou, and Dennis P.
Quinn. 2023. “S uc u al Re o ms and Elec ions: E idence om a Wo ld-Wide New
Da ase .” Jou nal o he Eu opean Economic Associa ion, 22 (4): 1936–80
Alexee , Michael, and Robe Con ad. 2009. “The Elusi e Cu se o Oil.” Re iew o
Economics and S a is ics 91 (3): 586–98.
A aujo, Jo ge A. De Thompson R., Eka e ina Vos oknu o a, Ma kus B ueckne , Ma eo
Cla ijo Munioz, and Kons an in M. Wacke . 2014. Beyond Commodi ies: The
g ow h challenge o La in Ame ica and he Ca ibbean. Wo ld Bank Repo No.
91015.
A ezki, Rabah, and Ma kus B ueckne . 2012. “Commodi y Wind alls, Democ acy and
Ex e nal Deb .” The Economic Jou nal 122, 561 (1): 848–66.
Aue bach, Alan J., and Yu iy Go odnichenko. 2012. “Fiscal Mul iplie s in Recession and
Expansion.” In Albe o Alesina and F ancesco Gia azzi, eds. Fiscal Policy A e he
Financial C isis. pp. 63–98. Na ional Bu eau o Economic Resea ch.
Bazzi, Samuel, and Ch is ophe Bla man. 2014. “Economic Shocks and Con lic :
E idence om Commodi y P ice Shocks.” Ame ican Economic Jou nal: Mac o 6: 1–
38.
Ben olila, Samuel, and Giuseppe Be ola. 1990. “Fi ing Cos s and Labou Demand: How
Bad Is Eu oscle osis?” The Re iew o Economic S udies, 57 (3): 381–402.
Blancha d, Oli ie , and F ancesco Gia azzi. 2003. “Mac oeconomic E ec s o Regula ion
and De egula ion in Goods and Labo Ma ke s.” The Qua e ly Jou nal o
Economics, 118 (3): 879–907.
Blancha d, Oli ie J., William D. No dhaus, and Edmund S. Phelps. 1997. “The Medium
Run.” B ookings Pape s on Economic Ac i i y, 1997 (2): 89–158.
49
B ueckne , Ma kus. 2017. “Ren Ex ac ion by Capi alis s.” Eu opean Jou nal o Poli ical
Economy, 50 (C), 157–70.
B ueckne , Ma kus, and An onio Ciccone. 2010. “In e na ional Commodi y P ices,
G ow h, and he Ou b eak o Ci il Wa in Sub-Saha an A ica.” Economic Jou nal,
120 (544): 519–34.
B ueckne , Ma kus, And ea Tesei, and An onio Ciccone. 2012. “Oil P ice Shocks,
Income, and Democ acy.” Re iew o Economics and S a is ics 94 (2): 389–99.
B unnschweile , C. and E win Bul e. 2008. “Linking Na u al Resou ces o Slow G ow h
and Mo e Con lic .” Science, 320 (5876): 616–7.
Caccia o e, Ma eo, and Giuseppe Fio i. 2016. “The Mac oeconomic E ec s o Goods
and Labo Ma ke s De egula ion.” Re iew o Economic Dynamics, 20: 1–24.
Cahuc, Pie e, Fabien Pos el-Vinay, and Jean-Ma c Robin. 2006. “Wage Ba gaining wi h
On- he-Job Sea ch: Theo y and E idence.” Econome ica, 74 (2): 323–64.
Ciminelli, Gab iele, Romain Du al, and Da ide Fu ce i. 2022. “Employmen P o ec ion
De egula ion and Labo Sha es in Ad anced Economies.” The Re iew o
Economics and S a is ics, 104 (6): 1174–90.
Ciminelli, Gab iele, and Da ide Fu ce i. Fo hcoming. “A New Da ase o Employmen
P o ec ion Legisla ion.”
Collie , Paul, and Benedik Gode is. 2012. “Commodi y P ices and G ow h: An Empi ical
In es iga ion.” Eu opean Economic Re iew, 56 (6): 1241–60.
Co e , Anca M., and Ke in K. Tsui. 2013. “Oil Disco e ies and Con lic : Wha Does he
C oss-Coun y E idence Really Show?” Ame ican Economic Jou nal:
Mac oeconomics, 5 (1): 49–80.
D azen, Allan, and William Eas e ly. 2001. “Do C ises Induce Re o m? Simple Empi ical
Tes s o Con en ional Wisdom.” Economics and Poli ics, 13 (2): 129–57.
Du al, Romain, Da ide Fu ce i, and Joao Jalles. 2020. “Job P o ec ion De egula ion in
Good and Bad Times.” Ox o d Economic Pape s, 72 (2): 370–90.
Du al, Romain, Da ide Fu ce i, and Jakob Mie he. 2021. “Robus Poli ical Economy
Co ela es o Majo P oduc and Labo Ma ke Re o ms in Ad anced Economies:
E idence om BAMLE o Logi Models.” Jou nal o Applied Econome ics, 36 (1):
98–124.
50
Esano , Ak am. 2006. “The G ow h-Po e y Nexus: E idence om Kazakhs an.” ADB
Ins i u e Discussion Pape No. 51. Asian De elopmen Bank Ins i u e (ADBI).
Fe nandez, And és, S ephanie Schmid -G ohe, and Ma ín U ibe. 2020. “Does he
Commodi y Supe Cycle Ma e ?” NBER Wo king Pape No. 27589. Na ional
Bu eau o Economic Resea ch (NBER)
Ga i ia, Alejand o U. 2005. La Re o ma Labo al de 2002: ¿Funcionó o no? Coyun u a
Económica, 35 (1): 73–100. Uni e sidad de los Andes.
Gelb, Alan. 1990. Oil Wind alls: Blessing o Cu se? A Compa a i e S udy o Six
De eloping Expo e s. Ox o d Uni e si y P ess, 1990.
G indle, Me ilee S., and John W. Thomas. 1991. Public Choices and Policy Change:
The Poli ical Economy o Re o m in De eloping Coun ies. Johns Hopkins
Uni e si y P ess.
Hagga d, S ephan, and Robe R. Kau man, eds. 1992. The Poli ics o Economic
Adjus men : In e na ional Cons ain s, Dis ibu i e Con lic s, and he S a e.
P ince on Uni e si y P ess, P ince on.
Hagga d, S ephan, and S e en B. Webb. 1994. Vo ing o Re o m: Democ acy, Poli ical
Libe aliza ion and Economic Adjus men . Ox o d Uni e si y P ess: New Yo k.
Jo dà, Òsca , and Alan M. Taylo . 2016. “The Time o Aus e i y: Es ima ing he A e age
T ea men E ec o Fiscal Policy.” Economic Jou nal, 126 (590): 219–55.
Jo dà, Òsca . 2005. “Es ima ion and In e ence o Impulse Responses by Local
P ojec ions.” Ame ican Economic Re iew, 95 (1): 161–82.
Ka l, Te y. 1997. The Pa adox o Plen y: Oil Booms and Pe o S a es. Uni e si y o
Cali o nia P ess.
Lae en, Luc, and Fabian Valencia. 2018. “Sys emic Banking C ises Re isi ed.” IMF
Wo king Pape No. 18/206. In e na ional Mone a y Fund (IMF).
Lancas e , Tony. 2000. “The Inciden al Pa ame e P oblem Since 1948.” Jou nal o
Econome ics, 95 (2): 391–413.
Lo a, Edua do A., and Mau icio Oli e a. 2004. “Wha Makes Re o ms Likely: Poli ical
Economy De e minan s o Re o ms in La in Ame ica.” Jou nal o Applied
Economics, 7 (1): 99–135.
Mehlum, Hal o , Ka l Moene, and Ragna To ik. 2006. “Ins i u ions and he Resou ce
Cu se.” Economic Jou nal, 116 (508): 1–20.
51
Olea, José Luis Mon iel, and Mikkel Plagbo g-Mølle . 2021. “Local P ojec ion In e ence
is Simple and Mo e Robus Than You Think.” Econome ica, 89 (2): 1789–823.
Mo ensen, Dale T., and Ch is ophe A. Pissa ides. 1994. “Job C ea ion and Job
Des uc ion in he Theo y o Unemploymen .” The Re iew o Economic S udies, 61
(3): 397–415.
Nelson, Joan M. 1990. Economic C isis and Policy Choice: The Poli ics o Adjus men in
he Thi d Wo ld. P ince on Uni e si y P ess, P ince on.
O’Kee e, John. 1997. “The Texas Banking C isis: 1980–1989.” Fede al Deposi Insu ance
Co po a ion.
Pe y, Guille mo, and Mau icio Oli e a. 2009. “El Impac o Del Pe óleo Y la Mine ía en El
Desa ollo Regional Y Local en Colombia.” CAF Documen o de abajo N° 2009/06.
Ca acas, Venezuela: CAF.
Pissa ides, Ch is ophe A. 2000. “Equilib ium Unemploymen Theo y.” MIT P ess.
Plagbo g-Mølle , Mikkel, and Ch is ian K. Wol . 2021. “Local P ojec ions and VARs
Es ima e he Same Impulse Responses.” Econome ica 89 (2): 955–80.
Pos el-Vinay, Fabien, and Jean--Ma c Robin. 2002. “Equilib ium Wage Dispe sion wi h
Wo ke and Employe He e ogenei y.” Econome ica, 70 (6): 2295–350.
Rod ik, Dani. 1996. “Unde s anding Economic Policy Re o m.” Jou nal o Economic
Li e a u e, 34 (1): 9–41.
Ross, Michael L. 2001. “Does Oil Hinde Democ acy?” Wo ld Poli ics, 53 (3): 325–61.
So bello, Paolo. 2023. “Hi ing, Fi ing, A omizing: Manpowe Agencies and P eca ious
Labo in Kazakhs an’s Oil Sec o .” In e na ional Labo and Wo king-Class 103: 8-
23. Camb idge Uni e si y P ess.
Teulings, Coen N., and Nikolay Zubano . 2014. “Is Economic Reco e y a My h? Robus
Es ima ion o Impulse Responses.” Jou nal o Applied Econome ics 29 (3): 497–
514.
Williamson, John, and S ephan Hagga d. 1994. “The Poli ical Condi ions o Economic
Re o m.” In John Williamson, ed. The Poli ical Economy o Policy Re o m. Ins i u e
o In e na ional Economics, pp. 525–96.
Woold idge, Je ey M. 2010. Econome ic Analysis o C oss Sec ion and Panel Da a (2nd
ed.). MIT P ess.
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. 752
No embe 2024
Ex e nal Shocks and Labo Ma ke Re o ms in Au oc acies and Democ acies
E idence om Oil P ice Wind alls
This pape explo es how oil p ice wind alls impac labo ma ke egula ion ac oss 83 coun ies om 1970
o 2014. Findings e eal ha wind all gains lead o de egula ion in au oc acies bu no democ acies, while
wind all losses igge de egula ion in democ acies. The pape examines how democ acies use wind all gains
o boos spending, whe eas au oc acies may ocus on en ex ac ion and inc easing economic e iciency.
De egula ion in democ acies du ing losses aligns wi h he c isis-induced e o m hypo hesis, linked o win
de ici s and heigh ened banking c isis isks.
Abou he Asian De elopmen Bank
ADB is commi ed o achie ing a p ospe ous, inclusi e, esilien , and sus ainable Asia and he Paci ic,
while sus aining i s e o s o e adica e ex eme po e y. Es ablished in 1966, i is owned by 69 membe s
—49 om he egion. I s main ins umen s o helping i s de eloping membe coun ies a e policy dialogue,
loans, equi y in es men s, gua an ees, g an s, and echnical assis ance.
EXTERNAL SHOCKS
AND LABOR MARKET
REFORMS IN AUTOCRACIES
AND DEMOCRACIES
EVIDENCE FROM OIL PRICE WINDFALLS
Ma kus B ueckne , Gab iele Ciminelli, and No man Loayza
