Munich Pe sonal RePEc A chi e
Single s. mul iple disclosu es in an
expe imen al asse ma ke wi h
in o ma ion acquisi ion
Ruiz-Bu o n, Alba and Al a ano, Simone and
Camacho-Cuena, E a and Mo one, And ea
Uni e si a Jaume I, Uni e si à degli S udi Aldo Mo o di Ba i
May 2020
Online a h ps://mp a.ub.uni-muenchen.de/101035/
MPRA Pape No. 101035, pos ed 17 Jun 2020 09:49 UTC
Single s. mul iple disclosu es in an expe imen al asse ma ke
wi h in o ma ion acquisi ion
Alba Ruiz-Bu o naSimone Al a anoa
E a Camacho-Cuenaa∗And ea Mo oneb
aEconomics Depa men , Uni e si y Jaume I, A enida Sos Bayna s/n 12071, Cas ell´on, Spain
bEconomics, Managemen and Co po a e Law Depa men , Uni e si `a degli S udi Aldo Mo o di Ba i, I aly
May 2020
Abs ac
We conduc labo a o y expe imen s o s udy whe he inc easing he numbe o
independen public signals in an economy wi h endogenous p i a e in o ma ion is
an e ec i e measu e o p omo e he acquisi ion o in o ma ion and o enhance p ice
e iciency. We obse e ha he elease o public in o ma ion c owds ou he ade s’
demand o p i a e in o ma ion unde a single disclosu e while a o ing p i a e in-
o ma ion acquisi ion unde mul iple disclosu es. The la e measu e imp o es p ice
accu acy in o ecas ing he asse undamen al alue. Howe e , mul iple disclosu es
do no elimina e he ad e se e ec o ma ke o e eac ion o public in o ma ion,
becoming a po en ial sou ce o agili y o he inancial sys em.
Keywo ds Expe imen s; Financial ma ke s; Public in o ma ion; In o ma ion acquisi-
ion; Mul iple disclosu es
JEL codes C92, D82, G14
∗CONTACT E a Camacho-Cuena. Email: [email p o ec ed]
1 In oduc ion
P io o he inancial c isis o 2008, he commonly accep ed pa adigm in Financial Eco-
nomics was cha ac e ized by a s ong belie e in he sel - egula ing o ces o un es ained
inancial ma ke s, he e iciency o asse -p ice o ma ion, and he inc eased e ec i eness
in isk alloca ion and sha ing h ough he in oduc ion o e e mo e complex inancial
ins umen s. Wi hin he e icien ma ke pa adigm, ading in s ock ma ke s has been
usually associa ed o he ansmission and dissemina ion o in o ma ion abou he isk
o he unde lying asse s. Be o e he c isis, he isk e alua ion o such p oduc s was ou -
sou ced o a ing agencies and ypically, he buye o CDO anches would no ha e spen
any e o himsel on in o ma ion acquisi ion conce ning his a away coun e pa s. The
in o ma ion ansmission and he e o o ga he ing independen “pieces” o in o ma ion,
he e o e, seemed o ha e b oken down in he case o s uc u ed inancial p oduc s. The
ansi ion om a decen alized in o ma ion ga he ing e o o inancial in es o s o i s
ou sou cing o a ing agencies o o he ele an public ins i u ions migh ha e con ibu ed
o enhance he agili y o he en i e inancial sys em be o e he c isis. The op imis ic
ecommenda ions o a ing agencies, o ins ance, we e “blindly” ollowed by he as
majo i y o in es o s, which led o a de imen al unde es ima ion o ce ain isks.
In o de o educe he agili y o he inancial sys em, he disclosu e o in o ma ion
has been a he o e on o egula o y e o s o imp o e inancial ma ke s abili y. Wi h
he aim a p o iding an en i onmen wi h mo e p ecise public in o ma ion, egula o y
measu es ha e been in oduced in o de o inc ease he eliabili y o c edi a ings and he
quali y and quan i y o disclosu es o mac oeconomic in o ma ion (see Golds ein e al.,
2014). Fo example, he Dodd-F ank Wall S ee Re o m and Consume P o ec ion Ac
o 2010 aims a inc easing he gene al le el o anspa ency, imp o ing inancial s abil-
i y and consume s p o ec ion (Baily e al., 2017). The Eu opean Cen al Bank u he
in ensi ied i s o wa d guidance ac i i y, disclosing mo e in o ma ion abou i s u u e
mone a y policy in en ions. Howe e , hese measu es may ha e po en ially unin ended
consequences, educing conside ably hei e ec i eness. In pa icula , he disclosu es o
egula o y ins i u ions may c owd ou ade s’ e o o ga he independen in o ma ion,
educing he in o ma ion a ailable in he ma ke (Golds ein and Yang, 2019). Fu he -
mo e, ma ke s can o e eac o disclosu es o public in o ma ion, dis o ing he asse
p ices and ansmi ing misleading in o ma ion (Golds ein and Yang, 2019).
One o he issues in he cu en discussion on he e o ms o he inancial a chi ec-
u e is ela ed o he way o disclose public in o ma ion by egula o y ins i u ions (Bank
o England, 2015). The academic li e a u e is qui e ambiguous abou he o e all e ec
o disclosu es. Al hough i is well unde s ood ha in o ma ion disclosu es can inc ease
ma ke e iciency, much has been w i en abou hei po en ial unin ended consequences.
The c owding ou o p i a e in o ma ion p oduc ion and he p omo ion o des abilizing
beau y-con es incen i es, a e wo iden i ied colla e al e ec s o disclosu es. In his e-
spec , ma ke o e - eac ion o public in o ma ion and ade s’ o e - eliance on disclosu es
ha e been in e p e ed as a sou ce o ma ke ins abili y and inancial isk.
E ec s o disclosu es ha e been modeled wi hin coo dina ion amewo ks in he he-
o e ical li e a u e (Mo is and Shin, 2002; Angele os and Pa an, 2004, 2007; Mya and
Wallace, 2011; Colombo e al., 2014) as well as wi hin ma ke amewo ks (Kool e al.,
2011; Golds ein and Yang, 2019). Se e al expe imen al s udies es ed he coo dina ion
models in he labo a o y, gi en he impossibili y o moni o ing in o ma ion held by eco-
nomic agen s in he eal wo ld. Thei main conclusions suppo he heo e ical conjec-
1
u es ha decision make s o e - ely on public in o ma ion (Co nand and Heinemann,
2014; Shapi o e al., 2014). Ne e heless, ela i ely li le expe imen al esea ch has been
conduc ed o analyze he e ec s o eleasing public in o ma ion in a ma ke en i onmen .
Middeldo p and Rosenk anz (2011) ind e idence ha disclosu es c owd ou p i a e in o -
ma ion and educe p ice in o ma i eness. They gi e suppo o he hypo hesis p oposed
by Kool e al. (2011) ha he educed p ice in o ma i eness is a di ec consequence o
he c owding ou o p i a e in o ma ion. Al e na i ely, Ruiz-Bu o n e al. (2020) show
ha he o e weigh ing o public in o ma ion is he main cause o he educ ion in p ice
in o ma i eness.
The academic li e a u e is no conclusi e abou he desi abili y and he e ec s ha
public disclosu es ha e on an economy wi h endogenous p i a e in o ma ion. In his e-
spec , ou pape con ibu es o he li e a u e by explo ing measu es o educe inancial
isks caused by public elease o in o ma ion in inancial ma ke s. Pa icula ly, we aim
a de ising mo e obus ma ke a chi ec u es ying o make su e ha p i a e in o ma-
ion (and p i a e in o ma ion acquisi ion) is no disca ded by ma ke pa icipan s in he
p esence o public in o ma ion. In ou con ibu ion, we y o shed some ligh on p os
and cons o di e en disclosu e con igu a ions, answe ing o he ques ions: Wha ypes
o disclosu e a e mos bene icial o p omo e ma ke e iciency and educe he inancial
isk ela ed o ma ke o e eac ion? Wha is he e ec o disclosu es in an economy wi h
endogenous p i a e in o ma ion?
Wi h his objec i e, we conduc labo a o y expe imen s using an in o ma ion ma ke
because o i s capabili y o agg ega ing in o ma ion and p o iding accu a e p edic ions on
undamen als (Wol e s and Zi zewi z, 2004; Healy e al., 2010; Page and Clemen, 2013). In
pa icula , we compa e di e en con igu a ions o disclosu es a ying he numbe and he
ela i e p ecision o independen public signals. Ou esul s sugges ha eleasing public
in o ma ion does no necessa ily imp o e ma ke e iciency when ade s ha e access o
cos ly p i a e in o ma ion. The way in o ma ion is disclosed a ec s p i a e in o ma ion
acquisi ion wi h he consequen isk o p ice dis o ions: a single disclosu e c owds ou
in o ma ion acquisi ion while mul iple disclosu es migh inc ease ade s’ in o ma ion
acquisi ion. This e ec on ade s’ in o ma ion acquisi ion ansla es in o a signi ican
o e - eac ion o disclosu es, ha migh lead o a educ ion in he p ice accu acy. We claim
ha s udying how ma ke eac s o di e en con igu a ions o in o ma ion disclosu e is
undamen al o unde s anding how o egula e inancial ma ke s.
We o ganize he emainde o he pape as ollows. Sec ion 2 b ie ly e iews he ela ed
li e a u e and in oduces he wo king hypo heses. Sec ion 3 desc ibes he expe imen al
design and p ocedu e. Sec ion 4 p esen s he esul s o he impac o he disclosu es on
he in o ma ion and asse ma ke , as well as on p ice accu acy. Sec ion 5 concludes.
2 Theo e ical backg ound and hypo heses
2.1 Theo e ical backg ound
G ossman and S igli z (1980) s a e ha , unde a ional expec a ions, decision make s
ha e nei he incen i es o acqui e cos ly in o ma ion no o ade when p ices a e ully
in o ma i e. Compe i i e ma ke s, he e o e, canno be in o ma ionally e icien i all
in o ma ion is ins an aneously agg ega ed in he p ice, as s a ed in he s ong e sion o
he E icien Ma ke Hypo hesis (EMH). G ossman and S igli z (1980) sol e his pa adox
2
by in oducing some noise in he ma ke . In hei e sion o a noisy a ional expec a ions
equilib ium model (REE), p ices a e jus pa ially in o ma i e, gi ing o ade s incen i es
o acqui e cos ly in o ma ion since hey can eco e hei cos s. In his seminal pape ,
Sunde (1992) es s expe imen ally some consequences o he noisy REE, showing ha
a double auc ion mechanism gene a es an endogenous noise o p e en p ices o ully
e eal in o ma ion. In pa icula , he shows ha he noisy REE can be a ained unde
qui e es ic i e condi ions, such as pe ec in o ma ion and a known numbe o inside s.
Many addi ional expe imen al con ibu ions ha e es ed he consequences o he noisy
REE, showing ha in o ma ion agg ega ion is impe ec , and, as a consequence, ma ke
p ices a e jus pa ial indica o s o he undamen al alue.1Rega ding he expe imen al
analysis o he condi ions enhancing ma ke e iciency, we ocus a en ion on he elease o
public in o ma ion as a possible ins umen o help p ice con e gence o he undamen al
alue. Un il now, e y li le esea ch e o has been de o ed o his impo an issue.
2.2 Hypo heses
We will analyze he impac o eleasing public in o ma ion in a labo a o y asse ma ke .
Ou esea ch ques ion p ima ily ocuses on he s udy o he e ec o di e en disclosu e
scena ios on ma ke pe o mance. In pa icula , we conside al e na i e con igu a ions
when in oducing public signals wi h di e en p ecision in o a inancial ma ke . In an
economy whe e in es o s ha e access o cos ly p i a e in o ma ion, we analyze how in-
c easing he numbe o independen eleased public signals a ec s: (i) ade s’ acquisi ion
o p i a e in o ma ion, (ii) e iciency and accu acy o ma ke p ices, and (iii) ade s’ p o -
i s.
Rega ding he ade s’ e o o ga he in o ma ion, we conjec u e ha i is he ag-
g ega e p ecision o public in o ma ion ha a ec s he ade s’ demand o p i a e in-
o ma ion. No e ha ou conjec u e is a di ec consequence o he EMH, since i is he
o e all in o ma ion a ailable o he ade s, and no he way is eleased, which de e mines
ma ke p ices.
Hypo hesis 1 The demand o p i a e in o ma ion depends on he Bayesian agg ega e
p ecision o public in o ma ion.
In o he wo ds, i is no ele an whe he a gi en agg ega e p ecision o public in-
o ma ion is achie ed by eleasing a single signal o mul iple (independen ) signals o
di e en p ecision. An impo an a e ma h o his hypo hesis implies ha he e is no
ole in he implemen a ion o al e na i e disclosing scena ios when managing he elease
o public in o ma ion.
The economic li e a u e p edic s ha ade s’ demand o in o ma ion s ongly de-
pends on hei ini ial in o ma ion. Se e al heo e ical models s a e ha in o ma ion dis-
closu es educe he p oduc ion o cos ly in o ma ion. See, o example, Diamond (1985);
Kool e al. (2011); Colombo e al. (2014); Han e al. (2016) and Golds ein and Yang
(2019). Mo eo e , ecen expe imen al s udies show ha ade s endowed wi h p ecise
and conclusi e ini ial in o ma ion educe hei e o in ga he ing new in o ma ion (Page
and Siem o h, 2017; Ruiz-Bu o n e al., 2020). We, he e o e, conjec u e ha , when con-
clusi e, he p esence o public in o ma ion has a nega i e e ec on he demand o p i a e
in o ma ion.
1See Sunde (1995) and Plo (2000) o a su ey.
3

Hypo hesis 2 The elease o conclusi e public in o ma ion c owds ou p i a e in o ma-
ion.
Conce ning he ade s’ beha io in he asse ma ke , expe imen al con ibu ions ha e
epea edly ound limi ed agg ega ion o in o ma ion in o p ices, which u ns in o p ice
de ia ions om undamen als. Ne e heless, many con ibu ions show ha mo e in-
o ma ion in he ma ke leads o highe p ice in o ma i eness and p ice accu acy (e.g.
Sunde (1992); Bossae s e al. (2014); Page and Siem o h (2017); Co gne e al. (2018);
Al a ano e al. (2020)). We de ine as p ice in o ma i eness a measu e o he in o ma ional
con en o p ices, while p ice accu acy measu es how well p ices p edic he asse alue.
Unde he EMH, we expec p ices o be close o he undamen als he mo e in o ma ion
is a ailable o he ade s.
Hypo hesis 3 Mo e in o ma ion in he ma ke imp o es p ice accu acy.
Fu he mo e, ollowing he EMH, we conjec u e ha he con igu a ion o disclosu es
do no a ec p ice in o ma i eness, since i should be independen o he quan i y o
quali y o he in o ma ion a ailable o he ade s.
Hypo hesis 4 P ice in o ma i eness does no depend on he con igu a ion o disclosu es.
On he o he hand, ma ke p ices may o e eac o he elease o public in o ma ion
and de ia e signi ican ly om undamen als, as conjec u ed by Allen e al. (2006) and
Golds ein and Yang (2019). An exis ing s and o he heo e ical as well as expe imen al
li e a u e ela es such de ia ions om undamen als o he o e weigh ing o public in-
o ma ion in ma ke p ices due o he o e - eliance o ade s o in o ma ion disclosu es.
While he heo e ical li e a u e coun s wi h se e al con ibu ions, he expe imen al li -
e a u e ega ding he o e weigh ing in labo a o y asse ma ke expe imen s is sca ce.2
Acke e al. (2004) epo he exis ence o ma ke o e eac ion o low p ecise disclo-
su es. In a ecen s udy, Ruiz-Bu o n e al. (2020) ind expe imen al e idence abou he
o e weigh ing o public in o ma ion when disclosed as a single signal. Mo eo e , hey
no e ha his e ec is ein o ced by he c owding ou on he ade s’ demand o p i a e
in o ma ion. Those expe imen s epo a non mono onic ela ion be ween he quan i y o
in o ma ion p esen in he ma ke and he p ice in o ma i eness. We o mula e, he e o e,
an al e na i e hypo hesis o Hypo heses 3 and 4:
Hypo hesis 5 Ma ke p ices o e weigh public in o ma ion, esul ing in a educ ion o
p ice in o ma i eness.
In a noisy REE amewo k, in o med ade s can make p o i s only o co e in o ma-
ion acquisi ion cos s, as p ices only pa ially e eal in o ma ion o unin o med ade s.
We should expec , he e o e, ha in o med ade s can eco e he cos o in o ma ion
acquisi ion, gaining he same ne p o i s as unin o med ade s. Ne e heless, some ex-
pe imen al con ibu ions epo ha unin o med ade s ou pe o m in o med ade s
(Hube e al., 2011; Page and Siem o h, 2017; Ruiz-Bu o n e al., 2020).
Hypo hesis 6 In o med ade s gain he same p o i as unin o med ade s.
2The as majo i y o expe imen al pape s on he o e weigh ing o public in o ma ion a e based on
he seminal pape o Mo is and Shin in a beau y con es amewo k (see, o example, Co nand and
Heinemann, 2014; Shapi o e al., 2014; Bae iswyl and Co nand, 2014, 2016), which is no a ma ke
en i onmen .
4
3 Expe imen al design and p ocedu e
Ou expe imen al se ing is simila o o he con ibu ions om he li e a u e on labo a-
o y inancial ma ke s and p edic ion ma ke s.3We implemen an asse ma ke popula ed
by 15 ade s.4A he beginning o each ma ke , ade s a e endowed wi h 1000 uni s
o expe imen al cu ency (ECU)5and 10 isky asse s, paying a di idend Dwhen ma -
ke close. The asse s ha e a one-ma ke li e. The asse ma ke is implemen ed as a 3
minu e double auc ion whe e ade s can submi bids and asks o di ec ly accep any
o he ade ’s ou s anding o e . E e y bid, ask o ansac ion conce ns only one uni o
he asse , bu e e y ade can handle as many as desi ed as long as he/she has enough
cash o asse s (no sho sale is allowed).
S a es o na u e Be o e he ma ke s a s, he compu e andomly de e mines one o
he wo equip obable s a es on na u e. In one s a e o na u e he di idend is 0 and in
he o he is 10, D∈ {0,10}. The asse alue a he end o he ma ke is equal o he
di idend. A he end o he ma ke , he ealiza ion o he s a e o na u e is e ealed and
asse s pay he co esponding di idend. I is common knowledge o all ade s ha he
wo s a es o na u e a e equip obable.
P i a e In o ma ion Endogenous p i a e in o ma ion is implemen ed in an in o ma-
ion ma ke . A any momen du ing he 3 minu es o he ma ke , ade s can acqui e
impe ec in o ma ion abou he alue o he di idend. They can acqui e (independen )
p i a e signals a he cos o 4 ECU each. Each p i a e signal akes he alue 0 o 10,
si∈ {0,10}, and i is co ec wi h p obabili y p, and inco ec wi h p obabili y q= 1−p.
We e e o pas he p ecision o he signal. We se p= 0.6 in all ea men s.
Disclosu es In ma ke s whe e public in o ma ion is eleased, ade s obse e, a he
beginning o each ma ke , ei he one o wo impe ec public signals abou he alue o
he di idend. We e e o hose public signals as disclosu es. Each public signal akes
he alue 0 o 10, S∈ {0,10}, being co ec wi h p obabili y P, and inco ec wi h
p obabili y Q= 1 −P. I is common knowledge among ade s he alue o Pand he
ealiza ions o he disclosu es. We e e o Pas he p ecision o he public signal.
T ea men s We implemen di e en ea men s a ying he numbe o disclosu es and
hei p ecision. Ou expe imen al se ing allows us o es he consequences ha di e en
disclosu e scena ios ha e on he ade s’ in o ma ion acquisi ion, he e iciency o he
ma ke in agg ega ing in o ma ion in o p ices, and ade s’ p o i s. Table 1 summa izes
he implemen ed ea men s. In he Baseline ea men (B), no public signal is eleased.
In he Single disclosu e ea men (S), one public signal wi h a p ecision o PA= 0.8
is eleased. Then, we implemen h ee addi ional ea men s whe e wo independen
public signals a e eleased: (i) Mul iple Symme ic disclosu es ea men (MS), whe e
wo public signals wi h equal p ecision (PA=PB= 0.66) a e eleased a he beginning o
each ma ke ; (ii) Mul iple Weak Asymme ic Disclosu es ea men (MWA), whe e he
p ecision o he wo public signals is PA= 0.64 and PB= 0.70; (iii) Mul iple S ong
Asymme ic Disclosu es ea men (MSA), whe e he p ecision o he wo public signals
is PA= 0.6 and PB= 0.75.
3See, o example, Acke e al. (2002); Hey and Mo one (2004); Fe i and Mo one (2008); Deck e al.
(2013); Fellne and Theissen (2014); Page and Siem o h (2017) and Halim e al. (2019).
410 ma ke s ou o he 20 ma ke s included in he Single disclosu e ea men a e popula ed by 10
ade s.
5Ea nings, as well as asse alue and di idend, du ing he expe imen s we e designa ed in expe imen al
cu ency uni s (ECU) and con e ed in o ea he end o he session.
5
Table 1: Expe imen al design and pa ame e s. Wi h pwe deno e he p ecision o a
p i a e signal, and PAand PBdeno e he p ecision o each indi idual public signal, when
p esen .
T ea men p PAPB♯o ma ke s
B 0.6 - - 20
S 0.6 0.8 - 20
MS 0.6 0.66 0.66 30
MWA 0.6 0.64 0.70 20
MSA 0.6 0.6 0.75 20
In he ea men s wi h mul iple disclosu es, he join p ecision o he public in o ma-
ion depends on whe he disclosu es a e con e gen o di e gen . We conside disclosu es
as con e gen when he wo public signals poin o he same di idend alue, whe eas we
conside disclosu es as di e gen , i hey poin o opposi e alues o he di idend. Using
he Bayesian in e ence, we compu e he p obabili y ha he asse alue is equal o 10 as
a unc ion o he ealiza ions o he public signals SAand SB:
P (D= 10|SA, SB) = "1 + QA
PASAQB
PBSB#−1
,(1)
whe e P (D= 10|SA, SB) is he p obabili y o he e en D= 10, gi en he ealiza ion
o he public signals SAand SB. The a iables SAand SB ake he alue 1 (−1) i he
public signal sugges s an asse alue equal o 10 (0), while hey ake he alue 0 i he e is
no public signal. Mu a is mu andis, we can compu e he p obabili y o he e en D= 0.
Table 2 shows he p obabili y o he e en D= 10 gi en he ealiza ions o he public
signals SAand SB o each one o he implemen ed ea men s.
Table 2: P obabili y o he e en D= 10 gi en he con igu a ion o he public signals.
PAand PB(SAand SB) deno e he p ecision (con igu a ions) o he disclosu es.
T ea men PAPBSASBP (D= 10|SA, SB)
B 0.6 - 0 0 0.5
MSd0.66 0.66 1 -1 0.5
MWAd0.64 0.70 1 -1 0.57
MSAd0.6 0.75 1 -1 0.66
S 0.8 - 1 0 0.8
MSc0.66 0.66 1 1 0.8
MWAc0.64 0.70 1 1 0.81
MSAc0.6 0.75 1 1 0.82
6
Ou expe imen al se ing allows us o s udy how he di e en con igu a ions o he dis-
closu es a ec he ou come o he in o ma ion ma ke and he asse ma ke . In pa icula ,
we can e alua e whe he he ela i e p ecision o he wo public signals, in he mul iple
disclosu e ea men s, a ec s he ma ke pe o mance. In he ea men s wi h asymme -
ic disclosu es, we can s udy di e en s ylized scena ios whe e he disclosu es eleased by
egula o y ins i u ions ha e di e en ela i e p ecision. In he Mul iple S ong Asymme -
ic ea men , in pa icula , we can analyze he e ec o eleasing one “dominan ” public
signal on ma ke pe o mance. Such s ylized scena ios desc ibe he p ominen ole o a
leading egula o y ins i u ion eleasing public in o ma ion, like cen al banks, pa allel o
o he sou ces eleasing noisie in o ma ion. No e ha he less p ecise signal has he same
p ecision as a single p i a e signal. Con a y o a single disclosu e, mul iple disclosu es
migh be di e gen , allowing us o s udy he eac ion o he ma ke o wo opposi e public
o ecas s abou he alue o he di idend.
Conside ing he Baseline ea men as a benchma k, we can s udy he impac o a
single disclosu e o a gi en p ecision (B s S). Fu he mo e, we can con on hose esul s
wi h he disclosu e o wo con e gen signals o equal o di e en p ecision (S s MSc,
S s MWAcand S s MSAc). No e ha in ea men s S, MSAc,MW Acand MSc,
he o e all p ecision o public in o ma ion is app oxima ely in a ian and close o 0.8
(see Table 2). Addi ionally, a di ec compa ison can be pe o med be ween MSdand B
ea men s.
P ocedu e The expe imen was conduc ed a he Labo a o i d’Economia Expe imen-
al a Uni e si y Jaume I in Cas ell´on. We ec ui ed 159 unde g adua e s uden s om
Economics, Finance, and Business Adminis a ion in a leas hei second yea o s udy.
Each subjec only pa icipa ed in one session ha consis ed o 10 ma ke s. When sub-
jec s a i ed a he labo a o y, ins uc ions we e dis ibu ed and explained aloud. This
was ollowed by one p ac ice ma ke o subjec s o ge amilia wi h he so wa e, which
was p og ammed using he Z-T ee so wa e (Fischbache , 2007).6A e each ma ke , di -
idends we e paid ou and subjec s’ p o i was compu ed as he di e ence be ween hei
ini ial money endowmen and he money held a he end o he ading pe iod. Each
subjec ’s inal payo was compu ed as he accumula ed p o i in all ma ke s, and paid
cash a he end o he session.7
4 Resul s
4.1 Summa y s a is ics
Table 3 p o ides summa y s a is ics o he ac i i y in he in o ma ion and asse ma ke s in
he i e ea men s. On a e age, ade s acqui e be ween 0.83 and 2.26 signals depending
on he ea men . The ewes acquisi ions co espond o ea men S, whe e he public
in o ma ion is eleased as a single disclosu e, while he mos acquisi ions a e in MS ea -
men , whe e wo independen and equally p ecise public signals a e eleased. Mo eo e ,
we obse e ha , depending on he ea men , be ween 42% and 67% o ade s acqui e
6T ansla ed ins uc ions and sc eensho s in he implemen ed ea men s a e a ailable upon eques .
7One expe imen al cu ency uni is equi alen o 2 cen s o e. The a e age payo was abou 20e
and each session las ed a ound 90 minu es. No e ha subjec s could make losses. To a oid some o
he p oblems associa ed wi h subjec s making eal losses in expe imen s, we endowed all subjec s wi h
a pa icipa ion ee o 5e. No subjec ea ned a nega i e inal payo in any session.
7
0
,2
,4
,6
,8
1
Di e gen
PI PP
Con e gen
PP75 PP60
MSA
Figu e 5: FR and public benchma ks in
T ea men MSA.
0
,2
,4
,6
,8
1
Di e gen
PI PP
Con e gen
PP70 PP64
MWA
Figu e 6: FR and public benchma ks in
T ea men MWA.
In igu es 7 and 8, we compa e he ime e olu ion o he p ice accu acy wi h he equi alen
ime e olu ion o he ma ke in o ma i eness pe ea men . The la e is compu ed as:
EMIj=1
T
jT
X
=1+(j−1)T
|FR −D|
10 j= 1,2, ..6.(7)
1 2 3 4 5
A g.MI
B S MSA MWA MS
T ea men
30 sec 60 sec 90 sec 120 sec 150 sec 180 sec
Figu e 7: Time e olu ion o ma ke in o -
ma i eness pe ea men . Ve ical ba s de-
pic 95% con idence in e al.
4 4.5 5 5.5
A g.DP
B S MSA MWA MS
T ea men
30 sec 60 sec 90 sec 120 sec 150 sec 180 sec
Figu e 8: Time e olu ion o p ice accu acy
pe ea men . Ve ical ba s depic 95%
con idence in e al.
A i s glance o bo h igu es shows a close co espondence be ween he quan i y o
in o ma ion p esen in he ma ke and he p ice accu acy. The e o e, we canno ejec
Hypo hesis 3, which s a es ha a highe p ice accu acy is achie ed wi h mo e in o ma-
ion in o he ma ke . In pa icula , we obse e ha he lowes le el o p ice accu acy is
achie ed in he Single disclosu e ea men , which is cha ac e ized by he highes magni-
ude o he c owding-ou e ec (see Figu e 1) and he s onges o e weigh ing e ec (see
Figu e 4). Bo h a e de imen al e ec s o p ice accu acy. The ea men s wi h mul iple
disclosu es exhibi he highes le el o p ice accu acy, signi ican ly highe han he Single
disclosu e and he B ea men , when we con on he co esponding alues in he las
ime subin e al (see he e o ba s in Figu e 8).
14

Resul 8 Mul iple disclosu es d i e ma ke p ices sys ema ically close o he undamen-
als han he single disclosu e.
In e es ingly, he ea men s wi h mul iple disclosu es exhibi s simila le els o p ice
accu acy, despi e a signi ican ly di e en le el o ma ke in o ma i eness (see he e o
ba s in Figu e 7). In pa icula , he c owding-in e ec we obse e in he MS ea men
does no ansla e in o a highe le el o p ice accu acy. Gi en he simila le el o p ice
accu acy o he ea men s wi h mul iple disclosu es, he MSA disclosu e gua an ees he
lowes p i a e in o ma ion ga he ing e o .
Resul 9 Gi en he p ice accu acy, he s ong asymme ic mul iple disclosu e is cha ac-
e ized by he lowes agg ega e cos in in o ma ion acquisi ion.
4.4 P o i s
Hypo hesis 6 s a es ha , in a noisy REE amewo k, in o med ade s can gene a e
a su icien le el o g oss p o i s o eco e he cos o in o ma ion acquisi ion. As a
consequence, we should obse e no signi ican di e ence in ade s’ ne p o i be ween
in o med and unin o med ade s.
The ne p o i o ade iin ma ke mis gi en by:
Ne P o i mi = (C180
mi −C0) + Dm·Asse s180
mi ,(8)
whe e (C180
mi −C0) is he cash held a he end o he ma ke a e paying back he ini ial
endowmen C0, and Asse s180
mi deno es he numbe o asse s held a he end o he ma ke .
Dmdeno es he alue o he asse a he end o ma ke m.
Table 4 epo s he mean and s anda d de ia ion o g oss and ne p o i s o in o med
and unin o med ade s in each ea men depending on he con igu a ions o disclosu es.
On a e age, in o med ade s gain sys ema ically highe g oss p o i s han unin o med
ade s, excep in MSd. Howe e , a e accoun ing in o ma ion cos s, hey gain lowe
ne p o i s han unin o med ade s in mos o he ea men s. These di e ences sugges
ha he con igu a ion o disclosu es a ec s in o med ade s’ abili y o eco e he cos
o acqui ing in o ma ion h ough ading.
Table 4: G oss p o i and ne p o i pe con igu a ion o disclosu es.
G oss P o i Ne P o i
In o med Unin o med In o med Unin o med
Mean S d. De . Mean S d. De . Mean S d. De . Mean S d. De .
B 54.2 74.1 46.9 77.7 41.7 73.6 46.9 77.7
S 48.1 100.5 37.2 68.92 35.4 97.4 37.2 68.9
MSAc60.5 75.7 35.0 51.1 53.4 75.8 35.0 51.1
MSAd80.4 98.5 57.8 46.8 74.1 97.8 57.8 46.8
MW Ac70.7 70.9 55.9 52.8 59.8 71.0 55.9 52.8
MW Ad35.8 66.5 29.0 53.0 26.6 66.4 29.0 53.0
MSc55.7 92.3 46.5 88.5 42.7 95.0 46.5 88.5
MSd22.4 89.1 23.9 74.9 8.4 91.7 23.9 74.9
15
To p o ide u he insigh s in o he e ec ha he con igu a ion o disclosu es has
on ade s’ ne p o i , we eg ess ade s’ ne p o i conside ing as explana o y a iables
dummies o he di e en con igu a ions o disclosu es, using as a baseline he ade s’
p o i s in ea men B. Table 5 epo s he esul s o he eg essions.
Models I and II show no signi ican di e ences in ne p o i s in mos o he con ig-
u a ions wi h con e gen disclosu es. Con e sely, in ma ke s wi h di e gen disclosu es,
ade s’ gain signi ican ly highe p o i s in case o s ong asymme ic signals, whe eas in
ma ke s wi h weak asymme ic o symme ic disclosu es, ade s p o i s a e signi ican ly
lowe compa ed o ea men B.
In Models III and IV, we eg ess ade s’ ne p o i conside ing as independen a i-
ables he numbe o acqui ed signals. We canno ejec Hypo hesis 6 in he con igu a ions
wi h con e gen disclosu es, since in o med ade s do no gain highe p o i s compa ed
o unin o med ade s. Recall ha ma ke in o ma i eness and p ice accu acy is highe
in he ma ke s wi h con e gen disclosu es and, he e o e, i is mo e di icul o hem o
ou pe o m unin o med ade s gaining highe ne p o i s. Ins ead, in hose ma ke s wi h
di e gen disclosu es, ade s’ p o i di e be ween in o med and unin o med ade s.
In ea men MSAdin o med ade s ou pe o m unin o med ade s gaining ne p o i s,
whe eas in ea men s MWAdand MSdin o med ade s gain lowe ne p o i s compa ed
o unin o med ade s. To explain hese esul s, we conside he di e en le el o ma ke
in o ma i eness in hose ma ke s. In ma ke s wi h di e gen and s ong asymme ic dis-
closu es, a signi ican c owding-ou o p i a e in o ma ion is obse ed. As a consequence,
ma ke in o ma i eness wo sens in hese ma ke s allowing in o med ade s ge an ad an-
age o e unin o med ade s. On he con a y, in ma ke s wi h di e gen and symme ic
disclosu es, a c owding-in in he demand o p i a e in o ma ion is obse ed, inc easing
signi ican ly ma ke in o ma i eness. As a esul , in hese ma ke s ade s’ in es men
on p i a e in o ma ion is oo high o eco e he cos o in o ma ion h ough ading
in he asse ma ke . Consequen ly, in o med ade s gain lowe ne p o i s compa ed o
unin o med ade s.
Resul 10 In ma ke s wi h con e gen disclosu es he e is a endency o obse e he
same ne p o i s o in o med and unin o med ade s.
Resul 11 In ma ke s wi h di e gen disclosu es in o med ade s’ migh ou pe o m un-
in o med ade s i ma ke in o ma i eness is low.
16
Table 5: OLS eg ession esul s o ne p o i s.
Dependen a iable: Ne P o i
(I) (II) (III) (IV)
S -8.333 -8.333
(6.849) (8.377)
MSAc-1.701 -1.701
(6.324) (6.449)
MSAd22.883** 22.883**
(9.067) (9.918)
MWAc13.205** 13.205**
(6.565) (6.608)
MWAd-17.270** -17.270**
(6.880) (6.701)
MSc-0.637 -0.637
(7.337) (9.102)
MSd-31.775*** -31.775***
(7.675) (9.510)
Acq.Signals in B -0.476 -0.476
(1.800) (1.933)
Acq.Signals in S 3.391 3.391
(6.473) (8.064)
Acq.Signals in MSAc3.250 3.250
(4.312) (4.188)
Acq.Signals in MSAd17.043** 17.043**
(7.658) (7.423)
Acq.Signals in MW Ac2.925 2.925
(2.265) (2.251)
Acq.Signals in MW Ad-4.713* -4.713*
(2.798) (2.496)
Acq.Signals in MSc-3.975 -3.975
(2.698) (3.255)
Acq.Signals in MSd-9.004*** -9.004***
(2.478) (1.643)
Cons an 44.733*** 44.733*** 43.139*** 43.139***
(4.388) (5.456) (2.225) (2.664)
Clus e SE No Yes No Yes
Obse a ions 1,590 1,590 1,590 1,590
Clus e s 169 169
R20.031 0.031 0.039 0.039
Robus s anda d e o s in pa en heses. ∗∗∗ p < 0.01; ∗∗ p < 0.05; ∗p < 0.1.
17
5 Conclusion
We expe imen ally in es iga e o wha ex en he new egula ion ela ed o he he dis-
closu e o public in o ma ion in inancial ma ke s, pu in place a e he 2008 inancial
c isis, is e ec i e. I s basic goals a e o enhance he p ecision o public in o ma ion e-
leased by egula o y ins i u ions and o incen i ize he e o o in es o s in ga he ing
al e na i e in o ma ion o de elop hei own sys ems o isk e alua ion and managemen .
The po en ial win-inc ease in he a ailabili y o in o ma ion due o he new egula ion,
om he public and p i a e domains, should, in p inciple, imp o e he s abili y o he
whole inancial sys em, a oiding o e eliance on ew exogenous signals (e.g. c edi a -
ing agencies). To es hose sound ideas, we conduc an asse ma ke expe imen wi h
endogenous p i a e in o ma ion. In o he wo ds, ade s decide how much in o ma ion
o acqui e a a gi en cos and, a he same ime, hey ade a isky asse , whose alue
depends on a andom s a e o na u e. Using his se ing, we compa e di e en disclo-
su e policies, keeping cons an i s p ecision, while a ying he con igu a ion o he public
signals eleased in he ma ke . We ind ha he way public in o ma ion is eleased has
ele an consequences o he ma ke pe o mance: i has a s ong impac on ade s’
demand o p i a e in o ma ion and a ec s signi ican ly p ice e iciency. The heo e ical
a gumen s based on he EMH sugges , on he con a y, ha he di e en con igu a ions
in disclosing public in o ma ion should no a ec nei he he asse ma ke e iciency no
he ade s’ e o in ga he ing p i a e in o ma ion.
Ou esul s show ha single disclosu es c owd ou p i a e in o ma ion, b inging ma -
ke p ices a om undamen als. Con e sely, we ind ha , when he elease o public
in o ma ion is pe o med in mul iple disclosu es, he ade s’ e o o acqui e p i a e in-
o ma ion inc eases. Howe e , mo e in o ma ion in he ma ke does no always ansla e
in o a mo e e icien p ice sys em.
In line wi h he p e ious li e a u e, we ind ha p ices only pa ially agg ega e he
in o ma ion a ailable o he ade s. We u he obse e ha in o ma ion disclosu es
migh ha e unin ended consequences, educing hei e ec i eness and inc easing he i-
nancial isk o pe sis en de ia ions om undamen als. The ma ke o e eac s o a single
disclosu e c ea ing a sou ce o p ice dis o ion. Ins ead, when inc easing he numbe o in-
s i u ions eleasing public in o ma ion, we obse e a educ ion in he o e weigh ing e ec
and an inc ease in p ice accu acy. In hese ma ke s, p ices a e, in ac , good p edic o s
o he ealiza ion o he u u e s a e o na u e, d i ing ma ke s close o undamen als.
We conclude ha , i he objec i e o he egula o is o educe he ou sou cing o he
in es o s’ in o ma ion ga he ing e o (like in he case o he ew c edi a ing agencies),
inc easing he numbe o ins i u ions eleasing public in o ma ion is an e ec i e policy
measu e.
Acknowledgmen
The au ho s a e g a e ul o he unding by he Uni e si a Jaume I unde he p ojec
UJI-B2018-77, and also he unding by he Spanish Minis y o Science, Inno a ion and
Uni e si ies unde he P ojec RTI2018-096927-B-I00.
18
Re e ences
Acke , L. F., Chu ch, B. K., and Zhang, P. (2002). Ma ke beha io in he p esence
o di e gen and impe ec p i a e in o ma ion: expe imen al e idence om canada,
china, and he uni ed s a es. Jou nal o Economic Beha io & O ganiza ion, 47(4):435–
450.
Acke , L. F., Chu ch, B yan, K., and Gille e, A. B. (2004). Immedia e disclosu e o
sec ecy? The elease o in o ma ion in expe imen al asse ma ke s. Financial Ma ke s,
Ins i u ions & Ins umen s, 13(5):219–243.
Al a ano, S., Banal-Es anol, A., Camacho-Cuena, E., Io i, G., and Kapa , B. (2020).
Cen alized s decen alized ma ke s in he labo a o y: The ole o connec i i y. MPRA
Pape 99129, Uni e si y Lib a y o Munich, Ge many.
Allen, F., Mo is, S., and Shin, H. S. (2006). Beau y con es s and i e a ed expec a ions
in asse ma ke s. The Re iew o Financial S udies, 19(3):719–752.
Angele os, G.-M. and Pa an, A. (2004). T anspa ency o in o ma ion and coo dina ion
in economies wi h in es men complemen a i ies. The Ame ican Economic Re iew,
94(2):91–98.
Angele os, G.-M. and Pa an, A. (2007). E icien use o in o ma ion and social alue o
in o ma ion. Econome ica, 75(4):1103–1142.
Bae iswyl, R. and Co nand, C. (2014). Reducing o e eac ion o cen al banks’ dis-
closu es: Theo y and expe imen . Jou nal o he Eu opean Economic Associa ion,
12(4):1087–1126.
Bae iswyl, R. and Co nand, C. (2016). The p edominan ole o signal p ecision in
expe imen al beau y con es s. The BE Jou nal o Theo e ical Economics, 16(1):267–
301.
Baily, M. N., Klein, A., and Scha din, J. (2017). The impac o he Dodd-F ank Ac on
inancial s abili y and economic g ow h. RSF: The Russell Sage Founda ion Jou nal
o he Social Sciences, 3(1):20–47.
Bank o England (2015). One bank esea ch agenda. Discussion pape .
Bossae s, P., F ydman, C., and Ledya d, J. (2014). The speed o in o ma ion e ela ion
and e en ual p ice quali y in ma ke s wi h inside s: Compa ing wo heo ies. Re iew
o Finance, 18(1):1–22.
Colombo, L., Femminis, G., and Pa an, A. (2014). In o ma ion acquisi ion and wel a e.
The Re iew o Economic S udies, 81(4):1438–1483.
Co gne , B., Deck, C., DeSan is, M., and Po e , D. (2018). In o ma ion (non) agg e-
ga ion in ma ke s wi h cos ly signal acquisi ion. Jou nal o Economic Beha io &
O ganiza ion, 154:286–320.
Co nand, C. and Heinemann, F. (2014). Measu ing agen s’ eac ion o p i a e and pub-
lic in o ma ion in games wi h s a egic complemen a i ies. Expe imen al Economics,
17(1):61–77.
19

Deck, C., Lin, S., and Po e , D. (2013). A ec ing policy by manipula ing p edic ion ma -
ke s: Expe imen al e idence. Jou nal o Economic Beha io & O ganiza ion, 85:48–62.
Diamond, D. W. (1985). Op imal elease o in o ma ion by i ms. The Jou nal o Finance,
40(4):1071–1094.
Fellne , G. and Theissen, E. (2014). Sho sale cons ain s, di e gence o opinion and asse
p ices: E idence om he labo a o y. Jou nal o Economic Beha io & O ganiza ion,
101:113–127.
Fe i, G. and Mo one, M. (2008). The e ec o a ing agencies on he d beha iou .
EERI Resea ch Pape Se ies 2008/21, Economics and Econome ics Resea ch Ins i-
u e (EERI).
Fischbache , U. (2007). z- ee: Zu ich oolbox o eady-made economic expe imen s.
Expe imen al Economics, 10(2):171–178.
Golds ein, I., Sap a, H., e al. (2014). Should banks’ s ess es esul s be disclosed? An
analysis o he cos s and bene i s. Founda ions and T ends R
in Finance, 8(1):1–54.
Golds ein, I. and Yang, L. (2019). Good disclosu e, bad disclosu e. Jou nal o Financial
Economics, 131(1):118–138.
G ossman, S. J. and S igli z, J. E. (1980). On he impossibili y o in o ma ionally e icien
ma ke s. The Ame ican Economic Re iew, 70(3):393–408.
Halim, E., Riyan o, Y. E., and Roy, N. (2019). Cos ly in o ma ion acquisi ion, social ne -
wo ks, and asse p ices: Expe imen al e idence. The Jou nal o Finance, 74(4):1975–
2010.
Han, B., Tang, Y., and Yang, L. (2016). Public in o ma ion and unin o med ading:
Implica ions o ma ke liquidi y and p ice e iciency. Jou nal o Economic Theo y,
163:604–643.
Healy, P. J., Lina di, S., Lowe y, J. R., and Ledya d, J. O. (2010). P edic ion ma ke s:
Al e na i e mechanisms o complex en i onmen s wi h ew ade s. Managemen Sci-
ence, 56(11):1977–1996.
Hey, J. D. and Mo one, A. (2004). Do ma ke s d i e ou lemmings o ice e sa? Eco-
nomica, 71(284):637–659.
Hube , J., Ange e , M., and Ki chle , M. (2011). Expe imen al asse ma ke s wi h endoge-
nous choice o cos ly asymme ic in o ma ion. Expe imen al Economics, 14(2):223–240.
Kool, C., Middeldo p, M., and Rosenk anz, S. (2011). Cen al bank anspa ency and
he c owding ou o p i a e in o ma ion in inancial ma ke s. Jou nal o Money, C edi
and Banking, 43(4):765–774.
Middeldo p, M. H. and Rosenk anz, S. (2011). Cen al bank anspa ency and he c owd-
ing ou o p i a e in o ma ion in an expe imen al asse ma ke . FRB o New Yo k S a
Repo , (487).
20
Mo is, S. and Shin, H. S. (2002). Social alue o public in o ma ion. The Ame ican
Economic Re iew, 92(5):1521–1534.
Mya , D. P. and Wallace, C. (2011). Endogenous in o ma ion acquisi ion in coo dina ion
games. The Re iew o Economic S udies, 79(1):340–374.
Page, L. and Clemen, R. T. (2013). Do p edic ion ma ke s p oduce well-calib a ed p ob-
abili y o ecas s? The Economic Jou nal, 123(568):491–513.
Page, L. and Siem o h, C. (2017). An expe imen al analysis o in o ma ion acquisi ion
in p edic ion ma ke s. Games and Economic Beha io , 101:354–378.
Plo , C. R. (2000). Ma ke s as in o ma ion ga he ing ools. Sou he n Economic Jou nal,
pages 2–15.
Ruiz-Bu o n, A., Camacho-Cuena, E., Mo one, A., and Al a ano, S. (2020). O e weigh -
ing o public in o ma ion in inancial ma ke s: A lesson om he lab. MPRA Pape
98472, Uni e si y Lib a y o Munich, Ge many.
Shapi o, D., Shi, X., and Zillan e, A. (2014). Le el-k easoning in a gene alized beau y
con es . Games and Economic Beha io , 86:308–329.
Sunde , S. (1992). Ma ke o in o ma ion: Expe imen al e idence. Econome ica: Jou -
nal o he Econome ic Socie y, 60(3):667–695.
Sunde , S. (1995). Handbook o Expe imen al Economics, chap e Expe imen al asse
ma ke s: A su ey. NJ: P ince on Uni e si y P ess.
Wol e s, J. and Zi zewi z, E. (2004). P edic ion ma ke s. Jou nal o Economic Pe spec-
i es, 18(2):107–126.
A Ma ke p ice accu acy
E e y panel plo s he e olu ion o he p ice accu acy. In o de o ende he ma ke s
wi h di e en di idends compa able, in ac , we plo he p ice accu acy DP =|D−p ice|
ins ead o ansac ion p ices (solid line), and smoo hed p ice accu acy ( hick-solid line).
We also show he ime e olu ion he di e ence be ween di idend and ully e ealing
p ice (do ed line), DFR =|D−F R|, and he di e ence be ween di idend and public
benchma k (dashed line), DPB =|D−PB|. The ho izon al axis shows he ime (in
seconds) a which he ansac ion ook place. The numbe a he cap ion o each panel
iden i ies he ma ke .
21
0 2 4 6 8 100 2 4 6 8 10
0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180
1 2 3 4 5
6 7 8 9 10
DP DFR DPB
Seconds
B (G oup 1)
Figu e 9: T ading ac i i y o e ime in T ea men B (g oup 1).
0 2 4 6 8 100 2 4 6 8 10
0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180
1 2 3 4 5
6 7 8 9 10
DP DFR DPB
Seconds
B (G oup 2)
Figu e 10: T ading ac i i y o e ime in T ea men B (g oup 2).
22
0 2 4 6 8 100 2 4 6 8 10
0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180
1 2 3 4 5
6 7 8 9 10
DP DPB DFR
Seconds
S (G oup 1)
Figu e 11: T ading ac i i y o e ime in T ea men S (g oup 1).
0 2 4 6 8 100 2 4 6 8 10
0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180 0 60 120 180
1 2 3 4 5
6 7 8 9 10
DP DPB DFR
Seconds
S (G oup 2)
Figu e 12: T ading ac i i y o e ime in T ea men S (g oup 2).
23