Consumer Information and Price Transmission: Empirical Evidence
Abstract
EconStor is a publication server for scholarly economic literature, provided as a non-commercial public service by the ZBW.
Full text
Loy, Jens‐Pe e ; Penne s o e , Die e ; R oshi, Daniela; Weiss, Ch is oph; Yon che a,
Biliana
A icle — Published Ve sion
Consume In o ma ion and P ice T ansmission: Empi ical
E idence
The Jou nal o Indus ial Economics
P o ided in Coope a ion wi h:
John Wiley & Sons
Sugges ed Ci a ion: Loy, Jens‐Pe e ; Penne s o e , Die e ; R oshi, Daniela; Weiss, Ch is oph;
Yon che a, Biliana (2022) : Consume In o ma ion and P ice T ansmission: Empi ical E idence, The
Jou nal o Indus ial Economics, ISSN 1467-6451, Wiley, Hoboken, NJ, Vol. 70, Iss. 3, pp. 631-683,
h ps://doi.o g/10.1111/joie.12300
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/266658
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 p://c ea i ecommons.o g/licenses/by/4.0/
THE JOURNAL OF INDUSTRIAL ECONOMICS 0022-1821
Volume LXX Sep embe 2022 No. 3
CONSUMER INFORMATION AND PRICE TRANSMISSION:
EMPIRICAL EVIDENCE*
JENS-PETER LOY†
DIETER PENNERSTORFER‡
DANIELA RROSHI§
CHRISTOPH WEISS¶
BILIANA YONTCHEVAΩ
We in es iga e how consume in o ma ion a ec s p ice adjus men in
he Aus ian e ail gasoline ma ke . Ou measu e o consume in o -
ma ion is ob ained om de ailed census da a on commu ing beha -
io , as commu e s can eely sample p ices on hei commu ing ou e
and a e hus be e in o med abou p ices. A h eshold e o -co ec ion
model sugges s ha p ices adjus mo e quickly i cos shocks exceed
ce ain h esholds. Pa ame ic and semi-pa ame ic eg essions show
ha a la ge sha e o in o med consume s inc eases bo h ansmission
speed and pass- h ough a e, bu has no e ec on he asymme y o cos
ansmission.
I. INTRODUCTION
We mus look a he p ice sys em as such a mechanism o
communica ing in o ma ion i we wan o unde s and i s eal
unc ion – a unc ion which, o cou se, i ul ils less pe ec ly as
p ices g ow mo e igid. (Hayek [1945], p. 526).
*We hank semina and con e ence pa icipan s a he Vienna Uni e si y o Economics and
Business, he Johannes Keple Uni e si y Linz, he Uni e si y o Ba i Aldo Mo o, he ISWG
Wo kshop ‘Empi ical Economics’ Innsb uck, he PSE Summe School IO Wo kshop, he 10 h
Geo ey J.D. Hewings Regional Economics Wo kshop Vienna, he Annual Mee ing & he 100 h
Anni e sa y o he Aus ian Economic Associa ion (NOeG), he Economic Resea ch Semina
B a isla a, and The Economy as a Spa ial Complex Sys em (ESCoS 2018) o help ul commen s.
†Au ho s’ a ilia ions: Kiel Uni e si y, Olshausens aße 40, Kiel, Ge many.
e-mail: [email protected]
‡Johannes Keple Uni e si y Linz, Al enbe ge S aße 69, Linz, Aus ia.
e-mail: die e .penne s o [email p o ec ed]
§Vienna Uni e si y o Economics and Business, Wel handelspla z 1, Vienna, Aus ia.
e-mail: daniela. [email p o ec ed].a
¶Vienna Uni e si y o Economics and Business, Wel handelspla z 1, Vienna, Aus ia.
e-mail: cw[email p o ec ed].a
ΩDüsseldo Ins i u e o Compe i ion Economics (DICE), Hein ich-Heine Uni e si y o
Düsseldo , Uni e si ä ss aße 1, Düsseldo , Ge many, and CEPR, London, UK.
e-mail: yon che [email protected]
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial
Economics and John Wiley & Sons L d.
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion
and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
631
632 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
FOLLOWING HAYEK’S EARLY OBSERVATION, economis s ega d he adjus men
o p ices as an impo an mechanism by which in o ma ion abou changes in
demand and cos s is communica ed o ma ke pa icipan s. Acco dingly, mea-
su es o he ex en and speed by which exogenous shocks a e ansmi ed in o
p ices a e equen ly used as a ya ds ick o assessing he unc ioning o ma -
ke s. The empi ical li e a u e on p ice ansmission (and cos pass- h ough) is
eno mous and co e s many di e en ma ke s and ime pe iods.1This li e a-
u e clea ly sugges s ha p ices adjus (a) slowly and (b) o en asymme ically
o exogenous shocks.
Di e en a gumen s ha e been p oposed o accoun o a slow, incom-
ple e and/o asymme ic p ice ansmission: ma ke powe (Bo ens ein
e al. [1997]; Weyl and Fabinge [2013]), menu cos s (Ball and Mankiw [1994]),
lags in adjus men o p oduc ion and in en o y managemen (Bo ens ein
e al. [1997]), habi o ma ion and consump ion ine ia (Xia and Li [2010])
and p oduc di e en ia ion (Loy and Weiss [2019]).
Recen ly, explana ions ela ed o consume sea ch beha io ha e ecei ed a
lo o o mal a en ion (see Yang and Ye [2008]; Tappa a [2009]; Lewis [2011];
Cab al and Fishman [2012]). While he echnical de ails di e , a common ea-
u e o all models is ha consume s’ sea ch beha io and i ms’ p ice se ing
a e de e mined simul aneously. Fi ms’ incen i es o adjus p ices (upwa ds
o downwa ds) o exogenous shocks a e de e mined by how well consume s
a e in o med abou p ices. A he same ime, he mo i a ion o consume s o
become in o med and lea n abou indi idual p ices depends on i ms’ p ice
se ing beha io (mo e de ails will be p o ided in he ollowing sec ion). In
hese models cos inc eases a e ansmi ed mo e quickly compa ed o cos
dec eases, because consume s sea ch mo e (and a e hus be e in o med) in
he i s case. These models he e o e p o ide a sea ch- heo e ic a ionale o
he ‘ ocke s and ea he s phenomenon.’2
Despi e he ecen wa e o heo e ical wo k on he impac o in o ma ion
and consume sea ch on p ice se ing, empi ical e idence is sca ce. The
easons o his a e wo- old: Fi s ly, consume s’ in o ma ion endowmen s o
consume s’ sea ch cos s usually canno be obse ed di ec ly and a e he e o e
di icul o quan i y. Secondly, as indica ed abo e, consume s’ sea ch beha io
is likely o be in luenced by i ms’ p icing decisions and is hus endogenous:
equen and subs an ial p ice changes educe consume s’ incen i es o sea ch
because he dep ecia ion a e o (p ice) in o ma ion is high (Ma el [1976]).
1Kouya é and on C amon-Taubadel [2016] unco e 492 ecen pape s using p ice ansmis-
sion as a sea ch e m. Excellen e iews o he oluminous empi ical li e a u e on p ice ansmis-
sion and cos pass- h ough a e p o ided in Meye and on C amon-Taubadel [2004], F ey and
Mane a [2007], Wolman [2007], Bakucs e al. [2014], Hassouneh e al. [2015] and Lloyd [2017].
The exis ing li e a u e o en uses he e ms ‘p ice ansmission’ and ‘cos pass- h ough’ in e -
changeably o cha ac e ize he impac o cos changes on e ail p oduc p ices.
2Bacon [1991] in oduces he e m ‘ ocke s and ea he s phenomenon’ o si ua ions in which
p ices espond mo e quickly o cos inc eases han o dec eases.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 633
Likewise, consume s’ gains om sea ch a e small i i ms cha ge simila
p ices and p ice dispe sion in a ma ke is low (Tappa a [2009]). This endo-
genei y o consume sea ch makes i di icul o iden i y he causal e ec
o in o ma ion on p ice se ing in gene al, and on p ice ansmission in
pa icula .
We con ibu e o his sca ce empi ical li e a u e in wo dimensions: Fi s ,
we apply a no el measu e o consume in o ma ion based on p ecise commu -
ing pa e ns ha is a guably independen o i ms’ p ice se ing beha io and
hus allows iden i ica ion o causal e ec s. We in es iga e p ice ansmission
in he e ail gasoline ma ke in Aus ia o a ime pe iod when websi es
epo ing comp ehensi e and up- o-da e in o ma ion on gasoline p ices
we e no ye a ailable. Ac ually going o a speci ic gasoline s a ion was in
ac he only way o consume s o lea n abou cu en gasoline p ices a
ha s a ion.3As poin ed ou by Ma el [1976], in o ma ion abou gasoline
p ices di e s signi ican ly be ween wo consume g oups: commu e s and
non-commu e s. Commu e s can eely sample all p ice quo es o gasoline
along hei commu ing ou e and a e he e o e ypically be e in o med
han non-commu e s. We ob ain a measu e o consume in o ma ion by
using de ailed da a on commu ing beha io om he Aus ian census o
calcula e he sha e o commu e s passing by each indi idual gas s a ion
(Penne s o e e al. [2020]). No e ha his measu e o consume in o -
ma ion is de e mined by consume s’ long- un decisions o commu e (i.e.,
whe e o li e and wo k), which is o hogonal o s a ions’ sho - un p icing
decisions.
Second, we use a lexible empi ical app oach o measu ing he deg ee
o p ice ansmission by es ima ing h eshold e o -co ec ion models
(TECM). This me hod de e mines he op imal h eshold alues endoge-
nously o each s a ion o classi y a s a ion’s p ice spell in o di e en
egimes, and es ima es sepa a e p ice adjus men pa ame e s o each egime.
This app oach akes in o accoun i ms’ ansmi ing cos changes a di -
e en speeds, depending on he size and he sign o he cos shock. We
can he e o e dis inguish be ween an asymme y in he speed o adjus -
men and an asymme y in h esholds (and hus he size o he di e en
egimes). This u ns ou o be impo an o in e p e ing ou empi i-
cal esul s in ligh o heo e ical models (see, o ins ance, Cab al and
Fishman [2012]).
3The a ailabili y o p ice compa ison websi es as well as sma phone applica ions and au o-
mobile global posi ioning sys ems, which p o ide he cu en p ice o gasoline a nea by e ail
loca ions, had a subs an ial e ec on consume sea ch cos s in he gasoline ma ke . A he same
ime, i also makes i easie o i ms o moni o each o he ’s p ices and could hus acili a e col-
lusion be ween i ms. In such a se ing, iden i ica ion o he e ec s o consume in o ma ion on
p ice dynamics is impeded. In e es ing empi ical s udies on p ice sea ch beha io on he basis o
hese echnologies include Lewis and Ma el [2011], De los San os e al. [2012] and By ne and de
Roos [2017].
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
634 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
Consis en wi h heo e ical p edic ions, we ind empi ical e idence ha
gasoline s a ions’ p ice ansmission is in luenced by consume s’ in o ma ion
endowmen s. A la ge sha e o in o med commu e s leads o a highe speed
o p ice ansmission and a highe pass- h ough a e. We do no obse e
a signi ican e ec o consume in o ma ion on he asymme y o p ice
adjus men , nei he in he speed o adjus men no in adjus men h esholds.
The emainde o his a icle is o ganized as ollows: Sec ion II b ie ly
e iews heo e ical models o consume sea ch and p ice dynamics and
discusses measu es o consume in o ma ion used in he exis ing empi ical
li e a u e. Sec ion III p esen s he da a and Sec ion IV epo s es ima ion
esul s. Sec ion Vdesc ibes esul s om al e na i e es ima ion expe imen s
and Sec ion VI concludes.
II. LITERATURE
II(i). Theo y
A numbe o heo e ical models a ibu e (asymme ic) p ice ansmission o
consume sea ch beha io . The cen al ea u e o hese models is ha p ice
igidi y o , con e sely, he speed o p ice adjus men is ela ed o consume s’
in o ma ion endowmen and sea ch in ensi y: I mo e consume s become
in o med, he in ensi y o compe i ion inc eases, p ice-cos ma gins decline,
and cos changes a e passed on o consume s mo e quickly.
Al hough he exac mechanisms di e , asymme ic p ice adjus men in
hese models is gene a ed by consume s sea ching mo e when cos s o p ices
inc ease han when cos s o p ices dec ease: In Yang and Ye [2008] consume s
do no obse e p oduc ion cos s di ec ly, bu lea n abou cos s by obse ing
i ms’ p ices. Consume s lea n abou posi i e cos shocks mo e quickly and
cos inc eases a e hus passed on as e han cos dec eases. In Tappa a [2009],
consume s sea ch mo e when cos s inc ease, and a ise in inpu p ices is he e-
o e passed on mo e quickly. In Lewis [2011] consume s sea ch mo e when
p ices (and hence cos s) a e inc easing, wi h simila consequences on he
asymme y o p ice dynamics. In Cab al and Fishman [2012] consume s lea n
abou cos shocks by obse ing p ice changes, which induces hem o sea ch
he ma ke . In o de o a oid consume sea ch, i ms e ain om passing on
small cos dec eases, leading o slowe pass- h ough o (small) cos dec eases
ela i e o cos inc eases.
To acili a e in e p e a ion o he exis ing empi ical e idence, we b ie ly
e isi Tappa a’s [2009] model, which is mos closely ela ed o ou empi i-
cal analysis.4In his model, a ini e numbe o n>1 i ms sell a homogeneous
p oduc . They ace cons an ma ginal cos s cand compe e in p ices. The e is
4A mo e de ailed e iew o Tappa a’s [2009] model is p esen ed in Online Appendix B, which
is a ailable a he Jou nal’s edi o ial web si e.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 635
a uni mass o consume s wi h uni demand o he p oduc and willingness
o pay >c.
Tappa a [2009] dis inguishes be ween di e en g oups o consume s. A
sha e 𝜆∈[0,1]o consume s obse es all p ices a no cos ; hese ‘shop-
pe s’ ob ain p ice in o ma ion unin en ionally (in he case o gasoline, o
example, commu e s obse e p ices om gasoline s a ions on hei commu e
ou e). The emaining sha e o consume s (1 −𝜆) ha e posi i e, bu he e oge-
neous sea ch cos s and will be e e ed o as ‘non-shoppe s.’ Non-shoppe s
decide whe he o no o sea ch by compa ing hei sea ch cos s wi h he
expec ed gains om sea ch E[p−pmin], wi h pmin being he lowes p ice
in he ma ke . I non-shoppe s choose o sea ch, hey obse e all p ices
in he ma ke (all-o -no hing nonsequen ial sea ch) and hus become ully
in o med. Toge he , shoppe s as well as hose non-shoppe s, who decide
o sea ch, cons i u e he g oup o in o med consume s; hei sha e among
all consume s is 𝜇. In o med consume s buy om he cheapes s o e, p o-
ided ha i s p ice does no exceed hei willingness o pay . The 1 −𝜇
uni o med consume s ( he non-shoppe s, who choose no o sea ch) buy
om a andomly selec ed s o e, as long as he p ice a his s o e does no
exceed .
In his model, consume s’ sea ch in ensi ies (consume s’ in o ma ion
endowmen s) a ec he deg ee o compe i ion among i ms: a highe sha e
o in o med consume s implies a mo e elas ic i m demand, which ansla es
in o highe cos pass- h ough a es. Tappa a [2009] shows o mally ha he
cos pass- h ough a e inc eases wi h he sha e o in o med consume s, i.e.,
𝜕2p
𝜕c𝜕𝜇 >0, whe e pdeno es he a e age ma ke p ice. Since 𝜕𝜇
𝜕𝜆 >0, his esul
also holds o he sha e o shoppe s (𝜕2p
𝜕c𝜕𝜆 >0).
In o de o explain asymme ies in p ice adjus men , Tappa a [2009]
in oduces a simple dynamic model in which consume s do no know he
ac ual p oduc ion cos . Thei sea ch decisions a e based upon pas cos
ealiza ions and he elas ici y o demand di e s be ween pe iods o ma ginal
cos inc eases o dec eases. Mo e speci ically, Tappa a [2009] assumes ha
ma ginal p oduc ion cos s can be high (cH)o low(cL), and a cos dec ease
(inc ease) is hus equi alen o cos s changing om cH o cL( om cL o cH).5
Consume s do no obse e he con empo aneous cos ealiza ion, bu o m
expec a ions abou p oduc ion cos s based on cos ealiza ions in he p e ious
pe iod. I ma ginal cos s we e high (cH) in he p e ious pe iod, consume s
s ill expec high cos s oday. In his case, p ice dispe sion is expec ed o
be low, which educes he incen i es o sea ch. Consequen ly he sha e o
5The assump ion o only wo ma ginal cos s a es (high and low) is c i icized in Lewis [2011],
since his makes i di icul o iden i y whe he po en ial e ec s o igina e o m high s. low ma gin
pe iods o om posi i e s. nega i e cos changes. Lewis [2011] de elops a heo e ical model
(and also p esen s empi ical e idence) showing ha p ices espond as e o cos changes du ing
pe iods when ma gins a e low.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
636 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
in o med consume s will be low, which educes he elas ici y o demand
aced by each i m. The ma ke ge s less compe i i e and he link be ween
cos s and p ices becomes weake . I , on he o he hand, cos s inc ease om
cL o cH, consume s (misleadingly) expec o be in a low-cos en i onmen ,
expec p ice dispe sion o be high and hus sea ch mo e in ensi ely. The
inc easing mass o in o med consume s inc eases he demand elas ici y and
makes he ma ke mo e compe i i e. Tappa a [2009] hus concludes ha
posi i e cos shocks a e passed on o p ices mo e quickly han nega i e
ones.
A di ec empi ical es o Tappa a’s [2009] model p edic ion ega ding
asymme ies in p ice adjus men is di icul , because e en i esea che s
obse e consume s’ ac ual sea ch beha io in ime pe iods o bo h cos
inc eases and dec eases, he sha e o in o med consume s will ne e heless
be endogenous.6To a oid endogenei y conce ns we use a measu e o he
(exogenous) sha e o shoppe s 𝜆in he empi ical applica ion. No e, howe e ,
ha Tappa a’s [2009] analysis does no p o ide clea p edic ions ega ding
he e ec o his consume g oup on he asymme y o p ice ansmission:
he ela ionship be ween he di e ence in he equilib ium sha e o in o med
consume s be ween a low-cos and a high-cos en i onmen , esponsible
o he asymme y in p ice ansmission, and he sha e o shoppe s 𝜆can
be ei he nega i e o e he en i e ange o 𝜆∈[0,1]o cha ac e ized by an
in e se-U-shaped ela ionship.7
Ano he implica ion o Tappa a’s [2009] model is ha he speed o cos
ansmission depends on he sign, bu no on he size o he cos shock.
In a di e en heo e ical se ing, Cab al and Fishman [2012] de elop a
sea ch- heo e ic model whe e p ices a e s icky o cos changes wi hin speci ic
anges. In his model p ice changes a e likely o induce consume sea ch,
which i ms wan o a oid, because be e in o med consume s make he
ma ke mo e compe i i e. I cos shocks a e posi i ely co ela ed ac oss
i ms (which is likely in he e ail gasoline ma ke ), he gains om adjus ing
p ices o mode a e cos dec eases a e small ela i e o he expec ed loss
due o inducing consume sea ch. P ices emain cons an i cos s dec ease
mode a ely, while la ge nega i e as well as posi i e cos shocks a e passed on
o consume s quickly, esul ing in h ee egimes wi h a highe speed o p ice
ansmission in he ou e egimes.
On he basis o his sho e iew o heo e ical models, we iden i y wo
issues ha a e pa icula ly impo an o he in e p e a ion o he empi ical
e idence. Fi s , he ex en o which consume s sea ch in o de o ob ain
6Empi ical a icles obse ing consume sea ch beha io di ec ly a he indi idual le el (De los
San os e al. [2012]) o a he ma ke le el (Lewis and Ma el [2011]; By ne and de Roos [2017])
a e indeed in e es ed in explaining consume sea ch a he han e alua ing he e ec o consume
in o ma ion on p ices.
7See Online Appendix B o a mo e de ailed discussion o his poin .
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 637
in o ma ion abou p ices is endogenous. This endogenei y o sea ch calls o
an adequa e s a egy o iden i y he causal e ec s o consume in o ma ion
on p ice ansmission. We accoun o his by p o iding a measu e o he
sha e o shoppe s 𝜆, a a iable exogenous in he heo e ical models. Second,
an adequa e measu emen o he deg ee and asymme y o p ice ansmission
needs o di e en ia e be ween di e en dimensions o p ice adjus men .
In Tappa a [2009] he asymme y in p ice adjus men s ems om di e ences
in he speed o ansmission o cos inc eases compa ed o cos dec eases. In
Cab al and Fishman [2012], on he o he hand, he asymme y comes om
he h esholds con ining he ‘inne egime’ (cha ac e ized by s icky p ices)
being asymme ic (i.e., no cen e ed a ound ze o). In ou empi ical analysis
we hus es ima e e y lexible h eshold e o -co ec ion models (TECM)
ha allow us o di e en ia e be ween asymme ies in he speed o p ice
ansmission and in h eshold le els.
II(ii). In o ma ion and P ices: E idence
An ob ious challenge in he empi ical li e a u e on he ela ionship be ween
in o ma ion and p ices is he measu emen o indi idual consume s’ in o -
ma ion endowmen s o buye s’ sea ch ac i i ies. In his seminal wo k on ‘The
Economics o In o ma ion,’ S igle [1961] a gues ha consume s will sea ch
mo e i he bene i s om sea ch inc ease and/o i sea ch cos s dec ease. In
he absence o a di ec measu e o consume s’ in o ma ion endowmen , he
p oxy a iables ypically used can be classi ied along hese lines, i.e., indica o s
ela ed o he bene i s om as well as he cos s o sea ch.
In one o he i s empi ical s udies on he impac o consume in o ma ion
and p ices in he gasoline ma ke , Ma el [1976] uses gasoline consump ion
pe ca o p oxy bene i s om sea ch. Gains will be la ge i pe capi a
consump ion is high. Median amily income and schooling a e used o mea-
su e cos s o sea ch. Ma el [1976] a gues ha an inc ease in amily income
aises oppo uni y cos s o ime (cos s o sea ch)8and ha be e educa ion
inc eases he e iciency o sea ch. So ensen [2000], in es iga ing he ma ke
o p esc ip ion d ugs, a gues ha pu chase equency is an impo an ele-
men o he sea ch decision. I p esc ip ions a e pu chased epea edly, p ice
in o ma ion ob ained om sea ching he ma ke s can be used mul iple imes
be o e his in o ma ion ‘expi es.’ These ma ke s should hus be cha ac e ized
by be e in o med consume s compa ed o ma ke s o p oduc s pu chased
less equen ly.
8Measu es o amily income a e also used as p oxies o sea ch cos s in Ba on e al. [2000], o
example. Simila ly, Chand a and Tappa a [2011], Reme [2015] and Chesnes [2016] asc ibe con-
sume s o al e na i e p oduc s o di e en ypes o s o es o di e en income le els, a guing ha
indi iduals consuming p emium gasoline (Chand a and Tappa a [2011]; Reme [2015]) o going
o b anded gas s a ions (Chesnes [2016]) ea n mo e, ha e highe sea ch cos s and he e o e sea ch
less. By es ima ing a s uc u al model, Nishida and Reme [2018] p o ide empi ical e idence ha
sea ch cos s and household income a e indeed closely ela ed.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
638 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
A no el app oach o measu ing he e ec s o sea ch cos s is adop ed by
She man and Weiss [2017]. On he basis o hand-collec ed da a om an
ou doo ma ke in Je usalem, he au ho s use c oss-sec ional and empo al
a ia ion in pedes ian conges ion as one p oxy o sea ch cos s. Indi ec
e idence on he e ec s o cos s and bene i s om sea ch o e a long ime
pe iod a e p o ided in Ecka d [2004]. Ecka d [2004] compa es p ice dis-
pe sion o he same commodi ies in 1901 and 2001 and inds ha p ice
dispe sion has inc eased o e ime, despi e he in oduc ion o signi ican
sea ch cos - educing echnologies in anspo a ion and communica ion. The
au ho a gues ha he p oduc s analyzed cons i u ed subs an ially smalle
sha es in he consume s’ budge s in 2001 compa ed o 1901. The decline in
sea ch cos s migh hus ha e been o se by a decline in he bene i s om
sea ch (associa ed wi h he educ ion in he budge sha es o hese p oduc s)
o e ime.
A guably he mos equen ly used indica o o consume in o ma ion
endowmen is consume access and use o he In e ne . Ellison and Elli-
son [2005], o ins ance, a gue ha ‘ he In e ne has p o ided esea che s
wi h he oppo uni y o s udy how ma ke s unc ion in no el and ex eme
ci cums ances. A i id example is ha wi h he g ow h o he In e ne , we
suddenly ha e ma ke s wi h essen ially no sea ch cos s’ (p. 140). The impac
o online sea ch and In e ne pu chases on p ices has been in es iga ed o
many di e en ma ke s, including au omobiles (Mo on e al. [2001]), li e
insu ance (B own and Goolsbee [2002]), books (Tang e al. [2010]), consume
elec onics p oduc s (Baye e al. [2004]), ai line icke s (O lo [2011]; Sen-
gup a and Wiggins [2014]), elec ici y (Gugle e al. [2018]) and e ail gasoline
(Luco [2019]).
While in es iga ing online ma ke s o d awing on In e ne usage o de i e
measu es o consume in o ma ion has p o ided many in e es ing insigh s,
some p oblems a e also associa ed wi h his app oach. Fi s , as Baye and
Mo gan [2001] poin ou , consume s’ decisions o use p ice compa ison
websi es a e endogenous and depend on i ms’ p icing decisions. The gains
om sea ch will be low i p ice dispe sion is low (Tappa a [2009]; Chand a
and Tappa a [2011]) o i p ice ola ili y is high (Ma el [1976]; Bo ens ein
e al. [1997]). Lewis and Ma el [2011] and By ne and de Roos [2017] p o ide
empi ical e idence ha consume s’ sea ch ac i i ies a e indeed in luenced by
i ms’ p icing decisions: Using web a ic da a om gasoline p ice epo ing
websi es, Lewis and Ma el [2011] ind ha consume s sea ch mo e when
p ices ise han when p ices all, and By ne and de Roos [2017] epo ha
consume s’ sea ch ac i i ies a e in luenced by bo h p ice dispe sion and p ice
ola ili y. Second, he a ailabili y o he In e ne o o In e ne compa ison
si es may no only p o ide a educ ion in consume s’ sea ch cos s, bu may
also ha e an i-compe i i e e ec s. I i ms can easily moni o hei i als’
ac ions, be e in o ma ion may acili a e coo dina ion in i ms’ p ice se ing
beha io . Luco [2019] p esen s a model o show ha p ice anspa ency can
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 645
We will e e o his equa ion as he s anda d e o -co ec ion model
(ECM).21 The e o -co ec ion e m ECT in he abo e equa ion ep esen s
he de ia ion o he e ail p ice om i s long- un ela ionship wi h he
c ude oil p ice o each gasoline s a ion, as desc ibed in Equa ion 1. Thus,
ECTi, ≡𝜖i, =Pi, −𝜅i−𝜌iC . The coe icien 𝛾imeasu es he speed o
adjus men o p ices owa ds he long- un equilib ium ( he a e a which he
e o s a e co ec ed) o gasoline s a ion i. Gi en co-in eg a ion, he speed
pa ame e 𝛾iis expec ed o be nega i e. In o de o es o e he equilib ium,
p ices dec ease in pe iods when hey a e abo e hei long- un ela ionship
wi h cos s, and a e expec ed o inc ease in pe iods when p ices a e below his
long- un ela ionship. The coe icien s 𝛿1 ep esen he sho - un esponses
o he e ail p ices o own changes in he pas wo days (A=2), and he
coe icien s 𝛿2measu e he sho - un esponses o he e ail p ice o changes
in he c ude oil p ice wi hin he las wo days (B=2).22 The in e cep and he
e o e m a e deno ed by 𝜏iand 𝜁i, , espec i ely.
The s anda d e o -co ec ion model (ECM) in Equa ion 2allows o di -
e en p ice dynamics ac oss s a ions. The weakness o his model, howe e , is
ha he adjus men p ocess o a pa icula gas s a ion is es ic ed o be he
same, i espec i e o (a) he sign and (b) he size o he cos shock.
Bo h, heo e ical models as well as a la ge numbe o empi ical s udies (see
Sec ion II) p o ide good easons o doub his implici symme y assump-
ion wi h espec o he sign o cos s shocks (posi i e o nega i e). Follow-
ing an ea ly empi ical analysis o p icing asymme ies in he gasoline ma ke
(Bo ens ein e al. [1997]), nume ous empi ical s udies es ima e a ia ions o
an e o co ec ion model o accoun o an asymme ic esponse in he speed
o adjus men o posi i e and nega i e cos s shocks.23
The second es ic ion o he s anda d e o -co ec ion model is ha he
adjus men pa e n is assumed independen o he size o he shock. This
implici assump ion is challenged in empi ical wo k (Godby e al. [2000])
as well as sea ch- heo e ic models (Cab al and Fishman [2012]). P e ious
empi ical con ibu ions in es iga ing asymme ic cos ansmission allow
p ices o espond di e en ly o cos inc eases e sus cos dec eases, and hus
equi e he h eshold o asymme ic esponses o be a ze o change in cos s.
21 Mo e de ails on empi ical applica ions o his model in he gasoline ma ke a e p o ided in
Ecke [2013].
22 We use he Akaike in o ma ion c i e ia (AIC) es s a is ics o selec he op imal lag o de
o each gasoline s a ion. Lue kepohl [1985] and Toda and Yamamo o [1995] show ha hese es
s a is ics will ha e he s anda d asymp o ic p ope ies e en i he a iables a e in eg a ed o o de
1, i.e., I(1). Paulsen [1984] and Nielsen [2001] also show ha he AIC can be used o bo h I(0)
and I(1) a iables.
23 Indeed, hese a icles a e pa o an e en la ge li e a u e examining (asymme ic)
pass- h ough o ups eam cos shocks in a a ie y o indus ies. F ey and Mane a [2007] and
Meye and on C amon-Taubadel [2004] p o ide excellen su eys o he b oade li e a u e wi h
a speci ic ocus on he echnical issues o modeling and es ima ion.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
646 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
Godby e al. [2000] ex end his li e a u e such ha his h eshold can ake
any alue, and hus allow small cos changes o be ansmi ed di e en ly
han la ge cos shocks. Th eshold-e ec s also play an impo an ole in ecen
sea ch- heo e ic models. Cab al and Fishman [2012], o example, sugges
ha small cos dec eases a e passed on less quickly han posi i e o la ge
nega i e cos shocks. While Godby e al. [2000] es ima e one h eshold only,
Cab al and Fishman [2012] sugges ha h esholds o p ice inc eases and
dec eases could di e . To accoun o his po en ial asymme y in h esholds,
an empi ical model wi h mo e han one h eshold is equi ed.
In he ollowing, we desc ibe and es ima e such a lexible h eshold
e o -co ec ion model, which allows o (a) an asymme ic esponse in he
speed o adjus men as well as (b) an asymme y in he le el o h esholds o
posi i e and nega i e shocks.
Tes ing o h eshold non-linea i y is a non-s anda d in e ence p ob-
lem since he nuisance pa ame e ( he h eshold) is no iden i ied unde
he null hypo hesis. Consequen ly he asymp o ic dis ibu ions o he es s
a e non-s anda d. This es p oblem is known as he Da ies P oblem
(Da ies [1987]) in non-linea ime se ies models and has been discussed
since by And ews and Plobe ge [1994] and Hansen [1996] in he con ex o
co-in eg a ion. Se e al app oaches a e a ailable o sol e his p oblem based
on he na u e o he ime se ies p ocess.24 We apply he app oach p oposed by
S ikholm and Te äs i a [2015] based on he smoo h ansi ion au o eg es-
sion o de e mine he numbe o egimes, and hus un a sequence o es s:
Fi s , a linea e o -co ec ion model is es ed agains an e o -co ec ion
model wi h wo egimes (one h eshold). Since he null hypo hesis o linea i y
is ejec ed o he majo i y o he s a ions, in he nex s ep we es he model
wi h wo egimes (one h eshold) agains a model wi h h ee egimes ( wo
h esholds). The wo- egimes model is also ejec ed, sugges ing ha he p ice
ansmission p ocess is cha ac e ized by h ee egimes ( wo h esholds).
To es ima e he h esholds we ollow he idea o Chan [1993], who showed
ha he alue o he h eshold minimizing he sum o squa ed e o s om
he long- un equilib ium is a supe -consis en es ima e. Howe e , since he e
a e wo h esholds ha need o be es ima ed, a g id sea ch p ocedu e as dis-
cussed in Ende s and Siklos [2001] is implemen ed. The es ima ed esiduals
om he long- un ela ionship o Equa ion 1a e candida es o he h esh-
olds. These esiduals a e so ed in ascending o de . To ge consis en es ima es
o he pa ame e s o he e o -co ec ion model, a su icien numbe o obse -
a ions in each egime is needed. The e o e, he smalles and he la ges 15%
o he es ima ed esiduals we e d opped and each o he emaining 70% o he
24 Tsay [1989] p oposed he use o esiduals om an a anged au o- eg ession o es o
non-linea beha io . Hansen [1996;1997] p opose an al e na i e me hod o uni a ia e p ocesses
which allows es ima ing (only) one h eshold and he model pa ame e s simul aneously. Fo he
mul i a ia e case, see Hansen and Seo [2002].
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 647
alues a e conside ed po en ial h esholds. No e ha he choice o d op he
smalles and la ges 15% is a he a bi a y, bu i is impo an o ensu e a
su icien numbe o obse a ions in each egime. Since la ge (absolu e) al-
ues o h esholds imply a small numbe o obse a ions in he ou e egimes,
we ollowed Ende s and Siklos [2001] and d opped 15%. Based on a ule o
humb o he minimum numbe o obse a ions in each egime p o ided by
Hansen [1999], we sea ch om he 20% o 45% o he emaining a anged
esidual alues o ind he lowe h eshold and om he 55% o 80% o he
a anged esidual alues o ind he uppe h eshold. The es ima ed h esh-
olds ha minimize he esidual sum o squa es in each egime a e he inal
h eshold alues ha we use o es ima e he TECM.
Based on he es esul s and he es ima ed h esholds, he ollowing h esh-
old e o -co ec ion model wi h h ee egimes is es ima ed o each gasoline
s a ion by means o o dina y leas -squa es (OLS):
(3)
ΔPi, =
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
𝜏+
i+𝛾+
iECTi, −1+∑A
a=1𝛿+
1,i,aΔPi, −a+∑B
b=0𝛿+
2,i,bΔCi, −b+𝜂+
i, ,
i ECT −1>𝜃
+
𝜏0
i+𝛾0
iECTi, −1+∑A
a=1𝛿0
1,i,aΔPi, −a+∑B
b=0𝛿0
2,i,bΔCi, −b+𝜂0
i, ,
i 𝜃+≥ECT −1≥𝜃−
𝜏−
i+𝛾−
iECTi, −1+∑A
a=1𝛿−
1,i,aΔPi, −a+∑B
b=0𝛿−
2,i,bΔCi, −b+𝜂−
i, ,
i 𝜃−>ECT −1
The pa ame e s 𝜃+and 𝜃−indica e he uppe and lowe h eshold alues
used o iden i y he h ee egimes. No e ha he e o co ec ion e m is pos-
i i e (nega i e)—and s a ions a e hus in he uppe (lowe ) egime—i hey
a e exposed o a la ge enough nega i e (posi i e) cos shock. The pa ame e s
𝛾+,𝛾0and 𝛾− e e o he speed o adjus men in he uppe , middle and lowe
egimes, while 𝜏+
i,𝜏0
iand 𝜏−
iindica e he espec i e cons an s. The pa ame e s
𝛿1(𝛿2) deno e he sho - un adjus men a es o changes in own p ices (cos s)
and a e es ima ed up o a lag o de o wo. The espec i e e o e ms a e
indica ed by 𝜂+,𝜂0and 𝜂−. Cos inc eases will be passed on mo e quickly o
p ices han cos dec eases i (a) he speed o p ice adjus men o cos inc eases
exceeds he speed o adjus men o cos dec eases (i.e., |𝛾−
i|>|𝛾+
i|) and/o (b)
he adjus men h eshold o p ice inc eases is smalle han o p ice dec eases
(i.e., |
𝜃−|<|
𝜃+|).
While he TECM is e y lexible, es ima ing a la ge numbe o coe icien s
o each gasoline s a ion migh lead o imp ecisely es ima ed pa ame e s.
We p o ide simple , bu less lexible models in he sensi i i y analysis, whe e
we allow o a smalle numbe o egimes only, o es ic he pa ame e s
o be he same ac oss s a ions. This enables us o es ima e he speed and
he pass- h ough a e mo e p ecisely, bu limi s he scope o in es iga ing
asymme ies in cos ansmission.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
648 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
TABLE III
DESCRIPTIVE STATISTICS OF ESTIMATED PARAMETERS
Mean S d. De . P5 P 95
Es ima ed Pa ame e s om TECM
Pass- h ough a e (𝜌)1.294 0.213 0.940 1.701
Posi i e h eshold (𝜃+)1.389 0.825 0.365 2.770
Nega i e h eshold (𝜃−)−1.443 0.610 −2.517 −0.667
Speed uppe egime (𝛾+)−0.050 0.034 −0.109 −0.006
Speed middle egime (𝛾0)−0.034 0.225 −0.421 0.266
Speed lowe egime (𝛾−)−0.090 0.047 −0.171 0.005
Speed and Asymme y in Cos T ansmission
A e age speed ou e egimes (|𝛾+|+|𝛾−|
2)0.071 0.029 0.028 0.132
Asymme y in speed (|𝛾−|−|𝛾+|)0.040 0.055 −0.466 0.131
Asymme y in h esholds (|𝜃+|−|𝜃−|)−0.053 0.869 −1.445 1.575
No es: The numbe o obse a ions is 281. P5 and P 95 deno e he 5 h and he 95 h pe cen ile, espec i ely.
IV(ii). Es ima ion Resul s
Equa ions 1and 3a e es ima ed by OLS o each indi idual gasoline s a ion.
Summa y s a is ics on he mos impo an pa ame e es ima es a e p o ided
in Table III. The a e age pa ame e alue o he long- un pass- h ough a e,
ob ained om es ima ing Equa ion 1,is 𝜌 =1.294.25 A e aging pa ame e
es ima es o e all gasoline s a ions, we ind ha he lowe h eshold alues
(
𝜃−=−1.443) a e nea ly iden ical o he uppe ones (
𝜃+=+1.389) in abso-
lu e e ms. The a e age speed o p ice inc eases (i.e., he pa ame e alue in he
lowe egime |𝛾−|=0.090) exceeds he speed o p ice dec eases ( he pa ame e
alue es ima ed o he uppe egime |𝛾+|=0.050). The es ima ed adjus men
speed o he middle egime (|𝛾0|=0.034) is smalle in absolu e e ms han he
co esponding pa ame e s o he ou e egimes, consis en wi h Cab al and
Fishman [2012].26 No e ha small pa ame e alues o he middle egime
could be an indica o o s icky p ices ( a he han he consequence o egula ,
bu small p ice changes).
Based on he pa ame e es ima es om he TECM we calcula e h ee a i-
ables measu ing he speed and he asymme y o cos pass- h ough: (i) The
speed o p ice ansmission is de ined as he a e age o he es ima ed speed
pa ame e s in he ou e egimes (|𝛾+|+|𝛾−|
2); (ii) The asymme y in he speed o
p ice ansmission is de ined by he di e ence be ween he lowe and he uppe
speed pa ame e s (|𝛾−|−|𝛾+|). (iii) Finally, he asymme y in he h esholds
is de ined as |
𝜃+|−|
𝜃−|.
25 No e ha we use g oss e ail p ices, so ully ansmi ed cos shocks lead o a pass- h ough
a e 𝜌o 1.2, due o VAT o 20%.
26 Cab al and Fishman [2012] p edic cos ansmission o be slowes o small cos dec eases,
sugges ing he middle egime o s a a |𝜃+|>|𝜃−|=0. Ou empi ical esul s sugges ha dly any
di e ence be ween he h esholds |𝜃+|and |𝜃−|and a lowe h eshold ha is subs an ially smalle
han ze o.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 649
TABLE IV
TEST RESULTS FOR THE SIGNIFICANCE OF PRICE ADJUSTMENT
Hypo hesis # o obse a ions Pe cen
Speed Es ima e in he Uppe Regime
Rejec H0:𝛾+=0 a 1% signi icance le el 107 38
Rejec H0:𝛾+=0 a 5% signi icance le el 158 56
Rejec H0:𝛾+=0 a 10% signi icance le el 191 68
Speed Es ima e in he Middle Regime
Rejec H0:𝛾0=0 a 1% signi icance le el 73 26
Rejec H0:𝛾0=0 a 5% signi icance le el 118 42
Rejec H0:𝛾0=0 a 10% signi icance le el 143 51
Speed Es ima e in he Lowe Regime
Rejec H0:𝛾−=0 a 1% signi icance le el 218 78
Rejec H0:𝛾−=0 a 5% signi icance le el 235 84
Rejec H0:𝛾−=0 a 10% signi icance le el 241 86
No es: The numbe o obse a ions is 281.
The a e age speed o adjus men in he ou e egimes is (|𝛾+|+|𝛾−|
2)=
0.071, which co esponds well wi h es ima es ob ained om a s anda d
e o -co ec ion model in Equa ion 2.27 Rega ding he asymme y in p ice
adjus men , Table III p o ides some suppo o he ‘ ocke s and ea he s phe-
nomenon’ in he speed o adjus men . The calcula ed absolu e di e ence in
he speed o p ice adjus men in he wo ou e egimes is |𝛾−|−|𝛾+|=0.040,
which indica es ha p ices adjus mo e quickly upwa ds han downwa ds
in esponse o cos shocks. A e aging o e all gasoline s a ions, Table III
sugges s only a small asymme y in he adjus men h esholds on a e age
(|
𝜃+|−|
𝜃−|=−0.053).
No e, howe e , ha he pa ame e es ima es o he indi idual gasoline s a-
ions a y conside ably (see Tables IV and V). Table IV sugges s ha he null
hypo hesis o no p ice adjus men (𝛾0=0) is ejec ed o abou hal he gaso-
line s a ions (51%) in he inne egime a he 10% signi icance le el. The same
null hypo hesis is ejec ed o 68% in he uppe egime ( o cos dec eases) and
o 86% o all s a ions in he lowe egime ( o cos inc eases).
Table V ocuses on he di e ence be ween pa ame e es ima es ob ained
om he di e en egimes o each gasoline s a ion. The null hypo hesis o no
di e ence in he speed o p ice adjus men be ween he middle egime and he
uppe egime (be ween he middle egime and he lowe egime) is ejec ed a
he 10% signi icance le el o 75% (69%) o all gasoline s a ions. A s a is ical
es o an asymme y in he speed o p ice adjus men be ween he uppe
and lowe egime ejec s he null hypo hesis (o no signi ican di e ence) o
43% o all gasoline s a ions, despi e he subs an ial di e ence be ween he
es ima ed pa ame e s 𝛾−and 𝛾+on a e age (p obably due o ela i ely la ge
27 The es ima ed speed pa ame e s 𝛾, ob ained om a s anda d e o -co ec ion model, ake a
mean o 0.064 wi h a s anda d de ia ion o 0.029.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
650 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
TABLE V
TEST RESULTS FOR ASYMMETRIC PRICE ADJUSTMENT
Hypo hesis # o obse a ions Pe cen
H0:𝛾+=𝛾0
Rejec H0a 1% signi ican le el 91 11
Rejec H0a 5% signi icance le el 172 61
Rejec H0a 10% signi icance le el 211 75
H0:𝛾−=𝛾0
Rejec H0a 1% signi icance le el 102 36
Rejec H0a 5% signi icance le el 158 56
Rejec H0a 10% signi icance le el 194 69
H0:𝛾−=𝛾+
Rejec H0a 1% signi icance le el 33 12
Rejec H0a 5% signi icance le el 79 28
Rejec H0a 10% signi icance le el 121 43
No es: The numbe o obse a ions is 281.
s anda d e o s). The ollowing sec ion aims a in es iga ing hese di e ences
be ween s a ions mo e sys ema ically.
IV(iii). Cos T ansmission and Consume Sea ch: C oss-Sec ion Analysis
In he second s age we es ima e c oss-sec ional eg essions, ela ing a iables
indica ing he speed and he asymme y o p ice ansmission o he con-
sume s’ in o ma ion endowmen s:
(4)
Yi=𝛼0+𝛼1𝜆i+Xi𝛼2+𝜉i
The a iable 𝜆ideno es he sha e o shoppe s o each gasoline s a ion. The
ec o Xiincludes measu es o local compe i ion and gasoline s a ion cha ac-
e is ics. 𝛼0,𝛼1and 𝛼2a e pa ame e s o be es ima ed and 𝜉ideno es he e o
e m.
The dependen a iable
Yi ep esen s ou es ima es om he i s s age, indi-
ca ing he speed o he asymme y o cos pass- h ough. In one a ian we use
ou combina ions o speci ic pa ame e es ima es o he i s s age, namely
he es ima ed speed o p ice ansmission (i.e., he a e age speed in he ou e
egimes, |𝛾+|+|𝛾−|
2), he es ima ed long- un pass- h ough a e (𝜌), he es ima ed
pa ame e s on he asymme y in speed (|𝛾−|−|𝛾+|) and he asymme y in
h esholds (|𝜃+|−|𝜃−|). Al e na i ely, we use cumula i e p ice esponse al-
ues (CRV’s) as he dependen a iable
Yiin he c oss-sec ional analysis.
The ou speci ica ions o eg ession (Equa ion 4) based on speci ic
i s -s age pa ame e s a e es ima ed using di e en es ima ion echniques.
Fi s , we es ima e he eg essions by OLS. In a second app oach, we ollow
he sugges ion o Lewis and Linze [2005] and p o ide weigh ed leas squa e
(WLS) es ima es o hese models, by weigh ing he obse a ions by he in e se
o he s anda d e o s o he dependen a iable es ima es. This accoun s o
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 651
he ac ha he dependen a iables in Equa ion 4a e es ima ed pa ame e s
a he han obse ed a iables and ensu es ha Equa ion 4is es ima ed
e icien ly. Thi d, we es ima e he eg ession semi-pa ame ically in o de o
a oid pa ame ic es ic ions o a linea unc ion o he ela ionship be ween
ou measu es o cos ansmission and he sha e o shoppe s 𝜆i. The modi ied
equa ion o he semi-pa ame ic c oss-sec ion model is:
(5)
Yi=𝛼0+ (𝜆i)+Xi𝛼2+𝜈i,
We use he wo-s ep p ocedu e ou lined in Robinson [1988]. We i s
ob ain non-pa ame ic es ima es o E(
Yi|𝜆i)and E(Xi|𝜆i)and hen
eg ess
Yi−E(
Yi|𝜆i)on Xi−E(Xi|𝜆) o ob ain a consis en es ima e o
𝛼2. Finally, we eg ess
Yi−Xi𝛼2agains 𝜆inon-pa ame ically o ob ain ou
es ima e
(.).
Res ic ing he a en ion o speci ic pa ame e s o he i s -s age eg essions
as indica o s o he speed o he asymme y o cos ansmission has wo
po en ial weaknesses: Fi s , he e ec s o he sho - un changes in p ices and
cos s on e ail p ices, cap u ed by he pa ame e s 𝛿1and 𝛿2, a e no aken in o
accoun . Second, he adjus men o p ices o exogenous cos shocks is a com-
bina ion o he di e en pa ame e s es ima ed by Equa ions 1and 3.E eni
we (do no ) obse e a signi ican impac o ou measu e o consume in o -
ma ion on one o hese pa ame e s, he agg ega e e ec o in o ma ion (when
aking all pa ame e s in o accoun simul aneously) migh be di e en . To
accommoda e hese po en ial weaknesses, we ollow Bo ens ein e al. [1997]
and Lewis [2011] and es ima e cumula i e esponse unc ions. These cumula-
i e esponse alues (CRV’s) measu e he adjus men o e ail p ices o a one
uni change in c ude oil p ices, and a e used as measu es o cos ansmission
Yiwhen es ima ing Equa ion 4.
Following he discussion in Sec ion II, we expec ha gasoline s a ions wi h
la ge sha es o shoppe s exhibi highe pass- h ough a es and highe speed
o p ice ansmission han gasoline s a ions wi h low sha es o shoppe s. The
e ec o he sha e o shoppe s on ou measu es o asymme y is heo e ically
ambiguous.
IV(iii)(a). Pa ame ic E idence
The pa ame ic esul s on he e ec s o consume s’ in o ma ion endowmen s
on he speed, he long- un pass- h ough a e and he asymme y o cos ans-
mission a e epo ed in Table VI. The i s ou columns epo OLS and
WLS eg ession esul s o he speed o p ice ansmission (columns 1 and
2) and he pass- h ough a e (columns 3 and 4). The es ima es show ha a
la ge sha e o in o med consume s is associa ed wi h a highe speed o p ice
ansmission and a highe pass- h ough a e. The pa ame e es ima es a e
signi ican ly di e en om ze o a he 1% signi icance le el o he speed o
adjus men and a he 5% le el o he pass- h ough a e. The esul s epo ed
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
652 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
TABLE VI
ESTIMATION RESULTS FOR SPEED,PASS-THROUGH RATE AND ASYMMETRY OF PRICE TRANSMISSION
Speed Pass- h ough
a e Asymme y
in speed Asymme y
in h eshold
OLS WLS OLS WLS OLS WLS OLS WLS
[1] [2] [3] [4] [5] [6] [7] [8]
Sha e o shoppe s (𝜆) 0.031*** 0.031*** 0.218** 0.223** −0.003 −0.001 0.589 0.202
(0.011) (0.011) (0.108) (0.094) (0.025) (0.025) (0.411) (0.421)
# o s a ions wi hin 2 km 0.179*** 0.180*** 0.135 0.164 −0.053 −0.015 0.921 1.560*
(in 100) (0.029) (0.023) (0.158) (0.181) (0.058) (0.050) (0.714) (0.915)
Dis ance o nea es i al s a ion −0.145*** −0.142** −0.922** −0.878* −0.351*** −0.338*** 1.457 1.186
(in 100 km) (0.048) (0.059) (0.439) (0.499) (0.117) (0.130) (1.884) (2.368)
S a ion is loca ed on highway −0.028*** −0.027*** −0.014 −0.002 −0.066*** −0.066*** 0.261 0.162
(0.004) (0.006) (0.031) (0.048) (0.011) (0.012) (0.159) (0.226)
Cons an 0.046*** 0.043*** 1.184*** 1.173*** 0.056*** 0.050*** −0.477* −0.223
(0.008) (0.007) (0.071) (0.062) (0.017) (0.016) (0.263) (0.277)
# o obse a ions 281 281 281 281 281 281 281 281
R20.327 0.344 0.036 0.042 0.118 0.144 0.015 0.012
No es: S anda d e o s a e epo ed in pa en heses. In OLS speci ica ions s anda d e o s a e based on a obus a iance es ima o . In WLS speci ica ions each obse a ion is
weigh ed by he in e se o he s anda d e o s o he dependen a iable es ima es. *** signi ican a 1%, ** signi ican a 5%, * signi ican a 10% le el.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 653
in Table VI sugges ha he deg ee o asymme y in cos ansmission is no
ela ed o consume s’ in o ma ion endowmen . Pa ame e es ima es o he
sha e o shoppe s a e no signi ican ly di e en om ze o in models o he
asymme y in speed (columns 5 and 6) as well as o he asymme y in h esh-
olds (columns 7 and 8). Ou esul s hus do no p o ide empi ical suppo
o heo e ical models sugges ing an impac o consume in o ma ion on he
asymme y o h esholds (such as Cab al and Fishman [2012], o example).28
The es ima ed e ec s o consume in o ma ion on p ice ansmission a e
no only s a is ically obus , bu also sizable: I he sha e o shoppe s inc eases
om ze o (all consume s a e ex-an e unin o med) o one (pe ec in o ma-
ion), he speed o p ice ansmission inc eases by abou 0.44 s anda d
de ia ions o he espec i e endogenous a iable. Fu he mo e, he inc ease
in consume in o ma ion o he same amoun causes he pass- h ough a e
o inc ease by 1.04 s anda d de ia ions ( e e ing o he esul s o he WLS
eg essions).
Wi h ega ds o he con ol a iables, Table VI sugges s ha he speed o
cos ansmission as well as he pass- h ough a e inc ease wi h he deg ee o
compe i ion (i.e., inc eases wi h he numbe o compe i o s wi hin he local
ma ke and dec eases wi h he dis ance o he nea es i al), al hough no all
pa ame e es ima es a e signi ican ly di e en om ze o. Rega ding he asym-
me y in speed, we ind ha compe i ion is associa ed wi h mo e pass- h ough
asymme y. Acco ding o columns 5 and 6, an inc ease in he dis ance o he
nea es i al (i.e., a educ ion in he in ensi y o compe i ion) signi ican ly
educes asymme y in he speed o adjus men ( he pa ame e es ima es a e
signi ican ly di e en om ze o a he 1% signi icance le el). The eg ession
esul s on he asymme y in he h esholds a e less clea . None o he pa am-
e e es ima es in column 7 is signi ican ly di e en om ze o and he posi i e
e ec o he numbe o i al s a ions on h eshold asymme y epo ed in
column8 issigni ican lydi e en omze oa he 10% le elonly. Acco ding o
Table VI, gasoline s a ions loca ed on he highway adjus p ices mo e slowly
(columns 1 and 2) bu ha e simila long- un pass- h ough a es (columns 3
and 4). Rega ding di e ences in p icing asymme ies be ween highway s a-
ions and o he gasoline s a ions, ou esul s a e ambiguous. Focusing on he
speed o adjus men , gasoline s a ions on he highway end o adjus p ices
mo e symme ically (pa ame e es ima es a e signi ican ly di e en om ze o
28 No e ha ou empi ical model is no a o mal es o Cab al and Fishman’s model. The
au ho s do no make speci ic p edic ions abou he sha e o shoppe s, because all consume s ha e
homogeneous (and s ic ly posi i e) sea ch cos s. In hei model, e ail p ice ola ili y induces
consume s o sea ch he ma ke and he e o e makes he ma ke mo e compe i i e, which implies
lowe p o i s. I he numbe o consume s wi h posi i e sea ch cos s declines (because he sha e
o shoppe s inc eases), discou aging non-shoppe s om sea ching becomes less impo an . Con-
sequen ly, he ange o cos changes o which p ices a e s icky should sh ink. The asymme y
o h esholds is he e o e expec ed o decline wi h be e in o med consume s, a esul we do no
ind in ou da a.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
654 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
a he 1% signi icance le el in columns 5 and 6). We do no obse e a sig-
ni ican e ec o his dummy a iable on p icing asymme y wi h espec o
adjus men h esholds.
IV(iii)(b). Semi-Pa ame ic E idence
In his sec ion we show ha ou esul s on he ela ionship be ween in o -
ma ion and p ice ansmission a e no d i en by he pa ame ic es ic ions
o a linea unc ion. The esul s ob ained o he non-pa ame ic compo-
nen s o Equa ion 5a e illus a ed in Figu e 1. The igu es a e based on a
ke nel-weigh ed local polynomial eg ession.29 The g aphs indica e a posi i e
ela ionship be ween he sha e o shoppe s and bo h he speed o cos ans-
mission (Figu e 1(a)) and he pass- h ough a e (b). The e ec s o consume s’
in o ma ion endowmen s on he asymme y in speed (c) is U-shaped, while
he asymme y in h esholds (d) does no sugges a clea ela ionship wi h
he sha e o in o med consume s. Compa ing panels (a) and (b) wi h (c) and
(d), we obse e ha he con idence bands o he speed and pass- h ough a e
a e a he small ela i e o he s eepness o he cu es, sugges ing ha his
ela ionship is s a is ically mo e obus han he e ec o in o ma ion on he
asymme y o cos ansmission (in panels (c) and (d)).
A es p o ided by Ha dle and Mammen [1993], epo ed in Table A.I
in Appendix A, shows ha he non-pa ame ic ela ionships be ween he
sha e o shoppe s and he a iables on he speed and he asymme y o cos
ansmission p o ide a be e model i compa ed o a simple cons an o all
endogenous a iables, excep he asymme y in speed (c). The es s a is ics
con i m ha he model i o a non-pa ame ic speci ica ion is no be e
han a linea ela ionship o he pass- h ough a e (b), while he ela ionship
be ween he sha e o shoppe s and he speed (a) and he asymme y in
h esholds (d) is nonlinea . Gene ally, hese esul s con i m he indings o
he pa ame ic speci ica ions epo ed and discussed abo e: cos changes a e
passed on mo e quickly and ully i consume s a e be e in o med, while he
in luence o consume in o ma ion on he asymme y o p ice ansmission
is less clea .
IV(iii)(c). In es iga ing Cumula i e Response Func ions
The p e ious c oss-sec ional analysis in es iga es indi idual ansmission
pa ame e s (adjus men speed, pass- h ough a e and asymme y in speed
and h esholds) ob ained om Equa ions 1and 3sepa a ely, while we ocus
on cumula i e p ice esponse alues (CRV’s) as measu es o cos ansmis-
sion he e. Mo e speci ically, we calcula e he alues o cumula i e esponse
29 The pa ame ic esul s on he con ol a iables a e simila o he non-pa ame ic eg essions
epo ed in Table VI. The esul s a e epo ed in Table A.I in Appendix A.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 661
TABLE VIII
ESTIMATION RESULTS FOR SPEED,PASS-THROUGH RATE AND ASYMMETRY OF PRICE TRANSMISSION CONTROLLING FOR COMMUTER
HETEROGENEITY
Speed Pass- h ough
a e Asymme y
in speed Asymme y
in h eshold
OLS WLS OLS WLS OLS WLS OLS WLS
[1] [2] [3] [4] [5] [6] [7] [8]
Sha e o shoppe s (𝜆) 0.032*** 0.032*** 0.224** 0.228** −0.004 −0.001 0.671 0.271
(0.011) (0.011) (0.109) (0.094) (0.025) (0.024) (0.410) (0.424)
A e age commu ing dis ance −0.000 0.000 0.005** 0.005** 0.002*** 0.002*** −0.005 −0.001
(condi ional on # o s a ions) (0.000) (0.000) (0.003) (0.002) (0.001) (0.001) (0.010) (0.010)
# o s a ions wi hin 2km 0.179*** 0.181*** 0.204 0.225 −0.030 0.005 0.913 1.608*
(in 100) (0.029) (0.023) (0.163) (0.182) (0.057) (0.050) (0.731) (0.917)
Dis ance o nea es i al s a ion −0.141*** −0.143** −1.164** −1.129** −0.440*** −0.430*** 1.775 1.338
(in 100 km) (0.049) (0.061) (0.468) (0.510) (0.129) (0.131) (1.947) (2.450)
S a ion is loca ed on highway −0.028*** −0.027*** −0.006 0.004 −0.062*** −0.063*** 0.243 0.152
(0.004) (0.006) (0.030) (0.048) (0.012) (0.012) (0.159) (0.226)
Cons an 0.045*** 0.043*** 1.181*** 1.172*** 0.056*** 0.050*** −0.521** −0.262
(0.008) (0.007) (0.071) (0.062) (0.017) (0.016) (0.262) (0.279)
# o obse a ions 280 280 280 280 280 280 280 280
R20.328 0.345 0.052 0.059 0.148 0.176 0.017 0.013
No es: S anda d e o s a e epo ed in pa en heses. In OLS speci ica ions s anda d e o s a e based on a obus a iance es ima o . In WLS speci ica ions each obse a ion is
weigh ed by he in e se o he s anda d e o s o he dependen a iable es ima es. *** signi ican a 1%, ** signi ican a 5%, * signi ican a 10% le el.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
662 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
addi ional a iable in all model speci ica ions. These esul s sugges ha ans-
po a ion cos s seem o play some ole in explaining p ice ansmission in
addi ion o he in o ma ion e ec . We ha e o acknowledge ha his p oce-
du e is a om ideal in sepa a ing he in o ma ion om he anspo a ion
cos e ec , because commu e s a e he e ogeneous in many dimensions. Com-
mu e s a elling long dis ances, o example, a e no only cha ac e ized by
low anspo a ion cos s, bu also by high demand. A en a i e in e p e a ion
o he esul s, howe e , poin s in he di ec ion ha consume s’ in o ma ion
endowmen s a e indeed impo an in i ms’ decisions on how o pass h ough
cos shocks o consume s.
V(ii). Resul s Using Less Flexible Model Speci ica ions
Compa ed o p e ious empi ical s udies, we use a TECM speci ica ion in he
i s s age o he p esen analysis, because his app oach is e y lexible: he
model es ima es a la ge numbe o pa ame e s sepa a ely o each s a ion,
allows o a la ge deg ee o he e ogenei y ac oss i ms, and hus enables us
o in es iga e p edic ions on he speed and he asymme y o cos ansmis-
sion de i ed om a a ie y o heo e ical pape s. The downside o his lexible
app oach is ha some pa ame e s migh be imp ecisely es ima ed. Table III
shows subs an ial di e ences in pa ame e es ima es ac oss s a ions in he i s
s age, and Table IV indeed sugges s la ge s anda d e o s o many o hem.
Fo example, he speed pa ame e s 𝛾 epo ed in Table IV a e no signi ican ly
di e en om ze o a he 10% signi icance le el o 32% (14%) o all s a ions
in he uppe (lowe ) egime. When we es ima e a s anda d ECM (Equa ion 2)
ins ead o he lexible TECM (Equa ion 3), he speed pa ame e s a e signi i-
can ly di e en om ze o o all 281 s a ions in ou sample, because s anda d
e o s o he pa ame e es ima es a e subs an ially smalle . In he ollowing
sensi i i y analysis, we es ima e eg ession models based on simple and less
lexible speci ica ions o he i s s ep o he analysis o ob ain pa ame e es i-
ma es which a e mo e p ecisely es ima ed.
We i s e alua e he speed o p ice adjus men and es ima e an asymme -
ic ECM wi h one h eshold exogenously se o ze o.34 The es ima ed speed
pa ame e s 𝛾+and 𝛾−allow us o in es iga e he a e age and he asymme y
o he adjus men speed. C oss-sec ion esul s using hese pa ame e es ima es
a e epo edinTableIX. Consis en wi hou p e ious esul s,columns 1and 2
o Table IX show ha he a e age speed o p ice adjus men inc eases wi h
he sha e o shoppe s 𝜆. Again, we do no obse e a signi ican ela ionship
be ween he sha e o shoppe s and he asymme y in he speed o p ice adjus -
men (see columns 3 and 4). In his less lexible model, any asymme ic p ice
adjus men is cap u ed by asymme ies in he adjus men speed since (asym-
me ic) h eshold e ec s a e no included in his speci ica ion. I we u he
34 This app oach is used by Galeo i e al. [2003] and G asso and Mane a [2007], o example.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 663
TABLE IX
ESTIMATION RESULTS FOR SPEED AND ASYMMETRY OF PRICE TRANSMISSION BASED ON SIMPLER ERROR CORRECTION MODELS
Speed
asymme ic ECM
(one h eshold: 𝜃=0)
Asymme y in speed
asymme ic ECM
(one h eshold: 𝜃=0)
Speed
s anda d ECM
(no h eshold)
OLS WLS OLS WLS OLS WLS
Model used in ime-se ies analysis [1] [2] [3] [4] [5] [6]
Sha e o shoppe s (𝜆) 0.027** 0.025* 0.005 0.006 0.038*** 0.037***
(0.013) (0.013) (0.025) (0.025) (0.011) (0.011)
# o s a ions wi hin 2 km 0.172*** 0.163*** −0.074 −0.013 0.181*** 0.180***
(in 100) (0.038) (0.028) (0.054) (0.051) (0.026) (0.023)
Dis ance o nea es i al s a ion −0.080 −0.094 −0.274** −0.261** −0.121** −0.104*
(in 100 km) (0.098) (0.069) (0.113) (0.128) (0.049) (0.054)
S a ion is loca ed on highway −0.020*** −0.018*** −0.061*** −0.060*** −0.026*** −0.025***
(0.004) (0.006) (0.009) (0.012) (0.005) (0.005)
Cons an 0.065*** 0.064*** 0.057*** 0.049*** −0.034*** −0.030***
(0.009) (0.009) (0.018) (0.016) (0.007) (0.007)
# o obse a ions 281 281 281 281 281 281
R20.193 0.201 0.092 0.121 0.313 0.345
No es: Reg ession esul s in columns 1 o 4 a e based on an asymme ic ECM wi h one h eshold exogenously se o ze o. Resul s in columns 5 and 6 a e based on a s anda d ECM.
S anda d e o s a e epo ed in pa en heses. In OLS speci ica ions s anda d e o s a e based on a obus a iance es ima o . In WLS speci ica ions each obse a ion is weigh ed
by he in e se o he s anda d e o s o he dependen a iable es ima es. *** signi ican a 1%, ** signi ican a 5%, * signi ican a 10% le el.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
664 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
simpli y he analysis and es ima e a s anda d ECM (wi hou any h esholds),
we ob ain only one speed pa ame e 𝛾 o each s a ion. Using hese pa ame-
e es ima es in a c oss-sec ion analysis, we again conclude ha he adjus men
speed depends posi i ely on he sha e o in o med consume s, as epo ed in
columns 5 and 6 o Table IX. No e ha he in luence o he sha e o shop-
pe s 𝜆on he speed o p ice ansmission is i ually iden ical o he esul s
ob ained om he lexible TECM, epo ed in Table VI.
Fu he mo e, we e alua e he long- un ela ionship be ween e ail p ices
andcos s (seeEqua ion1) oge he wi h he s anda dECM (Equa ion 2)using
a one-s ep app oach. We s a by applying a pooled mean-g oup (PMG) es i-
ma o , ou linedin Pesa an e al. [1999], which allows he sho - un coe icien s
(in pa icula he speed pa ame e s in he s anda d ECM) o a y ac oss s a-
ions, bu cons ains he long- unpass- h ough a e 𝜌 obe he same o alls a-
ions.35 The esul sob ained om his model, epo edin column1 o TableX,
show ha he pa ame e es ima es o he pass- h ough a e 𝜌( o C −1) and he
ansmission speed 𝛾( o ECT −1) a e simila o he a e age o he pa ame-
e alues ob ained when es ima ing he ime se ies sepa a ely o each s a ion
(as epo ed in Table III). Including an in e ac ion e m be ween cos s C −1
and he sha e o shoppe s 𝜆in column 2 shows ha he in e ac ion e m is
posi i e and signi ican ly di e en om ze o a he 1% le el. This sugges s
ha he pass- h ough a e is highe i consume s a e be e in o med, which
suppo s he esul s o ou main speci ica ion. To u he simpli y he speci-
ica ion, we ollow Del as [2008] and es ic he pa ame e s in he s anda d
ECM o be iden ical o all s a ions, and allow o di e en in e cep s only.
Wi h his dynamic ixed-e ec s es ima o , he long- un ela ionship be ween
cos s and p ices (Equa ion 1) and he s anda d ECM (Equa ion 2)a ees i-
ma ed simul aneously in one panel model.36 The esul s, epo ed in he inal
column (3) o Table X, again sugges a signi ican ly highe pass- h ough a e
𝜌 o s a ions acing be e in o med consume s.
V(iii). Addi ional Sensi i i y Analyses
In o de o con i m ha ou esul s a e no d i en by he pa icula model
speci ica ions o by a small sub-sample in ou da a, we p o ide a numbe o
obus ness exe cises. The esul s con i m he main indings epo ed abo e.
35 We employ he p ocedu e discussed by Blackbu ne and F ank [2007], who implemen he
PMG es ima o in S a a. We also pe o m speci ica ion es s o panel da a be o e es ima ion:
Tes ing he e o e ms o he dynamic panel models (pmg and d e) o weak c oss-sec ional
dependence, as p oposed by Pesa an [2015], indica es no c oss-sec ional dependence in he da a.
The coin eg a ion es s o Kao [1999], Ped oni [1999], Ped oni [2004] and Wes e lund [2005] sug-
ges ha we can ejec he null hypo hesis o no coin eg a ion in all panels.
36 The esul s a e based on a educed- o m ixed e ec s panel es ima ion, and he pa ame e s
o he long- un ela ionship a e eco e ed om his es ima ion. See Bo ens ein e al. [1997]o
Blackbu ne and F ank [2007] o de ails.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 665
TABLE X
PANEL MEAN GROUP (PMG) AND DYNAMIC FIXED EFFECTS (DFE) ESTIMATES
PMG PMG DFE
[1] [2] [3]
Long-Run Rela ionship
C −11.321*** 1.183*** 1.198***
(0.013) (0.044) (0.048)
Indica o Va . o yea 2004 0.033*** 0.033*** 0.034***
(0.001) (0.001) (0.001)
C −1×SIC 0.261*** 0.234***
(0.080) (0.085)
Sho -Run Rela ionship
ECT −1−0.059*** −0.059*** −0.055***
(0.002) (0.002) (0.001)
ΔP −1−0.104*** −0.104*** −0.125***
(0.004) (0.004) (0.002)
ΔP −2−0.022*** −0.022*** −0.034***
(0.003) (0.003) (0.002)
ΔC 0.112*** 0.104*** 0.108***
(0.004) (0.004) (0.007)
ΔC −10.078*** 0.078*** 0.087***
(0.006) (0.006) (0.007)
ΔC −20.047*** 0.047*** 0.058***
(0.004) (0.004) (0.007)
Cons an 0.039*** 0.039*** 0.036***
(0.001) (0.001) (0.001)
# o obse a ions 197,262 197,262 197,262
No es: Reg ession esul s in columns 1 and 2 a e ob ained by using he panel mean g oup es ima o . Resul s
in column 3 a e based on a dynamic ixed e ec s es ima o . S anda d e o s a e epo ed in pa en heses.
*** signi ican a 1%, ** signi ican a 5%, * signi ican a 10% le el.
Fo easons o space, hey a e only b ie ly men ioned he e and a e epo ed in
Appendix A.
Fi s , we use a a he pa simonious model in he main speci ica ions.
Table A.II epo s eg ession esul s including a la ge numbe o con ol
a iables o accoun o p oduc and s a ion he e ogenei y. The poin es i-
ma es and he signi icance le els o he pa ame e es ima es on he sha e o
shoppe s 𝜆a e ba ely a ec ed by his modi ica ion, indica ing ha he e is
no omi ed a iable bias wi h espec o s a ion he e ogenei y.
Second, he same analysis is ca ied ou o sub-samples only: In one
speci ica ion we es ima e he same eg ession as epo ed in Table VI,bu
exclude highway loca ions. S a ions loca ed on highways a e o en conside ed
o cons i u e a sepa a e ma ke ,37 di e conside ably om s a ions o he
highway ega ding compe i ion, demand and egula ions, and may he e o e
exhibi e y di e en p ice dynamics. Excluding highway s a ions ensu es
ha ou esul s a e no d i en by his small g oup o s a ions. The eg essions
37 The Aus ian compe i ion au ho i y inds ha s a ions loca ed on highways usually cha ge
highe p ices, and ha he e is li le compe i ion be ween s a ions on he highway and o -highway
s a ions (G ube and Puglisi [2010]).
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
666 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
wi hou hese s a ions, epo ed in Table A.III, howe e , indica e ha he
esul s a e ha dly a ec ed by his modi ica ion. In a second speci ica ion
we exclude s a ions ha do no change p ices e y o en. In he ime pe iod
unde sc u iny, p ices a e changed e e y 4.35 days on a e age. I we exclude
10% o s a ions ha adjus ed hei p ices leas o en, he pa ame e es ima es
o he sha e o shoppe s a e ha dly a ec ed, as epo ed in Table A.IV.In
ano he speci ica ion, we exclude s a ions wi h he highes o he lowes
es ima ed pass- h ough a es 𝜌, o ensu e ha ou esul s a e no d i en by
ou lie s. The eg ession esul s excluding he 5% o s a ions wi h ei he he
highes o he lowes pass- h ough a es a e epo ed in Table A.V. Again,
he poin es ima es a e e y simila compa ed o he main esul s ( epo ed
in Table VI).
Thi d, we include he numbe o commu e s and he numbe o non-
commu e s (in loga i hmic e ms) ins ead o he sha e o in o med consume s.
The esul s a e p o ided in Table A.VI in Appendix A. The numbe o com-
mu e s has a signi ican ly posi i e e ec on he speed and he pass- h ough
a e, as well as on he asymme y in speed. We ind he opposi e e ec s o
non-shoppe s. A la ge numbe o unin o med cus ome s educes he speed
o cos ansmission and he pass- h ough a e signi ican ly. No signi ican
e ec o non-shoppe s is iden i ied o he asymme y in speed and p icing
h esholds. This seems o sugges ha i does no ma e whe he a la ge
sha e o shoppe s 𝜆is due o many shoppe s o a ew non-shoppe s. Bo h
consume g oups a e no signi ican ly ela ed o he asymme y in h esholds,
in line wi h he esul s o he main speci ica ions epo ed in Table VI.
Fou h, we use wo al e na i e ways o cons uc he sha e o in o med
consume s 𝜆by weigh ing one o bo h consume g oups di e en ly: (i) In
he i s es ima ion expe imen , we e ain om weigh ing commu e lows
by he numbe o po en ial ou es when calcula ing he sha e o in o med
consume s 𝜆.38 (ii) In a second es ima ion exe cise, we weigh commu e s
(non-commu e s) by he numbe o gas s a ions along hei commu ing ou e
(in hei municipali y o esidence). This is mo i a ed by he no ion ha
long-dis ance commu e s may pass by a la ge numbe o gasoline s a ions
and a e hus unlikely o e uel a a pa icula one. Simila ly, he p oba-
bili y o a speci ic s a ion’s a ac ing a non-commu e declines wi h he
numbe o i al s a ions loca ed in he same municipali y. The eg ession
esul s o bo h es ima ion expe imen s a e epo ed in Table A.VII and
Table A.VIII, espec i ely. In bo h cases, he impac o he sha e o shoppe s
on he speed and he pass- h ough a e o p ice ansmission is posi i e
and signi ican ly di e en om ze o, while we do no obse e a signi ican
e ec o ou measu es o consume in o ma ion on he asymme y o p ice
adjus men .
38 No e ha in he main speci ica ion, ansi commu e s a e weigh ed by he sha e o possible
ou es passing by a pa icula gasoline s a ion (see oo no e 16).
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 667
And inally, we examine he sensi i i y o ou esul s o da a equency. I is
well known ha agg ega ion o e ime can lead o biased pa ame e es ima es
i he in e empo al lag dis ibu ion is no p ope ly speci ied (Geweke [1978];
Bachmeie and G i in [2003]; Chesnes [2016]). Geweke [1978] sugges s a
‘p ac ical p ocedu e which will indica e in a quali a i e way he ex en o he
empo al agg ega ion p oblem in a pa icula applica ion,’ which in ol es
es ima ing he ela ion o in e es ‘a wo di e en le els o agg ega ion,
he ines possible and one o he ’ (p. 654). Following his sugges ion, we
e-es ima e ou model using weekly ins ead o daily p ice obse a ions. Mo e
speci ically, we use p ices o each gasoline s a ion on e e y Tuesday39 and
es ima e i s he long- un ela ionship be ween e ail p ices and cos s o
Equa ion 1and hen he e o -co ec ion models o Equa ions 2and 3.40 The
sequen ial es p ocedu e o es o he numbe o egimes, as ou lined in
Sec ion IV(i), suppo s a model wi h h ee egimes o 224 ou o 281 gasoline
s a ions a he 5% signi icance le el, whe eas o 57 s a ions he es sugges s
a linea e o -co ec ion model.
Using weekly p ice da a, he pa ame e es ima es o he pass- h ough
a e (𝜌) and a ec ed, as epo ed in Table A.IV he h esholds (𝜃+and 𝜃−)
a e simila o hose based on daily e ail p ices, epo ed in Table III.The
speed pa ame e s 𝛾a e subs an ially la ge in absolu e e ms when using
lowe - equency da a, because hey desc ibe he adjus men p ocess wi hin
a week a he han wi hin a day (see Table C.IV in Online Appendix C o
de ails). Based on hese pa ame e es ima es, we e-es ima e he c oss-sec ion
eg ession (4). Compa ing he esul s based on weekly da a, epo ed in
Table A.IX, wi h he analysis based on daily p ice obse a ions (see Table VI)
sugges s ha da a equency does no in luence he impac o consume in o -
ma ion on mos dimensions o p ice ansmission. The sha e o shoppe s
𝜆exe s a signi ican and posi i e impac on he speed o p ice adjus men
as well as on he pass- h ough a e, and does no con ibu e signi ican ly o
he a ia ion in he asymme y in speed. In con as o Table VI,howe e ,
Table A.IX sugges s a signi ican and nega i e impac on he asymme y
in h esholds: a la ge sha e o in o med consume s signi ican ly educes
asymme ies in h esholds, which unde lines he impo ance o in es iga ing
p icing h esholds (as sugges ed by Cab al and Fishman [2012], o example).
Da a equency hus seems o ma e when in es iga ing p icing asymme ies
in h esholds.
39 We selec p ices on e e y Tuesday, since i u ns ou ha p ices a e mos likely o adjus on
his day o he week. An al e na i e p ocedu e (no chosen he e) in ol es calcula ing (mo ing)
a e ages o e a pe iod o se en days. This would imply ha he calcula ed p ice possibly ne e
ac ually has been chosen. Using he weekly p ice se ies based on a pa icula day o he week
a oids his issue.
40 Simila ly o he analysis wi h daily da a, we use he Akaike in o ma ion c i e ia (AIC) es
s a is ics o selec he op imal lag o de o each gasoline s a ion. Fo he majo i y o he gasoline
s a ions we use a lag o de o wo.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
668 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
VI. DISCUSSION AND CONCLUSIONS
This a icle in es iga es he ela ionship be ween consume in o ma ion and
p ice dynamics. We u ilize high- equency p ice da a o indi idual e ail
gasoline s a ions o a leading e ically in eg a ed company o ob ain es i-
ma es o cos shocks (changes in c ude oil p ices) on he a ious dimensions
o p ice ansmission (i.e., he pass- h ough a e as well as he speed and
asymme y o p ice adjus men ). Ou measu e o consume in o ma ion
is cons uc ed by using de ailed da a on commu ing pa e ns. Commu e s
can eely sample p ices a gasoline s a ions along hei commu ing pa h
and hus end o be be e in o med han non-commu e s. We use da a
o a ime pe iod when websi es p o iding comp ehensi e and up- o-da e
in o ma ion on gasoline p ices we e no ye a ailable and isi ing a speci ic
gasoline s a ion was he only way o mos consume s o lea n abou cu en
gasoline p ices a ha s a ion. The iden i ica ion s a egy o he causal e ec
o consume in o ma ion on p ice adjus men elies on he ac ha ou
measu e o consume in o ma ion is de e mined by consume s’ long- un
decisions o commu e, which a e a guably independen o sho - un p ice
dynamics.
The esul s show ha gasoline s a ions wi h a highe sha e o in o med
consume s ha e la ge cos pass- h ough a es and adjus p ices mo e
quickly o exogenous cos shocks. This inding is consis en wi h implica-
ions o sea ch- heo e ic models (see Tappa a [2009]; Yang and Ye [2008];
Lewis [2011]). I also ela es o he li e a u e on p ice anspa ency and
i s e ec s on compe i ion by showing ha he p esence o be e in o med
consume s makes a ma ke mo e compe i i e. Recen empi ical e idence on
in o ma ion disclosu e in he e ail gasoline ma ke (Dewen e e al. [2017];
Luco [2019]) sugges s ha he an icompe i i e e ec o p ice anspa ency,
due o acili a ing collusion among i ms, ou weighs he compe i i e e ec
due o lowe consume sea ch cos s. No e ha he p esen s udy p o ides
an indica o o an ‘in o ma ion clea inghouse’ ha canno be accessed
by i ms and hus isola es and iden i ies he e ec o be e in o med
consume s.41
Ou analysis u he con ibu es o he la ge empi ical li e a u e on asym-
me ies in p ice adjus men s ( he ‘ ocke s and ea he s phenomenon’). While
a common s a egy o exis ing s udies o examine possible asymme ies
is o p e-speci y wo egimes in an ad-hoc manne , he p esen analysis
endogenously iden i ies mul iple egimes by applying a sequen ial model
41 I is wo h emphasizing ha we measu e in o ma ion abou p ices on he consume s’ side.
Access o p ice compa ison apps and websi es, in con as , would no only imp o e consume
in o ma ion bu would also make i easie o i ms o moni o each o he ’s p ices. In such se -
ings, inc eased anspa ency may hus acili a e collusion be ween i ms which would impede he
iden i ica ion o he e ec s o consume in o ma ion on p ice dynamics.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 669
selec ion app oach. This p ocedu e sugges s ha he p ice adjus men p o-
cess is bes cha ac e ized by h ee egimes ( wo h esholds). Es ima ing
mul iple h eshold e o -co ec ion models allows us o di e en ia e be ween
an asymme y in he speed o p ice ansmission and an asymme y in
h esholds. The empi ical analysis e eals (a) signi ican asymme ies in p ice
adjus men o a la ge numbe o gasoline s a ions and (b) subs an ial he e o-
genei y be ween gasoline s a ions wi h espec o bo h measu es o adjus men
asymme ies. Howe e , ou esul s do no p o ide unambiguous suppo o
heo e ical models sugges ing ha p icing asymme ies a e ela ed o he
deg ee o consume in o ma ion. We do no obse e a signi ican impac o
he sha e o commu e s (in o med consume s) on asymme ies in speed, and
he ela ionship be ween consume in o ma ion and asymme ies in p icing
h esholds seems o depend on da a equency.
Finally, ou s udy complemen s he empi ical li e a u e on he e ec s
o commu ing (as an indica o o consume s’ in o ma ion endowmen s) on
i ms’ p icing beha io in he e ail gasoline ma ke . Penne s o e e al. [2020]
d aw on simila measu es o consume in o ma ion based on he same da a on
commu ing,and ind ha p icele els a e lowe in egions wi h a la ge sha eo
commu e s, while he ela ionship be ween in o ma ion and p ice dispe sion
is cha ac e ized by an in e se-U. Thei empi ical in es iga ion—co e ing a
longe ime pe iod ( om Oc obe , 1999, o Ma ch, 2005) and a la ge numbe
o gas s a ions—d aws on p ice in o ma ion collec ed much less equen ly
and uses qua e ly da a. Penne s o e e al. [2020] a e in e es ed in long- un
di e ences in p ice le els and p ice dispe sion be ween local ma ke s wi h
poo ly and well in o med consume s only. In con as , he p esen a icle uses
high- equency in o ma ion on bo h p ices and ma ginal cos s, which allows
us o in es iga e sho - un (in ac , daily) p ice esponses o cos changes as
well as hei ela ion o consume in o ma ion. Based on he indings o bo h
s udies, we conclude ha consume in o ma ion no only in luences he le el
o p ices and p ice dispe sion, bu also he dynamics o p ice adjus men and
hus he unc ioning o ma ke s.
In he end, we poin ou h ee impo an issues in his esea ch a ea which
ha e no been con incingly add essed so a (including he p esen analysis).
Fi s , while commu ing is a sou ce o in o ma ion o consume s ( a he han
i ms) and gene a es an a guably exogenous measu e o consume in o ma-
ion, i is impo an o no e ha commu e s and non-commu e s also di e
in he anspo a ion cos s hey incu when isi ing gasoline s a ions. We
p o ide a i s a emp o di e en ia e be ween he impac o anspo a ion
cos s and consume in o ma ion on p ice ansmission in his a icle. Mo e
de ailed empi ical and heo e ical wo k on his issue mus be pos poned o
u u e esea ch. Second, consume in o ma ion could also in luence ma ke
s uc u e. A highe sha e o in o med consume s inc eases he p ice elas ici y
o demand, lowe s i ms’ p ice-cos ma gins and hus educes he numbe o
i ms in equilib ium. In he con ex o he p esen pape , his would imply
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
670 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
ha consume in o ma ion could in luence p ice ansmission indi ec ly ia a
‘ma ke s uc u e channel.’42 While he impac o in o ma ion ( anspa ency)
on ma ke s uc u e (numbe o i ms and p oduc a ie y) is well unde s ood
in heo y (Schul z [2009]; Gu and Wenzel [2011]), a ca e ul empi ical analysis
(ex ending Seim’s [2006] analysis, o example) is s ill lacking. Thi d, his
analysis applies a ious measu es o he g adual adjus men o indi idual
p ices. Howe e , we do no explici ly add ess one impo an dimension o
p ice adjus men : he equency o p ice changes. P ices do change in e-
quen ly o many e ail ansac ions and a ious explana ions ha e been
p omo ed o explaining di e en spells o unchanged p ices (Wolman [2007];
Nakamu a and S einsson [2008]). In es iga ing he impac o consume
in o ma ion on he equency o p ice adjus men would u he imp o e
ou unde s anding abou he dynamics o p ice adjus men and hus he
unc ioning o ma ke s.
APPENDIX A
ADDITIONAL REGRESSION TABLES
TABLE A.I
PARAMETRIC RESULTS OF SEMI-PARAMETRIC REGRESSIONS
Speed Pass- h ough
a e Asymme y
in speed Asymme y
in h eshold
# o s a ions wi hin 2 km 0.181*** 0.167 −0.059 1.430*
(in 100) (0.024) (0.208) (0.052) (0.851)
Dis ance o nea es i al s a ion −0.158** −1.356** −0.505*** −0.598
(in 100 km) (0.074) (0.653) (0.163) (2.668)
S a ion is loca ed on highway −0.027*** −0.007 −0.062*** 0.264
(0.006) (0.056) (0.014) (0.227)
# o obse a ions 281 281 281 281
R20.292 0.025 0.124 0.015
Ha dle and Mammen [1993] speci ica ion es s a is ic
polynomial o deg ee 0 4.034 1.832 0.535 2.119
[p =0.00] [p =0.06] [p =0.48] [p =0.04]
polynomial o deg ee 1 1.877 0.979 1.132 2.776
[p =0.05] [p =0.34] [p =0.32] [p =0.00]
polynomial o deg ee 2 1.468 1.045 0.429 2.418
[p =0.12] [p =0.29] [p =0.76] [p =0.00]
No es: S anda d e o s a e epo ed in pa en heses. *** signi ican a 1%, ** signi ican a 5%,
* signi ican a 10% le el. Ha dle and Mammen [1993] speci ica ion es s e alua e whe he he non-pa ame ic
i can be app oxima ed by a pa ame ic adjus men o o de ze o, one o wo, espec i ely (H0: pa ame ic
and non-pa ame ic i s a e no di e en ). In e ence is based on a wild boo s ap es ing p ocedu e wi h 100
eplica ions.
42 We a e hank ul o a e e ee o making us awa e o his po en ial e ec .
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 677
TABLE A.VIII
ESTIMATION RESULTS FOR SPEED,PASS-THROUGH RATE AND ASYMMETRY OF PRICE TRANSMISSION,USING ALTERNATIVE WEIGHTS
Speed Pass- h ough
a e Asymme y
in speed Asymme y
in h eshold
OLS WLS OLS WLS OLS WLS OLS WLS
Sha e o shoppe s (𝜆al e na i e weigh s) 0.040*** 0.041*** 0.306*** 0.281*** 0.013 0.015 0.202 −0.066
(0.009) (0.010) (0.082) (0.080) (0.021) (0.021) (0.383) (0.381)
# o s a ions wi hin 2 km 0.138*** 0.138*** −0.166 −0.136 −0.055 −0.020 0.340 1.399*
(in 100) (0.026) (0.021) (0.143) (0.165) (0.051) (0.046) (0.701) (0.842)
Dis ance o nea es i al s a ion −0.131*** −0.128** −0.797* −0.794 −0.330*** −0.317** 0.981 0.812
(0.040) (0.058) (0.408) (0.491) (0.118) (0.129) (1.780) (2.359)
S a ion is loca ed on highway −0.029*** −0.029*** −0.026 −0.013 −0.067*** −0.067*** 0.276* 0.180
(0.005) (0.006) (0.033) (0.048) (0.012) (0.012) (0.163) (0.227)
Cons an 0.054*** 0.052*** 1.242*** 1.241*** 0.051*** 0.046*** −0.164 −0.082
(0.003) (0.003) (0.035) (0.029) (0.008) (0.008) (0.121) (0.135)
# o obse a ions 281 281 281 281 281 281 281 281
R20.349 0.367 0.063 0.064 0.119 0.146 0.009 0.011
No es: S anda d e o s a e epo ed in pa en heses. In OLS speci ica ions s anda d e o s a e based on a obus a iance es ima o . In WLS speci ica ions each obse a ion is
weigh ed by he in e se o he s anda d e o s o he dependen a iable es ima es. *** signi ican a 1%, ** signi ican a 5%, * signi ican a 10% le el.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
678 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
TABLE A.IX
ESTIMATION RESULTS FOR SPEED,PASS-THROUGH RATE AND ASYMMETRY OF PRICE TRANSMISSION BASED ON WEEKLY DATA
Speed Pass- h ough
a e Asymme y
in speed Asymme y
in h eshold
OLS WLS OLS WLS OLS WLS OLS WLS
[1] [2] [3] [4] [5] [6] [7] [8]
Sha e o shoppe s (𝜆) 0.080** 0.082* 0.364*** 0.342*** 0.135 0.126 −1.063** −1.326***
(0.039) (0.042) (0.094) (0.091) (0.120) (0.128) (0.411) (0.434)
# o s a ions wi hin 2 km 0.267*** 0.281*** 0.250 0.280 −0.413* −0.363 −2.322*** −2.731***
(in 100) (0.071) (0.082) (0.154) (0.180) (0.218) (0.248) (0.692) (0.898)
Dis ance o nea es i al S a ion −0.059 −0.050 0.132 0.139 −0.392 −0.387 −6.593** −5.686*
(in 100 km) (0.327) (0.324) (0.397) (0.508) (1.029) (0.985) (3.035) (3.022)
S a ion is loca ed on he Highway −0.022 −0.022 0.030 0.036 −0.038 −0.039 0.254 0.189
(0.025) (0.022) (0.039) (0.049) (0.094) (0.067) (0.238) (0.190)
Cons an 0.219*** 0.212*** 1.100*** 1.103*** 0.077 0.073 0.951*** 1.108***
(0.026) (0.027) (0.066) (0.060) (0.078) (0.083) (0.267) (0.281)
# o obse a ions 224 224 281 281 224 224 224 224
R20.067 0.071 0.051 0.053 0.030 0.023 0.051 0.064
No es: S anda d e o s a e epo ed in pa en heses. In OLS speci ica ions s anda d e o s a e based on a obus a iance es ima o . In WLS speci ica ions each obse a ion is
weigh ed by he in e se o he s anda d e o s o he dependen a iable es ima es. *** signi ican a 1%, ** signi ican a 5%, * signi ican a 10% le el.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 679
REFERENCES
And ews, D. W. K. and Plobe ge , W., 1994, ‘Op imal Tes s when a Nuisance Pa ame e
Is P esen Only unde he Al e na i e,’ Econome ica, 62(6), pp. 1383–1414.
ARBÖ, 2000, ‘Wi sagen Ihnen wo man billig ank und wie man Geld spa en kann,’ p ess
elease OTS0124, See h ps://www.o s.a /p esseaussendung/OTS_20000922_OTS0124,
las accessed June 8 h, 2020.
Bachmeie , L. J. and G i in, J. M., 2003, ‘New E idence on Asymme ic Gasoline P ice
Responses,’ The Re iew o Economics and S a is ics, 85(3), pp. 772–776.
Bacon, R. W., 1991, ‘Rocke s and Fea he s: The Asymme ic Speed o Adjus men o
U.K. Re ail Gasoline P ices o Cos Changes,’ Ene gy Economics, 13(3), pp. 211–218.
Bakucs, Z.; Falkowsky, J. and Fe ö, I., 2014, ‘Does Ma ke S uc u e In luence P ice
T ansmission in he Ag o-Food Sec o ? A Me a-Analysis Pe spec i e,’ Jou nal o Ag i-
cul u al Economics, 65(1), pp. 1–25.
Balke, N. and Fomby, T., 1997, ‘Th eshold Coin eg a ion,’ In e na ional Economic
Re iew, 38(3), pp. 627–45.
Ball, L. and Mankiw, N. G., 1994, ‘Asymme ic P ice Adjus men and Economic Fluc-
ua ions,’ Economic Jou nal, 104(423), pp. 247–61.
Ba on, J.; Taylo , B. and Umbeck, J. R., 2004, ‘Numbe o Selle s, A e age P ices,
and P ice Dispe sion,’ In e na ional Jou nal o Indus ial O ganiza ion, 22(8-9),
pp. 1041–1066.
Ba on, J. M.; Taylo , B. A. and Umbeck, J. R., 2000, ‘A Theo y o Quali y-Rela ed
Di e ences in Re ail Ma gins: Why The e Is a “P emium” on P emium Gasoline,’
Economic Inqui y, 38(4), pp. 550–569.
Baye, M. R. and Mo gan, J., 2001, ‘In o ma ion Ga ekeepe s on he In e ne and he
Compe i i eness o Homogeneous P oduc Ma ke s,’ Ame ican Economic Re iew,
91(3), pp. 454–474.
Baye, M. R.; Mo gan, J. and Schol en, P., 2004, ‘P ice Dispe sion in he Small and in
he La ge: E idence om an In e ne P ice Compa ison Si e,’ Jou nal o Indus ial
Economics, 52(4), pp. 463–496.
Blackbu ne, E. F. I. and F ank, M. W., 2007, ‘Es ima ion o Nons a iona y He e oge-
neous Panels,’ S a a Jou nal, 7(2), pp. 197–212.
Bo ens ein, S.; Came on, A. C. and Gilbe , R., 1997, ‘Do Gasoline P ices Respond
Asymme ically o C ude Oil P ice Changes?’ The Qua e ly Jou nal o Economics,
112(1), pp. 305–339.
B own, J. R. and Goolsbee, A., 2002, ‘Does he In e ne Make Ma ke s Mo e Compe i-
i e? E idence om he Li e Insu ance Indus y,’ Jou nal o Poli ical Economy, 110(3),
pp. 481–507.
By ne, D. P. and de Roos, N., 2017, ‘Consume Sea ch in Re ail Gasoline Ma ke s,’ The
Jou nal o Indus ial Economics, 65(1), pp. 183–193.
Cab al, L. and Fishman, A., 2012, ‘Business as Usual: A Consume Sea ch Theo y o
S icky P ices and Asymme ic P ice Adjus men ,’ In e na ional Jou nal o Indus ial
O ganiza ion, 30(4), pp. 371–376.
Chambe lin, E., 1948, The Theo y o Monopolis ic Compe i ion, 6 h edi ion (Ha a d
Uni e si y P ess, Camb idge, Massachuse s, U.S.A.).
Chan, K. S., 1993, ‘Consis ency and Limi ing Dis ibu ion o he Leas Squa es Es ima-
o o a Th eshold Au o eg essi e Model,’ The Annals o S a is ics, 21(1), pp. 520–533.
Chand a, A. and Tappa a, M., 2011, ‘Consume Sea ch and Dynamic P ice Dispe -
sion: An Applica ion o Gasoline Ma ke s,’ The RAND Jou nal o Economics, 42(4),
pp. 681–704.
Chen, L.-H.; Finney, M. and Lai, K. S., 2005, ‘A Th eshold Coin eg a ion Analysis
o Asymme ic P ice T ansmission om C ude Oil o Gasoline P ices,’ Economics
Le e s, 89(2), pp. 233–239.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
680 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
Chesnes, M., 2016, ‘Asymme ic Pass-Th ough in U.S. Gasoline P ices,’ The Ene gy Jou -
nal, 37(1).
Claycombe, R. J., 1991, ‘Spa ial Re ail Ma ke s,’ In e na ional Jou nal o Indus ial O ga-
niza ion, 9(2), pp. 303–313.
Coope , Thomas E. and Jones, John T., 2007, ‘Asymme ic Compe i ion on Com-
mu e Rou es: The Case o Gasoline P icing,’ Sou he n Economic Jou nal, 74(2),
pp. 483–504.
Da ies, R. B., 1987, ‘Hypo hesis Tes ing when a Nuisance Pa ame e Is P esen Only
unde he Al e na i es,’ Biome ika, 74(1), pp. 33–43.
De los San os, B.; Ho acsu, A. and Wildenbees , M. R., 2012, ‘Tes ing Models o Con-
sume Sea ch Beha io Using Da a on Web B owsing and Consume Pu chases,’ The
Ame ican Economic Re iew, 102(6), pp. 2955–2980.
Del as, G., 2008, ‘Re ail Gasoline P ice Dynamics and Local Ma ke Powe ,’ The Jou nal
o Indus ial Economics, 56(3), pp. 613–628.
Dewen e , R.; Heimesho , U. and Lü h, H., 2017, ‘The Impac o he Ma ke T ans-
pa ency Uni o Fuels on Gasoline P ices in Ge many,’ Applied Economics Le e s,
24(5), pp. 302–305.
Dickey, D. A. and Fulle , W. A., 1979, ‘Dis ibu ion o he Es ima o s o Au o eg essi e
Time Se ies wi h a Uni Roo ,’ Jou nal o he Ame ican S a is ical Associa ion, 74(366),
pp. 427–431.
Ecka d, E. W., 2004, ‘The “Law o One P ice” in 1901,’ Economic Inqui y, 42(1),
pp. 101–110.
Ecke , A., 2013, ‘Empi ical S udies o Gasoline Re ailing: A Guide o he Li e a u e,’
Jou nal o Economic Su eys, 27(1), pp. 140–166.
Ecke , A. and Wes , D. S., 2005, ‘P ice Uni o mi y and Compe i ion in a Re ail Gasoline
Ma ke ,’ Jou nal o Economic Beha io & O ganiza ion, 56(2), pp. 219–237.
Ellison, G. and Ellison, S. F., 2005, ‘Lessons abou Ma ke s om he In e ne ,’ Jou nal
o Economic Pe spec i es, 19(2), pp. 139–158.
Ellison, G. and Ellison, S. F., 2009, ‘Sea ch, Ob usca ion, and P ice Elas ici ies on he
In e ne ,’ Econome ica, 77(2), pp. 427–452.
Ende s, W. and Siklos, P. L., 2001, ‘Coin eg a ion and Th eshold Adjus men ,’ Jou nal
o Business & Economic S a is ics, 19(2), pp. 166–176.
Engle, R. F. and G ange , C. W. J., 1987, ‘Co-In eg a ion and E o Co ec ion: Rep e-
sen a ion, Es ima ion, and Tes ing,’ Econome ica, 55(2), pp. 251–276.
F ey, G. and Mane a, M., 2007, ‘Econome ic Models o Asymme ic P ice T ansmis-
sion,’ Jou nal o Economic Su eys, 21(2), pp. 349–415.
Galeo i, M.; Lanza, A. and Mane a, M., 2003, ‘Rocke s and Fea he s Re isi ed: an
In e na ional Compa ison on Eu opean Gasoline Ma ke s,’ Ene gy Economics, 25(2),
pp. 175–190.
Geweke, J., 1978, ‘Tempo al Agg ega ion in he Mul iple Reg ession Model,’ Econome -
ica, 46(2), pp. 643–661.
Godby, R.; Lin ne , A. M.; S engos, T. and Wandschneide , B., 2000, ‘Tes ing o Asym-
me ic P icing in he Canadian Re ail Gasoline Ma ke ,’ Ene gy Economics, 22(3),
pp. 349–368.
G asso, M. and Mane a, M., 2007, ‘Asymme ic E o Co ec ion Models o he
Oil-Gasoline P ice Rela ionship,’ Ene gy Policy, 35(1), pp. 156–177.
G ube , J. and Puglisi, L., 2010, ‘En wicklung de Ve kau sma gen ü Diesel und Supe -
benzin an Au obahn- und Nich -Au obahn anks ellen in Ös e eich,’ echnical epo
(Bundeswe bewe bsbehö de [Fede al Compe i ion Au ho i y], Vienna, Aus ia).
Gu, Y. and Wenzel, T., 2011, ‘T anspa ency, P ice-Dependen Demand and P oduc Va i-
e y,’ Economics Le e s, 110, pp. 216–219.
Gugle , K.; Heim, S.; Janssen, M. and Liebens eine , M., 2018, ‘Ma ke Libe aliza ion:
P ice Dispe sion, P ice Disc imina ion and Consume Sea ch in he Ge man Elec ici y
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 681
Ma ke s,’ ZEW discussion pape s no. 18-042 (Zen um ü Eu opäische Wi scha s-
o schung, Mannheim, Ge many).
Hansen, B., 1996, ‘In e ence When a Nuisance Pa ame e Is No Iden i ied unde he
Null Hypo hesis,’ Econome ica, 64(2), pp. 413–430.
Hansen, B., 1997, ‘In e ence in TAR Models,’ S udies in Nonlinea Dynamics & Econo-
me ics, 2(1), pp. 1–16.
Hansen, B., 1999, ‘Tes ing o Linea i y,’ Jou nal o Economic Su eys, 13(5),
pp. 551–576.
Hansen, B. E. and Seo, B., 2002, ‘Tes ing o Two-Regime Th eshold Coin eg a ion in
Vec o E o -Co ec ion Models,’ Jou nal o Econome ics, 110(2), pp. 293–318 (Long
memo y and nonlinea ime se ies).
Ha dle, W. and Mammen, E., 1993, ‘Compa ing Nonpa ame ic Ve sus Pa ame ic
Reg ession Fi s,’ Annals o S a is ics, 21(4), pp. 1926–1947.
Hassouneh, I.; Hols , C.; Se a, T.; on C amon-Taubadel, S. and Gil, J. M., 2015,
‘O e iew o P ice T ansmission and Reasons o Di e en Adjus men Pa e ns
ac oss EU Membe S a es,’ McCo is on, S., (ed.), Food P ice Dynamics and P ice
Adjus men in he EU (Ox o d Uni e si y P ess, Ox o d, England).
Has ings, J. S., 2004, ‘Ve ical Rela ionships and Compe i ion in Re ail Gasoline Ma -
ke s: Empi ical E idence om Con ac Changes in Sou he n Cali o nia,’ Ame ican
Economic Re iew, 94(1), pp. 317–328.
Has ings, J. S. and Gilbe , R. J., 2005, ‘Ma ke Powe , Ve ical In eg a ion and he
Wholesale P ice o Gasoline,’ The Jou nal o Indus ial Economics, 53(4), pp. 469–492.
Hayek, F., 1945, ‘The Use o Knowledge in Socie y,’ The Ame ican Economic Re iew,
35(4), pp. 519–530.
Hosken, D. S.; McMillan, R. S. and Taylo , C. T., 2008, ‘Re ail Gasoline P icing: Wha
Do We Know?’ In e na ional Jou nal o Indus ial O ganiza ion, 26(6), pp. 1425–1436.
Houde, J.-F., 2012, ‘Spa ial Di e en ia ion and Ve ical Me ge s in Re ail Ma ke s o
Gasoline,’ Ame ican Economic Re iew, 102(5), pp. 2147–2182.
Johnson, R. N., 2002, ‘Sea ch Cos s, Lags and P ices a he Pump,’ Re iew o Indus ial
O ganiza ion, 20(1), pp. 33–50.
Kao, C., 1999, ‘Spu ious Reg ession and Residual-Based Tes s o Coin eg a ion in Panel
Da a,’ Jou nal o Econome ics, 90(1), pp. 1–44.
Kouya é, C. and on C amon-Taubadel, S., 2016, ‘Dis ance and Bo de E ec s o
P ice T ansmission: A Me a-Analysis,’ Jou nal o Ag icul u al Economics, 67(2),
pp. 255–271.
Lewis, J. B. and Linze , D. A., 2005, ‘Es ima ing Reg ession Models in which he Depen-
den Va iable Is Based on Es ima es,’ Poli ical Analysis, 13(04), pp. 345–364.
Lewis, M. S., 2008, ‘P ice Dispe sion and Compe i ion wi h Di e en ia ed Selle s,’ Jou -
nal o Indus ial Economics, 56(3), pp. 654–678.
Lewis, M. S., 2011, ‘Asymme ic P ice Adjus men and Consume Sea ch: An Examina-
ion o he Re ail Gasoline Ma ke ,’ Jou nal o Economics & Managemen S a egy,
20(2), pp. 409–449.
Lewis, M. S. and Ma el, H. P., 2011, ‘When Do Consume s Sea ch?’ The Jou nal o
Indus ial Economics, 59(3), pp. 457–483.
Lloyd, T., 2017, ‘Fo y Yea s o P ice T ansmission Resea ch in he Food Indus-
y: Insigh s, Challanges and P ospec s,’ Jou nal o Ag icul u al Economics, 68(1),
pp. 3–21.
Loy, J.-P. and Weiss, C. R., 2019, ‘P oduc Di e en ia ion and Cos Pass-Th ough,’ Jou -
nal o Ag icul u al Economics, 70(3), pp. 840–858.
Luco, F., 2019, ‘Who Bene i s om In o ma ion Disclosu e? The Case o Re ail Gaso-
line,’ Ame ican Economic Jou nal: Mic oeconomics, 11(2), pp. 277–305.
Lue kepohl, H., 1985, ‘Compa ison o C i e ia o Es ima ing he O de o a Vec o
Au o eg essi e P ocess,’ Jou nal o Time Se ies Analysis, 6(1), pp. 35–52.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
682 J.-P. LOY, D. PENNERSTORFER, D. RROSHI, C. WEISS, B. YONTCHEVA
Ma el, H. P., 1976, ‘The Economics o In o ma ion and Re ail Gasoline P ice Beha io :
An Empi ical Analysis,’ Jou nal o Poli ical Economy, 84(5), pp. 1033–1060.
Meye , J. and on C amon-Taubadel, S., 2004, ‘Asymme ic P ice T ansmission: A Su -
ey,’ Jou nal o Ag icul u al Economics, 55(3), pp. 581–611.
Mo on, F. S.; Ze elmeye , F. and Sil a-Risso, J., 2001, ‘In e ne Ca Re ailing,’ The
Jou nal o Indus ial Economics, 49(4), pp. 501–519.
Nakamu a, E. and S einsson, J., 2008, ‘Fi e Fac s abou P ices: A Ree alua ion o Menu
Cos Models,’ The Qua e ly Jou nal o Economics, 123(4), pp. 1415–1464.
Nielsen, B., 2001, ‘O de De e mina ion in Gene al Vec o Au o eg essions,’ Economics
pape s 2001-W10 (Economics G oup, Nu ield College, Uni e si y o Ox o d, Ox o d,
England).
Nishida, M. and Reme , M., 2018, ‘The De e minan s and Consequences o Sea ch
Cos He e ogenei y: E idence om Local Gasoline Ma ke s,’ Jou nal o Ma ke ing
Resea ch, 55(3), pp. 305–320.
ÖAMTC, 2003, ‘Sp i p eis-Ve gleich lohn sich im Ku zu laub zu P ings en,’ p ess elease
OTS0067, See h ps://www.o s.a /p esseaussendung/OTS_20030604_OTS0067,las
accessed June 8 h, 2020.
O lo , E., 2011, ‘How Does he In e ne In luence P ice Dispe sion? E idence om he
Ai line Indus y,’ Jou nal o Indus ial Economics, 59(1), pp. 21–37.
Paulsen, J., 1984, ‘O de De e mina ion o Mul i a ia e Au o eg essi e Time Se ies wi h
Uni Roo s,’ Jou nal o Time Se ies Analysis, 5(2), pp. 115–127.
Ped oni, P., 1999, ‘C i ical Values o Coin eg a ion Tes s in He e ogeneous Panels wi h
Mul ipleReg esso s,’Ox o dBulle in o Economicsand S a is ics,61(S1), pp. 653–670.
Ped oni, P., 2004, ‘Panel Coin eg a ion: Asymp o ic and Fini e Sample P ope ies o
Pooled Time Se ies Tes s wi h an Applica ion o he PPP Hypo hesis,’ Econome ic
Theo y, 20(3), pp. 597–625.
Penne s o e ,D., 2009, ‘Spa ial P ice Compe i ion in Re ail Gasoline Ma ke s: E idence
om Aus ia,’ The Annals o Regional Science, 43(1), pp. 133–158.
Penne s o e , D.; Schmid -Dengle , P.; Schu z, N.; Weiss, C. and Yon che a, B., 2020,
‘In o ma ion and P ice Dispe sion: Theo y and E idence,’ In e na ional Economic
Re iew, 61(2), pp. 871–899.
Pesa an, M. H., 2015, ‘Tes ing Weak C oss-Sec ional Dependence in La ge Panels,’
Econome ic Re iews, 34(6-10), pp. 1089–1117.
Pesa an, M. H.; Shin, Y. and Smi h, R. P., 1999, ‘Pooled Mean G oup Es ima ion o
Dynamic He e ogeneous Panels,’ Jou nal o he Ame ican S a is ical Associa ion,
94(446), pp. 621–634.
Phillips, P. C. B. and Pe on, P., 1988, ‘Tes ing o a Uni Roo in Time Se ies Reg ession,’
Biome ika, 75(2), pp. 335–346.
Rai h, M., 1996, ‘Spa ial Re ail Ma ke s wi h Commu ing Consume s.’ In e na ional
Jou nal o Indus ial O ganiza ion, 14(4), pp. 447–463.
Reme , M., 2015, ‘An Empi ical In es iga ion o he De e minan s o Asymme ic P ic-
ing,’ In e na ional Jou nal o Indus ial O ganiza ion, 42(C), pp. 46–56.
Robinson, P. M., 1988, ‘Roo -N-Consis en Semipa ame ic Reg ession,’ Econome ica,
56(4), pp. 931–954.
Schul z, C., 2009, ‘T anspa ency and P oduc Va ie y,’ Economics Le e s, 102,
pp. 165–168.
Seim, K., 2006, ‘An Empi ical Model o Fi m En y wi h Endogenous P oduc -Type
Choices,’ The RAND Jou nal o Economics, 37(3), pp. 619–640.
Sengup a, A. and Wiggins, S. N., 2014, ‘Ai line P icing, P ice Dispe sion, and Ticke
Cha ac e is ics on and o he In e ne ,’ Ame ican Economic Jou nal: Economic Policy,
6(1), pp. 272–307.
She man, J. and Weiss, A., 2017, ‘On F ui ul and Fu ile Tes s o he Rela ionship
be ween Sea ch and P ice Dispe sion,’ Economic Inqui y, 55(4), pp. 1898–1918.
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.
CONSUMER INFORMATION AND PRICE TRANSMISSION 683
So ensen, A. T., 2000, ‘Equilib ium P ice Dispe sion in Re ail Ma ke s o P esc ip ion
D ugs,’ Jou nal o Poli ical Economy, 108(4), pp. 833–850.
S a is ics Aus ia, 2019, ‘IKT-Einsa z in Haushal en: Einsa z on In o ma ions-und
Kommunika ions echnologien in Haushal en 2018,’ echnical epo (S a is ics Aus ia,
Guglgasse 13, Vienna, Aus ia).
S igle , G., 1961, ‘The Economics o In o ma ion,’ Jou nal o Poli ical Economy, 69.
S ikholm, B. and Te äs i a, T., 2015, ‘De e mining he Numbe o Regimes in a Th esh-
old Au o eg essi e Model Using Smoo h T ansi ion Au o eg essions,’ SSE/EFI Wo k-
ing Pape s in Economics and Finance 578 (Economic Resea ch Ins i u e, S ockholm
School o Economics, S ockholm, Sweden).
Tang, Z.; Smi h, M. D. and Mon gome y, A., 2010, ‘The Impac o Shopbo Use on
P icesand P iceDispe sion:E idence om OnlineBook Re ailing,’ In e na ional Jou -
nal o Indus ial O ganiza ion, 28(6), pp. 579–590.
Tappa a, M., 2009, ‘Rocke s and Fea he s: Unde s anding Asymme ic P icing,’ The
RAND Jou nal o Economics, 40(4), pp. 673–687.
Toda, H. Y. and Yamamo o, T., 1995, ‘S a is ical In e ence in Vec o Au o eg essions
wi h Possibly In eg a ed P ocesses,’ Jou nal o Econome ics, 66(1), pp. 225–250.
Tsay, R. S., 1989, ‘Tes ing and Modeling Th eshold Au o eg essi e P ocesses,’ Jou nal
o he Ame ican S a is ical Associa ion, 84(405), pp. 231–240.
Wes e lund, J., 2005, ‘New Simple Tes s o Panel Coin eg a ion,’ Econome ic Re iews,
24(3), pp. 297–316.
Weyl, E. G. and Fabinge , M., 2013, ‘Pass-Th ough as an Economic Tool: P inciples
o Incidence unde Impe ec Compe i ion,’ Jou nal o Poli ical Economy, 121(3),
pp. 528–583.
Wolman, A. L., 2007, ‘The F equency and Cos s o Indi idual P ice Adjus men ,’ Man-
age ial and Decision Economics, 28(6), pp. 531–552.
Xia, T. and Li, X., 2010, ‘Consump ion Ine ia and Asymme ic P ice T ansmission,’
Jou nal o Ag icul u al and Resou ce Economics, 35(2).
Yang, H. and Ye, L., 2008, ‘Sea ch wi h Lea ning: Unde s anding Asymme ic P ice
Adjus men s,’ The RAND Jou nal o Economics, 39(2), pp. 547–564.
SUPPORTING INFORMATION
Addi ional suppo ing in o ma ion may be ound in he online e sion o his a icle a he
publishe ’s web si e: h p://wileyonlinelib a y.com/jou nal/joie o ia The Jou nal’s websi e,
h p://www.jindec.o g
© 2022 The Au ho s. The Jou nal o Indus ial Economics published by The Edi o ial Boa d o The Jou nal o Indus ial Eco-
nomics and John Wiley & Sons L d.