An international comparison of incentives for earnings management in order to meet analysts forecasts

Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1) (2021) 75-89 REVISTA DE CONTABILIDAD SPANISH ACCOUNTING REVIEW e is as.um.es/ csa An in e na ional compa ison o incen i es o ea nings managemen in o de o mee analys s o ecas s Susana Callao Gas óna, José Ignacio Ja ne Ja neb a, b)Depa men o Accoun ing and Finance, Facul y o Economics and Business, Uni e si y o Za agoza. Za agoza, España. bCo esponding au ho . E-mail add ess: jija ne@uniza .es ARTICLE INFO A icle his o y: Recei ed 10 Janua y 2019 Accep ed 29 Oc obe 2019 A ailable online 1 Janua y 2021 JEL classifica ion: M40 M41 G17 Keywo ds: Ea nings managemen Analys s o ecas s S ock ma ke s Incen i e o manage ea nings In e na ional compa ison ABSTRACT P e ious in es iga ions ha e shown ha when ea nings mee analys s o ecas s, he ma ke eac s posi i ely, and when hese o ecas s a e no eached, he ma ke eac s nega i ely. Fo his eason, o ecas s become goals o bea , and as he li e a u e has e ealed, his c ea es an incen i e o ea nings managemen . The p esen pape ex ends p io esea ch, p o iding e idence ha he magni ude o he ewa d (penal y) o companies mee ing ( ailing o mee ) ea nings o ecas s di e s depending on he ma ke . In addi ion, i e eals ha he s eng h o he incen i e o companies o manage hei ea nings di e s depending on he ma ke in which hey a e lis ed. We compa e six s ock ma ke s and find ha he mos powe ul incen i es a ise when he ewa d o mee ing he o ecas s o he penal y o no doing so is g ea e . ©2021 ASEPUC. Published by EDITUM - Uni e sidad de Mu cia. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). Códigos JEL: M40 M41 G17 Palab as cla e: Manipulación del esul ado P onós icos de los analis as Me cados bu sá iles Incen i os a manipula Compa ación in e nacional Una compa ación in e nacional del incen i o pa a manipula el esul ado con el obje i o de alcanza los p onós icos de los analis as RESUMEN In es igaciones p e ias han e idenciado que cuando los esul ados alcanzan, o supe an, los p onós icos de los analis as, el me cado eacciona posi i amen e, mien as que, cuando dichos p onós icos no son alcanzados, la eacción del me cado es la con a ia. Po ello, los p onós icos se con ie en en me as a ba i y, al como la li e a u a ya ha con as ado, su ge un incen i o pa a la manipulación del esul ado. En es e con ex o, el p esen e abajo a más allá, e idenciando que la eacción posi i a o nega i a es de di e en e magni ud según los me cados y que ello explica que la ue za con la que las emp esas pe ciben el incen i o a manipula los esul ados pa a alcanza dichos p onós icos ambién sea di e en e según el me cado en el que co iza. Compa ando seis me cados, ob enemos que el incen i o es más po en e cuan o mayo es el p emio po cumpli los p onós icos, o la penalización po no hace lo. ©2021 ASEPUC. Publicado po EDITUM - Uni e sidad de Mu cia. Es e es un a ículo Open Access bajo la licencia CC BY-NC-ND (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). h ps://www.doi.o g/10.6018/ csa .357771 ©2021 ASEPUC. Published by EDITUM - Uni e sidad de Mu cia. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc- nd/4.0/). 76 S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 1. In oduc ion Ea nings announcemen s a e e y impo an e en s o in- es o s because, when hey ake place, in es o s can know whe he he company’s ac ual ea nings ma ch analys s’ o e- cas s. Mee ing ( ailing o mee ) hese o ecas s has a pos- i i e (nega i e) impac on he company’s s ock p ices and e u n, as ound by Edmonds, Edmonds, Fu, and Jenkins (2018),Kinney, Bu gs ahle , and Ma in (2002),López and Rees (2002) and Skinne and Sloan (2002), among o he s. This is why o ecas s a e an impo an a ge o each and, like o he a ge s, mee ing o ecas s may be an incen i e o ea nings managemen . The link be ween analys s’ o ecas s and ea nings managemen has been highligh ed in se e al s udies, including hose by Aba banell and Leha y (2003a, 2003b),Bu gs ahle and Eames (2006),Callao and Ja ne (2018),Cheng and Wa field (2005),Das and Zhang (2003), Dechow, Ge, and Sch and (2000),Ma sumo o (2002),Payne and Robb (2000) and Plumme and Mes (2001). We hink ha he impac o mee ing o ailing o mee ea n- ings o ecas s on he s ock e u n— ha is, he magni ude o he ewa d o mee ing o ecas s o he penal y o no doing so—may di e be ween s ock ma ke s and ha he s eng h o he incen i e o ea nings managemen in o de o mee o ecas s may be linked o he magni ude o he ewa d o penal y. These opics ha e no been in es iga ed un il he p esen pape . We conside six indexes ha ep esen six s ock ma ke s ha a e loca ed in di e en geog aphic a eas and ha ha e di e en cha ac e is ics (legal o igin, financial de elopmen , analys s ollowing he companies, accu acy o he o ecas s and g oss domes ic p oduc GDP pe capi a). These aspec s may influence he ma ke s’ esponse o compliance o non- compliance wi h o ecas s as well as fi ms’ a i udes ega d- ing ea nings managemen . The six indexes a e Bo espa (B azil), DAX (Ge many), Dow Jones (US), FTSE (UK), Hang Seng (Hong Kong) and IBEX (Spain). Ou esul s show ha he magni ude o he ma ke eac ion (i.e. he ewa d o punishmen ) di e s be ween ma ke s. A code law sys em, less financial de elopmen and less weal h d i e g ea e ma ke eac ions. In addi ion, he ma ke eac- ion inc eases when he numbe o analys s ollowing a fi m inc eases and when hese analys s o ecas wi h mo e accu - acy. Fu he mo e, ou esul s show ha he magni ude o he ma ke eac ion explains di e ences in he s eng h o he pe cei ed incen i es o companies o pe o m ea nings managemen o mee o ecas s. The g ea e he ewa d o penal y, he s onge he incen i e o ea nings managemen . Tha is, when mee ing o ecas s has a highe impac on he s ock e u n, he e is a highe p obabili y ha a company will manage i s ea nings o mee o ecas s. P e ious s udies ha e highligh ed he impo ance o ea n- ings managemen in lis ed companies (see, o example, Di- che , G aham, Ha ey, & Rajgopal, 2016). Ou s udy con ib- u es o he li e a u e on ea nings managemen and i s impac on he s ock ma ke . Fi s , he esul s o an in e na ional com- pa ison e eal di e ences in he magni ude o he ma ke ’s ewa d o mee ing o ecas s. Second, he esul s show ha he incen i e o manage ea nings in o de o mee o ecas s is no independen o he ma ke in which he company is lis ed. The esul s we ob ained may be use ul o in es o s, audi - o s, analys s and capi al ma ke s in gene al, as hey p o ide in o ma ion abou he p ocess ha is se in mo ion when ana- lys s’ ea nings o ecas s a e published and how his can influ- ence he eliabili y o financial in o ma ion. In es o s and analys s may find ou esul s pa icula ly use ul when in e - p e ing financial in o ma ion o make in es men decisions o c ea e o ecas s and ecommenda ions. The findings in- dica e ha audi o s should be mo e igilan in hei wo k o fi ms lis ed in s ock ma ke s, o which he incen i e o man- age ea nings is g ea e due o he ma ke ’s s onge eac ion when hese fi ms mee o ail o mee o ecas s. The nex sec ion e iews he li e a u e on his issue. Sec- ion 3 desc ibes he sample, and Sec ion 4 explains he me h- odology. In Sec ion 5, we colla e he esul s. Sec ion 6 p esen s ou sensi i i y analysis, and in Sec ion 7, we discuss he ele an conclusions we can d aw om he esul s. 2. Li e a u e e iew Financial analys s’ opinions and ecommenda ions ega d- ing in es o s and hei in es men decisions ha e an impo - an impac on he sha e p ices o lis ed companies. Analys s’ o ecas s a e highly ele an o in es o s, companies and he ma ke in gene al, because, as La án and Rees (1999) s a e, hey ep esen an app oxima ion o he ma ke ’s expec a ions o he company. When a company’s ac ual ea nings a e announced and he ma ke knows whe he i me i s ea nings o ecas , he com- pany’s sha e p ices will change. P e ious s udies (e.g. Ba o , Gi oly, & Hayn, 2002;Edmonds e al., 2018;Kasznik & Mc- Nichols, 2002;Kinney e al., 2002;López & Rees, 2002;Skin- ne & Sloan, 2002) ha e analysed hese ma ke eac ions and ound ha a nega i e adjus men o sha e p ices occu s when ea nings ail o mee o ecas s and a posi i e one occu s when ea nings mee o exceed o ecas s. In absolu e e ms, he penal y o ailing o mee o ecas s is significan ly g ea e han he ewa d o exceeding hem. In addi ion, fi ms ha con inuously mee expec a ions a e alued mo e highly han hose ha only mee hem occasionally (Ba o e al., 2002; Kasznik & McNichols, 2002;López & Rees, 2002). This ma - ke beha iou jus ifies he ac ha ea nings o ecas s a e be- coming goals o companies o mee . Analysis o he e olu ion o ea nings o e ime e ealed a dec easing endency o epo ea nings ha all sligh ly sho o analys s’ o ecas s and an inc easing endency owa ds pos- i i e ea nings de ia ions (Ba ua, Hoon, & Yi, 2019;B own, 2003;López & Rees, 2002). When he e is a goal o be me , such as ea nings o ecas s, he e is an incen i e o manage ea nings when non-managed ea nings do no mee he goal (Caneghem, 2002;Embong & Hosseini, 2018;Ga cía Osma, Gill de Albo noz, & Gisbe , 2005;Niskanen & Keloha ju, 2000)1. P e ious s udies examining he link be ween ea n- ings o ecas s and ea nings managemen ha e p o ided e id- ence ha ea nings a e managed upwa d o mee ea nings o ecas s (Callao & Ja ne, 2018;Dechow e al., 2000;Ma - sumo o, 2002;Payne & Robb, 2000;Zhang, Pe ols, Robinson, & Smi h, 2018). In addi ion, i has been ound ha compan- ies manage ea nings o mee ea nings o ecas s, bu show- ing sligh posi i e ea nings de ia ions (Aba banell & Leha y, 2003b;Cheng & Wa field, 2005). To da e, he li e a u e has examined he impac o mee ing o ailing o mee ea nings o ecas s on he ma ke , including ewa ds and punishmen s o he company. The esul s ha e shown ha mee ing o ecas s is a goal o companies and his leads o an incen i e o ea nings managemen . How- e e , s udies ha e no in es iga ed whe he he magni ude o 1The incen i es o manage ea nings ha ha e been s udied in he li e - a u e include, among o he s, hose ela ed o deb con ac s, managemen emune a ion, ax e ec s and ea nings benchma ks (posi i e ea nings, pos- i i e ea nings a ia ion, ea nings o ecas s). S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 77 he ewa d (penal y) o mee ing ( ailing o mee ) o ecas s di e s be ween ma ke s o , i so, he implica ions o hese di - e ences on he p obabili y o eeling an incen i e o manage ea nings. The p esen wo k aims o help fill his gap. The ma ke s unde s udy a e loca ed in di e en geo- g aphic a eas, in coun ies wi h di e en legal sys ems, le els o financial de elopmen , weal h, di e en numbe s o ana- lys s ollowing companies and di e en manne s o wo king. A p io i, i is expec ed ha compa ison o ma ke s wi h di - e en cha ac e is ics will e eal di e ences in he magni ude o ewa ds/penal ies be ween ma ke s. I his is he case, we hink ha companies a e mo e likely o eel an incen i e o manage ea nings in o de o mee o ecas s when hey ade in ma ke s in which he magni ude o he ewa d o mee ing hese o ecas s, o he penal y o no doing so, is highe . Based on his, we o mula ed and es ed he ollowing hy- po heses, s a ed in he al e na e o m: H1:The e a e significan di e ences be ween ma ke s in he magni ude o he ewa d (penal y) o mee ing ( ailing o mee ) ea nings o ecas s. H2:The magni ude o he ewa d (penal y) o e ed by he ma ke o mee ing ( ailing o mee ) ea nings o ecas s influ- ences he p obabili y ha companies will pe cei e an incen - i e o manage ea nings. 3. Sample To es H1and H2, we c ea ed a sample composed o com- panies lis ed in six s ock indexes: Bo espa, DAX, Dow Jones, FTSE, Hang Seng and IBEX. In o al, we iden ified 306 com- panies2, bu a e emo al o 82 financial companies, he final sample was comp ised o 224 companies. The pe iod o ana- lysis co e s he yea s om 2006 o 2015, p o iding us wi h 2,240 obse a ions. The ma ke dis ibu ion o hese obse - a ions is p esen ed in Table 1. Table 1 Sample Wo king wi h hese indexes, we will be able o compa e ma ke s ha a e loca ed in di e en geog aphic a eas and ha ha e di e en cha ac e is ics, which may influence hei esponse o compliance o non-compliance wi h o ecas s as well as fi ms’ a i udes ega ding ea nings managemen . Spe- cifically, we conside ed he ollowing aspec s: legal o igin (common law o code law), le el o financial de elopmen (low, medium, high), numbe o analys s ollowing he com- panies, accu acy o o ecas s and GDP pe capi a. Table 2 shows hese aspec s o he di e en coun ies and ma ke s. As shown in he able, we included h ee code law coun- ies (Ge many, Spain and B azil), which ha e a g ea e need 2Six y-one companies om he Bo espa Index, 30 om he DAX, 30 om he Dow Jones, 100 om he FTSE, 50 om he Hang Seng Index and 35 om he IBEX. Table 2 Cha ac e is ics o he ma ke s and coun ies 1Taken om Djanko e al. (2007). Code (Common) indica es a code-law (commom- law) coun y 2As in Degeo ge e al. (2013), he le el o financial de elopmen is measu ed by Finance-Agg ega e (Beck & Le ine, 2002) 3I is he mean o analys s ollowing he companies lis ed in he Index om 2006 o 2015 (da a om I/B/E/S) 4I is he mean o he absolu e di e ence be ween ac ual ea nings and ea nings o ecas scaled by ea nings o ecas o companies lis ed in he Index om 2006 o 2015 (ac ual ea nings om DATSTREAM and ea nings o ecas om I/B/E/S) 5I is he mean o G oss Domes ic P oduc (GDP) pe capi a in $ USA om 2006 o 2015 (da a om Wo ld Bank) o p o ec ion and s a e con ol o e law en o ce s (Siems, 2007). The o he h ee coun ies ( he US, he UK and Hong Kong) ha e common law coun ies in which he legal sys em elies on judges’ decisions a he han s ic codes. We selec ed coun ies wi h di e en le els o financial de- elopmen , measu ed by he finance agg ega e a iable de- eloped by Beck and Le ine (2002) and used by Degeo ge e al. (2013). Specifically, we conside ed coun ies wi h he highes le el o financial de elopmen (US), a medium le el o de elopmen (UK, Ge many and Spain) and a low le el o de elopmen (B azil)3. The numbe o analys s ollowing companies also di e s be ween ma ke s. F om 2006–2015, he a e age anged om 10 analys s in he B azilian ma ke o mo e han 25 in he Ge man ma ke . Fo ecas accu acy is measu ed by he de ia ion o ac ual ea nings om he ea nings o ecas . Thus, lowe de ia ion means highe accu acy. In ou sample, based on he com- panies lis ed in he indexes unde s udy, he ma ke s wi h a highe deg ee o o ecas accu acy a e US and UK, and hose wi h lowe accu acy a e Spain and B azil. Finally, he sample is comp ised o ma ke s loca ed in coun- ies wi h di e en le els o economic de elopmen , meas- u ed by GDP pe capi a. F om 2006–2015, his anged om $11,000 (B azil) o almos $50,000 (US). 4. Me hodology This sec ion desc ibes he me hodology we employ in he s udy. Fi s , we es H1: he e a e significan di e ences be ween ma ke s in he magni ude o he ewa d (penal y) o mee ing ( ailing o mee ) ea nings o ecas s. The a i- ables used o es H1a e defined in Annex 1. To measu e he magni ude o he ewa d (penal y) o mee ing ( ailing o mee ) ea nings o ecas s, we examine he cumula i e abno mal e u n (CAR) a ound he ea nings an- nouncemen da e, when he ea nings de ia ion om he o e- cas is known4. The a iable we use is he mean o he abso- lu e CAR alues o he sha es lis ed in each ma ke in a ±1 window a ound ea nings announcemen da e. We use abso- lu e alues o CAR because we aim o quan i y he ma ke ’s 3Hong Kong is no included in Beck and Le ine’s (2002) a iable. 4Cumula i e abno mal e u n has been used in p e ious esea ch o measu e he impac o an e en on he s ock e u n (see Ba o e al., 2002; Kinney e al., 2002;López & Rees, 2002;Skinne & Sloan, 2002). 78 S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 ewa d (in which case CAR would ha e a posi i e alue) o penal y (in which case CAR would ha e a nega i e alue). Annex 2 shows how he (mean|CAR|) a iable is compu ed. The da a a e ob ained om Da as eam (ac ual da a) and I/B/E/S ( o ecas ed da a). To es H1, we use he K uskal-Wallis non-pa ame ic es (K uskal & Wallis, 1952) since he esul s o he Kolmogo o - Smi no and Shapi o-Wilk es s (Kolmogo o , 1933;Sha- pi o & Wilk, 1965;Smi no , 1939) showed ha he a iable mean|CAR|does no ollow a no mal dis ibu ion (Table 3). We also es he significance o di e ences in his a iable o ma ke pai s using he non-pa ame ic Mann-Whi ney es (Mann & Whi ney, 1947). Table 3 Resul s om No mali y es s │ │ mean|CAR|is defined as he mean o he absolu e alues o cumula i e abno mal e- u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in he indexes unde s udy. I measu es he ewa d (o penal y) o mee ing (o ailing) o mee ea nings o ecas s. To explain di e ences in he ma ke s’ eac ions o com- panies mee ing o no mee ing ea nings o ecas s, we pe - o m a eg ession (equa ion 1) in which mean|CAR|is he dependen a iable and he explana o y a iables a e hose conside ed ele an o he sample selec ion: meanCARj =α0+α1LEGj+α2FINDEV j+α3AFOLLi +α4ACCi +α5lGDPj +ei , (1) whe e: meanCARj is he dependen a iable, which quan ifies he a e age ewa d (penal y) in ma ke jdu ing pe iod o mee ing ( ailing o mee ) ea nings o ecas s. I is defined as he mean o he absolu e CAR alues in a ± 1 window a ound he ea nings announcemen da e o fi ms lis ed in index j in pe iod LEGjis a dicho omous a iable ep esen ing he legal sys- em o coun y j, which akes a alue o 0 o a com- mon law sys em and a alue o 1 o a code law sys- em, in line wi h he classifica ion p oposed by Djanko , McLiesh, and Shlei e (2007). FINDEVjmeasu es he le el o financial de elopmen o coun y jbased on he finance agg ega e a iable de- eloped by Beck and Le ine (2002). AFOLLi is he numbe o analys s ollowing company iin pe iod (da a ex ac ed om I/B/E/S). ACCi is he absolu e di e ence be ween ac ual ea nings and an ea nings o ecas scaled by he ea nings o ecas o company iin pe iod (ac ual da a ex ac ed om Da a- s eam and o ecas ed da a ex ac ed om I/B/E/S). lGDPj is he loga i hm o GDP pe capi a o coun y jin pe iod (da a ob ained om Wo ld Bank). Below, we explain he a iables used in he eg ession as well as he expec ed signs ega ding he ela ionship be ween he independen a iables and he dependen one. Dependen a iable (mean|CARj |): This a iable is defined as he mean o he absolu e alues o he cumula - i e abno mal e u n o he sha es lis ed in a ce ain ma ke in a ±1 window a ound ea nings announcemen da e, when i is known whe he a company’s ea nings me he o ecas s. I measu es he a e age ewa d (penal y) in ma ke jdu ing pe iod o mee ing ( ailing o mee ) ea nings o ecas s. The a iable is calcula ed o each s udied yea acco ding o he p ocess explained in Annex 2. Explana o y a iables: We in oduce fi e explana o y a iables in he model: • Legal o igin (LEG): Code law sys ems ollow he Roman legal adi ion, which is cha ac e ised by ac -finding con- duc ed by s a e-employed judges, au oma ic e iew o de- cisions and eliance on codes a he han judicial disc e- ion. In con as , common law sys ems ely on ac -finding by ju ies, independen judges, in equen appeals and flex- ible codes. In es o s in code law coun ies usually ely on financial e- po ing egula ed by law, and analys s’ o ecas s could be less alued han in common law coun ies. Thus, we could expec a weake ma ke eac ion o mee ing o ailing o mee ea nings o ecas s in code law coun ies; ha is, we could expec he sign o he coe ficien o LEG o be nega - i e. On he o he hand, code law coun ies ea u e less o a s ock ma ke adi ion han common law coun ies and in es o s ha e ewe decision ools, gi ing g ea e alue o accessible ools like analys s’ o ecas s. This would cause he ma ke eac ion o be g ea e and he expec ed sign o he coe ficien o LEG o be posi i e. Consequen ly, we a e no able o p edic he sign o he ela ionship be ween LEG and he dependen a iable. • Financial de elopmen (FINDEV): This a iable is based on he finance agg ega e a iable (Beck & Le ine, 2002). I is a measu e o he deg ee o which na ional financial sys- ems a e used o assess fi ms, moni o manage s, acili a e isk managemen , and mobilise sa ings. This a iable has wo main componen s: finance ac i i y, which is a meas- u e o he o e all ac i i y o he financial in e media ies and ma ke s, and finance size, which is a measu e o he o e all size o he financial sec o . The expec ed sign o he coe ficien o FINDEV is nega i e because a highe le el o financial de elopmen implies ha mo e in o m- a ion is a ailable in he ma ke o make decisions. This makes he ea nings o ecas s ela i ely less ele an o he ma ke , and he e o e he eac ion o mee ing (o ailing o mee ) o ecas s is lowe . • Numbe o analys s ollowing (AFOLL): This a iable is defined as he numbe o analys s ollowing companies in he ma ke . We expec a posi i e ela ionship be ween AFOLL and he dependen a iable. When he e a e mo e analys s ollowing companies, he analys s’ o ecas s a e mo e ele an and eliable o in es o s, and so he sha e p ice eac ion o mee ing o ailing o mee ea nings o e- cas s will be g ea e . • Accu acy (AC C): When he accu acy o o ecas s is highe , he ma ke ’s su p ise a de ia ions om o ecas s is g ea e , causing a s onge ma ke eac ion. Gi en ha we meas- u e his a iable as he de ia ion o ac ual ea nings wi h espec o ea nings o ecas s, lowe alues indica e g ea e accu acy. Thus, i is expec ed ha he sign o he coe ficien o his a iable will be nega i e. • Loga i hm o GDP pe capi a (lGDP): This a iable meas- u es he economic de elopmen o each coun y. The highe he GDP pe capi a, he mo e weal h is a ailable o S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 79 in es o s o make in es men s. In ela i e e ms, he im- pac o posi i e news (compliance wi h o ecas s) o neg- a i e news (non-compliance) will ha e a lowe impac on weal hie in es o s, while in es o s wi h less weal h will be mo e sensi i e o his news and he ma ke eac ion will be highe . Thus, we expec he sign o he coe ficien o lGDP o be nega i e. Finally, we es hypo hesis H2: he magni ude o he e- wa d (penal y) o e ed by he ma ke o mee ing (no mee - ing) ea nings o ecas s influences he p obabili y ha a com- pany will pe cei e an incen i e o manage ea nings. To do so, we pe o m a logis ic eg ession (equa ion 2)in which ITVE is he dependen a iable, mean|CAR|is he ex- plana o y a iable, and he o he a iables a e con ol a i- ables. The a iables used o es H2a e defined in Annex 1. ITVEi =α0+α1meanCARj −1+α2LIQi +α3SOLVi +α4DEBTi +α5ROIi ++α6SIZEi +α7 a GDP +α8 o MKTj +ei (2) whe e: ITVEi is he dependen a iable, which ep esen s he in- cen i e o fi m iin pe iod o manage ea nings upwa d o mee ea nings o ecas s. I s alue is 1 o fi ms wi h he incen i e o manage ea nings and 0 o he wise. I s compu- a ion is p esen ed in Annex 2. meanCARj −1is he explana o y a iable, which quan ifies he a e age ewa d (o penal y) in ma ke jin pe iod −1 o mee ing (o ailing o mee ) ea nings o ecas s. I is defined as he mean absolu e CAR alue in a ±1 window a ound he ea nings announcemen da e o fi ms lis ed in index j in pe iod -1. I s compu a ion is p esen ed in Annex 2. LIQi is he liquidi y a io o fi m iin pe iod , which we define as he quo ien be ween cu en asse s and cu en liabili ies (da a ex ac ed om Da as eam). SOLVi is he sol ency a io o fi m iin pe iod , which we define as he quo ien be ween o al asse s and o al liabil- i ies (da a ex ac ed om Da as eam). DEBTi is he deb a io o fi m iin pe iod , which we define as he quo ien be ween liabili ies and equi y (da a ex ac ed om Da as eam). ROIi is he e u n on in es men o fi m iin pe iod , which we define as he quo ien be ween ope a ing p ofi s and o al asse s (da a ex ac ed om Da as eam). SIZEi measu es he size o fi m iin pe iod , which we define using he asse loga i hm (da a ex ac ed om Da a- s eam). a GDP is he pe cen age a ia ion in GDP be ween yea and −1 in he coun y in which ma ke jis loca ed (da a ob ained om Wo ld Bank) o MKTj is a p oxy o he liquidi y o ma ke jin pe iod , which we define as he quo ien be ween ade olume and ma ke capi alisa ion a he end o yea (da a ex ac ed om Da as eam) Below, we explain he a iables in he eg ession as well as he expec ed signs o he ela ionships be ween independen a iables and he dependen one. Dependen a iable (ITVE): This a iable iden ifies he in- cen i e o upwa dly manage ea nings o mee ea nings o e- cas s. I is a dicho omous a iable whose alue is 1 o fi ms ha ha e he incen i e o manage ea nings upwa ds and hey do so. The alue is 0 o fi ms wi hou he incen i e o man- age ea nings upwa ds o mee o ecas s because hey me hem wi hou ea nings managemen . To assign a alue o 1 o 0 o he a iable, we ha e o know whe he he company has managed ea nings upwa d o no . In o de o do so, we use disc e iona y acc uals (DA), which is he mos common me hodology in he li e a u e o de ec - ing ea nings managemen (Ga cía Osma e al., 2005;McNich- ols, 2000). Acc uals a e defined as he pa o ea nings ha does no in ol e cash flow and he e o e a e mo e likely o be manipula ed by manage s. Howe e , no all acc uals can be managed, so we can dis inguish be ween non-disc e iona y acc uals (NDA), which a e no manipula ed by managemen since hey depend on he economic ci cums ances o he com- pany, and disc e iona y acc uals, which a e subjec o he dis- c e ion o he managemen and he e o e ulne able o being managed5. Since i is possible o obse e o al acc ual (TA), he non- disc e iona y acc uals (NDA) a e es ima ed o subsequen ly calcula e he disc e iona y componen (DA) as he di e - ence be ween he o al acc uals and he es ima ed non- disc e iona y acc uals6. A posi i e alue o DA indica es up- wa ds ea nings managemen , and a nega i e one indica es downwa ds ea nings managemen . We d aw upon Dechow, Sloan, and Sweene (1995) model, as we explain in Annex 2. The desc ip i e s a is ics o DA ob ained om his model a e shown in Table 4. Table 4 Desc ip i e s a is ics o DA (disc e iona y acc uals) om Dechow e al. (1995) model TAi Ai −1 =α1 1 Ai −1 +α2 ΔSALEi −ΔRECi Ai −1 +α3 PPEi Ai −1 +ei ܶܣ௜௧ ܣ௜௧ିଵ ൌߙ ଵ ͳ ܣ௜௧ିଵ ൅ߙ ଶ οܵܣܮܧ௜௧ െοܴܧܥ ௜௧ ܣ௜௧ି ൅ߙ ଷ ܲܲܧ௜௧ ܣ௜௧ିଵ ൅݁ ௜௧ ܦܣ௜௧ ܣ௜௧ିଵ ൌܶܣ௜௧ ܣ௜௧ିଵ െሺܽ ଵ ͳ ܣ௜௧ିଵ ൅ܽ ଶ οܵܣܮܧ௜௧ െοܴܧܥ ௜௧ ܣ௜௧ିଵ ൅ܽ ଷ ܲܲܧ௜௧ ܣ௜௧ିଵ ሻ α α DA (INDEX): Disc e iona y acc uals ob ained om he es ima ion o Dechow e al. (1995) model o he fi ms lis ed in each INDEX DAi Ai −1 =TAi Ai −1 −(a1 1 Ai −1 +a2 (ΔSALEi −ΔRECi ) Ai −1 +a3 PPEi Ai −1 ) DAi is he disc e iona y acc uals o fi m i in pe iod and a1,a2and a3a e he es ima ed alues o pa ame e s α1 o α3 TAi is he o al acc uals o fi m i in pe iod , which has been calcula ed using he di e ence be ween ac ual ea nings (AE) and cash flow om ope a ions (CFO): TAi =AEi −CFO i . ΔSALEi is he change in sales o fi m i in pe iod compa ed wi h -1. ΔRECi is he change in ecei ables o fi m i in pe iod compa ed wi h -1. PPEi is p ope y, plan and equipmen o fi m i in pe iod . Ai −1is he o al asse s figu e o fi m i in pe iod -1 and we ha e used i as a defla o o p e en he e oscedas ici y p oblems Da a a e ex ac ed om DATASTREAM The fi ms ha me hei ea nings o ecas s may (o no ) ha e pe cei ed an incen i e o manage ea nings in o de 5The concep s o disc e iona y and non-disc e iona y acc uals a e ex- plained by Dechow e al. (2010),F ancis e al. (2004) and Ko ha i e al. (2005), among o he s. 6Di e en models ha e been used in he li e a u e o es ima e non- disc e iona y acc uals. A comp ehensi e o e iew o hese models is p esen- ed by Callao e al. (2014). 80 S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 o mee hese o ecas s. Consequen ly, he alue o ITVE is 1 o fi ms who mee ea nings o ecas s when hei non- disc e iona y ea nings (NDE) a e below he ea nings o e- cas (EF) and hei disc e iona y acc uals (DA) a e posi i e (NDE<EF and DA >0) Tha is, his alue is assigned when fi ms ha e an incen i e o manage ea nings upwa ds o each o ecas s and do so. The alue o ITVE is 0 o fi ms who mee ea nings o ecas s when hei non-disc e iona y ea n- ings a e abo e he ea nings o ecas (NDE >EF). In hese cases, fi ms lack an incen i e o manage ea nings upwa ds o mee ea nings o ecas s (no e ha ea nings may be man- aged, bu he incen i e o doing so is no o mee ea nings o ecas s). The p ocess used o assign alues o 1 o 0 o ITVE is explained in Annex 1. Explana o y a iable (mean|CAR|): As indica ed abo e, his a iable is defined as he mean o he absolu e alues o he cumula i e abno mal e u n o he sha es lis ed in a ce ain ma ke in a ±1 window a ound ea nings announce- men da e, which is when we know whe he companies me hei ea nings o ecas s. We in oduce he ewa d (o pen- al y) wi h a delay pe iod, because he manage ’s incen i e o mee analys s’ o ecas s will depend on he eac ion o he ma ke in he pas ( −1). The a iable is calcula ed o each s udied yea acco ding o he p ocess explained in Annex 2. The s a is ical significance o he meanCARj −1 a iable indica es ha he cumula i e abno mal e u n a ound he ea nings announcemen da e, as a measu e o he ewa d (o penal y) he fi m ecei es o mee ing (o ailing o mee ) ea nings o ecas s, is ela ed o a fi m’s mo i a ion o man- age i s ea nings o mee hese o ecas s. The expec ed coe - ficien sign is posi i e since he g ea e he ewa d (penal y) o mee ing ( ailing o mee ) o ecas s, he mo e likely he fi m will pe cei e an incen i e o manage ea nings. Con ol a iables: We in oduce se en con ol a iables in he model: • Economic-financial a ios measu ing he liquidi y (LIQ), sol ency (SOLV), indeb edness (DEBT) and e u n on in- es men (ROI) o he fi ms: Since he economic and finan- cial posi ion o fi ms influences hei ea nings managemen (Bikky & Picheng, 2002;Cha i ou e al., 2012;DeFond & Jiambal o, 1994;Ia idis & Kado inis, 2009;Rosne , 2003; Sweeney, 1994), we expec ha a wo se economic and fin- ancial si ua ion (lowe liquidi y, sol ency and p ofi abili y and highe indeb edness) will be associa ed wi h g ea e incen i es o manage ea nings. So, he expec ed sign o LIQ,SOLV and ROI is nega i e and ha o DEBT is pos- i i e. • Fi m size (SIZE): Many p io s udies ha e analysed he ela ionship be ween fi m size and ea nings managemen , p oposing di e se conclusions (e.g. Ba on & Simko, 2002; Bu gs ahle & Diche , 1997;Kim e al., 2003;Llukani, 2013;Swas ika, 2013). In ou case, we ha e o conside ha la ge companies a e mo e isible in he ma ke and a e usually ollowed by mo e analys s. This may inc ease companies’ ea o no mee ing o ecas s and hus inc ease he p obabili y ha hey will manage ea nings. Hence, he expec ed sign is posi i e. • Va ia ion o GDP ( a GDP): Th ough his a iable, we con ol o he e ec o changes in economic si ua ion on he incen i e o manage ea nings. We expec o find a neg- a i e ela ionship, indica ing ha good e olu ion o GDP may limi he incen i e o manage ea nings (Chih, Shen, & Kang, 2007, and Shen & Chih, 2005, poin ou ha hose iche coun ies a e gene ally less likely o manage ea n- ings). • Ma ke liquidi y ( o MKT): This a iable ep esen s he deg ee o ease o di ficul y o finding a coun e pa , buye o selle o a sha e. Ascioglu, Hegde, K ishnan, and Mc- De mo (2012) sugges ha fi ms exhibi ing g ea e ea n- ings managemen a e associa ed wi h lowe ma ke liquid- i y. Howe e , Huang, Lao, and McPhee (2017) show ha s ock liquidi y inc eases acc ual-based ea nings manage- men . Thus, we canno p edic he sign o he ela ionship be ween o MKT and he incen i e o manage ea nings. The desc ip i e s a is ics o he a iables in eg ession (2) a e p esen ed in Table 5, and he co ela ions be ween hese a iables a e shown in Table 6. As can be seen, he co ela- ions be ween he a iables a e low o mode a e and he signs o he co ela ions a e as expec ed. The nega i e sign o he co ela ion be ween mean|CAR|and a GDP may be su p is- ing. This indica es ha when he economy ge s wo se, he e- wa d (penal y) o companies ha mee ( ail o mee ) hei ea nings o ecas s is highe . In es o s hink ha when he economy is imp o ing, companies a e able o mee ea nings o ecas s mo e easily. We mus conside ha CAR is meas- u ed in a ±1 window a ound he ea nings announcemen . Thus, CAR does no measu e he e olu ion o s ock p ices, bu he eac ion o a pa icula e en wi hin a sho pe iod. The nega i e ela ionship be ween a GDP and o MKT may also d aw a en ion ap io i, bu his is no su p ising i we conside ha o MKT is he quo ien be ween ade olume and ma ke capi alisa ion. In ising pe iods, bo h ade olume and capi alisa ion can inc ease, bu i capi al- isa ion inc eases p opo ionally mo e han ade olume (be- cause, o example, in es men s a e mo e s able), he quo- ien will dec ease. In he o he hand, in ising pe iods no always he olume ade inc eases because some imes he in- es men s a e mo e s able. Table 5 Desc ip i e s a is ics o a iables in equa ion 2 │ │ L IQi is he liquidi y a io o i m i in pe iod , which we de ine as he quo ien be ween asse s and cu en liabili ies (da a om DATASTREAM). SOLVi is he sol ency a io o i m i in pe iod , which we de ine as he quo ien be we e asse s and o al liabili ies (da a om DATASTREAM). D EBTi is he deb a io o i m i in pe iod , which we de ine as he quo ien be ween he lia b equi y (da a om DATASTREAM). ROIi is he e u n-on-in es men o i m i in pe iod , which we de ine as he quo ien b e ope a ing p o i s and o al asse s (da a om DATASTREAM). SIZEi measu es he size o i m i in pe iod , and we de ine i using he asse loga i hm ( DATASTREAM). a GDPj is he pe cen age a ia ion o G oss Domes ic P oduc (GDP) be ween yea an d coun y in which ma ke j is loca ed (da a om Wo l Bank) o MKTj is a p oxy o he liquidi y o ma ke j in pe iod , which is de ined as he quo ien b e ade olume and he ma ke capi aliza ion a he end o yea (da a om DATASTREAM). meanCARj −1is he explana o y a iable quan i ying he a e age ewa d (o pen- al y) in ma ke j in pe iod -1 o mee ing (o ailing) o mee ea nings o ecas s. I is defined as he mean o he absolu e alues o cumula i e abno mal e u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in index j in pe iod -1. See compu a ion in Annex 2. LIQi is he liquidi y a io o fi m i in pe iod , which we define as he quo ien be ween he cu en asse s and cu en liabili ies (da a om DATASTREAM). SOLVi is he sol ency a io o fi m i in pe iod , which we define as he quo ien be ween he o al asse s and o al liabili ies (da a om DATASTREAM). DEBTi is he deb a io o fi m i in pe iod , which we define as he quo ien be ween he liabili ies and equi y (da a om DATASTREAM). ROIi is he e u n-on-in es men o fi m i in pe iod , which we define as he quo ien be ween he ope a ing p ofi s and o al asse s (da a om DATASTREAM). SIZEi measu es he size o fi m i in pe iod , and we define i using he asse loga i hm (da a om DATASTREAM). a GDPj is he pe cen age a ia ion o G oss Domes ic P oduc (GDP) be ween yea and -1 in he coun y in which ma ke j is loca ed (da a om Wo l Bank) o MKTj is a p oxy o he liquidi y o ma ke j in pe iod , which is defined as he quo ien be ween he ade olume and he ma ke capi aliza ion a he end o yea (da a om DATASTREAM). S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 81 Table 6 Pea son Co ela ions ( a iables in equa ion 2) │ │ ∗Significan a 0.05 ∗∗ Significan a 0.01 meanCARj −1is he explana o y a iable quan i ying he a e age ewa d (o pen- al y) in ma ke j in pe iod -1 o mee ing (o ailing) o mee ea nings o ecas s. I is defined as he mean o he absolu e alues o cumula i e abno mal e u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in index j in pe iod -1. See compu a ion in Annex 2. LIQi is he liquidi y a io o fi m i in pe iod , which we define as he quo ien be ween he cu en asse s and cu en liabili ies (da a om DATASTREAM). SOLVi is he sol ency a io o fi m i in pe iod , which we define as he quo ien be ween he o al asse s and o al liabili ies (da a om DATASTREAM). DEBTi is he deb a io o fi m i in pe iod , which we define as he quo ien be ween he liabili ies and equi y (da a om DATASTREAM). ROIi is he e u n-on-in es men o fi m i in pe iod , which we define as he quo ien be ween he ope a ing p ofi s and o al asse s (da a om DATASTREAM). SIZEi measu es he size o fi m i in pe iod , and we define i using he asse loga i hm (da a om DATASTREAM). a GDPj is he pe cen age a ia ion o G oss Domes ic P oduc (GDP) be ween yea and -1 in he coun y in which ma ke j is loca ed (da a om Wo ld Bank) o MKTi is a p oxy o he liquidi y o ma ke j in pe iod , which is defined as he quo ien be ween he ade olume and he ma ke capi aliza ion a he end o yea (da a om DATASTREAM). 5. Resul s This pape in es iga es whe he he ewa d (penal y) o mee ing ( ailing o mee ) analys s’ o ecas s di e s be ween ma ke s (H1) and whe he he ma ke ’s eac ion influences he incen i e o companies o manage ea nings o mee o e- cas s (H2). We find significan di e ences in he ewa d (penal y) o mee ing ( ailing o mee ) ea nings o ecas s be ween he ma - ke s. The esul s o he non-pa ame ic K uskal-Wallis es o k independen samples (Table 7) highligh significan di e - ences in mean|CAR|ac oss he six indices we analysed. In o he wo ds, we can confi m ha he eac ion o sha e p ices o mee ing (o ailing o mee ) an ea nings o ecas is signific- an ly di e en a he 1% le el. P e ious s udies ha e al eady in es iga ed he ma ke ’s eac ion when i is known whe he companies eached hei ea nings o ecas s o no (Ba o e al., 2002;Edmonds e al., 2018;Kasznik & McNichols, 2002; Kinney e al., 2002;López & Rees, 2002;Skinne & Sloan, 2002). Howe e , hese s udies ocused only on one ma ke , mos o en he US ma ke , so we canno compa e ou esul s wi h hei s. Howe e , we can a fi m ha he esul s coincide wi h ou expec a ions, since changes in sha e p ices, which de e mine CAR, depend on many ac o s ha di e be ween he ma ke s we s udy. The anks gene a ed by he K uskal-Wallis es (Table 8) show ha mee ing (o ailing o mee ) o ecas s has he mos impac on CAR in he B i ish ma ke , ollowed by he Ge - man, Ame ican, B azilian, Spanish and Hong Kong ma ke s. In o he wo ds, he eac ion o sha e p ices o ea nings an- nouncemen s, which p o ide in o ma ion abou whe he ana- lys s’ o ecas s we e me o no , is s onges in he B i ish and Ge man ma ke s and weakes in he Spanish and Hong Kong ma ke s. Fu he mo e, he esul s o he Mann-Whi ney es , which was conduc ed o compa e he eac ion o each pai o ma ke s (Table 9), indica e ha , in gene al, each ma ke Table 7 K uskal-Wallis es esul s │ │ mean|CAR|is defined as he mean o he absolu e alues o cumula i e abno mal e- u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in he indexes unde s udy. I measu es he ewa d (o penal y) o mee ing (o ailing) o mee ea nings o ecas s. See compu a ion in Annex 2. Table 8 K uskal-Wallis es anks Table 8. K uskal-Wallis es anks INDEX N A e age ank BOVESPA 440 646.83 mean | CAR | DAX 240 1330.06 DOW JONES 250 626.78 FTSE 770 1602.44 HANG-SENG 280 470.50 IBEX 260 511.17 TTo al 22,240 mean Ň CAR Ňis de ined as he mean o he absolu e alues o cumula i e abno mal e u n in a a ound he ea nings announcemen da e o i ms lis ed in he indexes unde s udy. I measu (o penal y) o mee ing (o ailing) o mee ea nings o ecas s. See compu a ion in Annex 2. mean|CAR|is defined as he mean o he absolu e alues o cumula i e abno mal e- u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in he indexes unde s udy. I measu es he ewa d (o penal y) o mee ing (o ailing) o mee ea nings o ecas s. See compu a ion in Annex 2. Table 9 Mann-Whi ney es esul s Table 9: Mann-Whi ney es esul s mmean | CAR | DAX DOW JONES FTSE HANG SENG IBEX Mann-Whi ney U 0.000 35200 0.000 34496 34320 BOVESA Wilcoxon W 78606 60625 78606 66374 61815 Z -20.500 -4.359 -27.535 -6.641 -5.452 Asymp o ic sig. 0.000 0.000 0.000 0.000 0.000 Mann-Whi ney U 4800 34416 0.000 1872 DAX Wilcoxon W 30225 54852 31878 29367 Z -14.597 -12.926 -18.689 -17.004 Asymp o ic sig. 0.000 0.000 0.000 0.000 Mann-Whi ney U 15400 23800 25350 DOW JONES Wilcoxon W 40825 55678 52845 Z -18.155 -3.032 -0.687 Asymp o ic sig. 0.000 0.002 0.492 Mann-Whi ney U 0.000 0.000 FTSE Wilcoxon W 31878 27495 Z -23.594 -22.958 Asymp o ic sig. 0.000 0.000 Mann-Whi ney U 283920 HANG SENG Wilcoxon W 60270 Z -0.707 Asymp o ic sig. 0.480 mean Ň CAR Ňis de ined as he mean o he absolu e alues o cumula i e abno mal e u n in a -1+1 win d a ound he ea nings announcemen da e o i ms lis ed in he indexes unde s udy. I measu es he e w (o penal y) o mee ing (o ailing) o mee ea nings o ecas s. See compu a ion in Annex 2. mean|CAR|is defined as he mean o he absolu e alues o cumula i e abno mal e- u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in he indexes unde s udy. I measu es he ewa d (o penal y) o mee ing (o ailing) o mee ea nings o ecas s. See compu a ion in Annex 2. eac s in a significan ly di e en way o he o he ma ke s. To iden i y which cha ac e is ics o he ma ke s and coun- ies may explain di e ences in ma ke eac ions, we pe - o med eg ession (1). The esul s a e shown in Table 10. The able shows ha all he a iables a e significan a he 1% le el, excep AC C, which is significan a 10%. The pos- i i e sign o he coe ficien o LEG indica es ha he eac- ion o a company’s sha e p ices o mee ing o ailing o mee an ea nings o ecas is highe in s ock ma ke s in coun ies wi h a Roman Ge manic legal adi ion (i.e. a code law sys- em). The ac ha hese coun ies ha e less o a s ock ma ke adi ion han hose based on Anglo-Saxon legal adi ions 82 S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 Table 10 Linea eg ession esul s (equa ion 1) meanCARj =α0+α1LEGj+α2FINDEVj+α3AFOL Li +α4AC Ci +α5lGDPj +ei T Table 10. Linea eg ession esul s (equa ion 1) mean Ň CAR j Ň = α 0 + α 1 LEG j + α 2 FINDEV j + α 3 AFOLL i + α 4 ACC j + α 5 lGDP j + +e i B S anda d e o Sig. LEG 0.014 0.001 0.000 FINDEV -0.020 0.002 0.000 AFOLL 0.000 0.000 0.000 ACC -0.369 0.011 0.084 lGDP -0.133 0.010 0.000 Cons an 0.679 0.046 0.419 a n Ň CAR j Ň is he dependen a iable and quan i ying he a e age ewa d (o penal y) in ke j in pe iod o mee ing (o ailing) o mee ea nings o ecas s. I is de ined as he mean h e absolu e alues o cumula i e abno mal e u n in a -1+1 window a ound he ea nings ouncemen da e o i ms lis ed in index j in pe iod . See compu a ion in Annex 2. G j is a dicho omous a iable ep esen ing he legal sys em o coun y j, aking alue 0 o m mon law and alue 1 o code law, acco ding o he classi ica ion by Djanko e al. (2007). ND EV j measu es he le el o inancial de elopmen o coun y j by he a iable Finance- g ega e aken om Beck and Le ine, (2002). O LL i is he numbe o analys s ollowing he company i in pe iod (da a ex ac ed om / E/S). C i is he absolu e di e ence be ween ac ual ea nings and ea nings o ecas scaled by n ings o ecas o company i in pe iod (ac ual da a ex ac ed om DATASTREAM and e cas ed da a ex ac ed om I/B/E/S). D P j is he loga i hm o GDP pe capi a o coun y j in pe iod (da a om Wo d Bank) meanCARj is he dependen a iable and quan i ying he a e age ewa d (o pen- al y) in ma ke j in pe iod o mee ing (o ailing) o mee ea nings o ecas s. I is defined as he mean o he absolu e alues o cumula i e abno mal e u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in index j in pe iod . See compu a ion in Annex 2. LEGjis a dicho omous a iable ep esen ing he legal sys em o coun y j, aking alue 0 o common law and alue 1 o code law, acco ding o he classifica ion by Djanko e al. (2007). FINDEVjmeasu es he le el o financial de elopmen o coun y j by he a iable Finance-Agg ega e aken om Beck and Le ine, (2002). AFOLLi is he numbe o analys s ollowing he company i in pe iod (da a ex ac ed om I/B/E/S). ACCi is he absolu e di e ence be ween ac ual ea nings and ea nings o ecas scaled by ea nings o ecas o company i in pe iod (ac ual da a ex ac ed om DATA- STREAM and o ecas ed da a ex ac ed om I/B/E/S). lGDPj is he loga i hm o GDP pe capi a o coun y j in pe iod (da a om Wo d Bank) (i.e. hose wi h a common law sys em) may explain his e- la ionship. Analys s’ o ecas s can be conside ed a mo e el- e an sou ce o in o ma ion in coun ies wi h a code law sys- em, whe e he e a e no as many sophis ica ed analy ical elemen s as in mo e de eloped ma ke s. Thus, he ma ke will ha e a s onge eac ion o mee ing o ailing o mee ea nings o ecas s will be s onge . As expec ed, he e is a nega i e ela ionship be ween FINDEV and he ma ke ’s eac ion. The highe le el o in- o ma ion a ailable in mo e de eloped ma ke s means ha analys s’ o ecas s a e no as impo an o in es o s’ decision- making. Fo his eason, he impac o mee ing ea nings o e- cas s on he cumula i e abno mal e u n o sha es will no be as significan as in he ma ke s wi h less a ailabili y o in o m- a ion. AFOLL has a posi i e ela ionship wi h he dependen a i- able. A g ea e numbe o analys s ollowing a company im- plies ha he company is mo e isible o he ma ke . This leads in es o s o ha e a g ea e eac ion o companies ha mee o ail o mee ea nings o ecas s. As expec ed, he coe ficien o ACC is nega i e. In ma - ke s in which o ecas s a e mo e accu a e and he e is lowe ea nings de ia ion om o ecas s (i.e. lowe ACC alues), in- es o s a e accus omed o companies mee ing ea nings o e- cas s. Failu e o mee hese o ecas s will be ecei ed wi h su p ise by hese ma ke s, and hey will eac mo e s ongly han ma ke s mo e accus omed o less accu a e ea nings o e- cas s. Finally, he ma ke eac ion is g ea e in coun ies wi h lowe GDP pe capi a, as in es o s ha e ewe a ailable e- sou ces o make in es men s and he lowe e u n on hei in es men s has a g ea e impac on hei weal h. Thus, mee - ing ( ailing o mee ) ea nings o ecas s, which is conside ed posi i e (nega i e) news by in es o s, will gene a e a g ea e ewa d (penal y) wi hin he ma ke . A e es ablishing ha he e a e significan di e ences in he way ma ke s beha e, we ocused on assessing whe he he impac o mee ing (o ailing o mee ) o ecas s influences fi ms’ pe cei ed incen i e o mee hei o ecas s h ough ea nings managemen . We conduc ed a logis ic eg ession (equa ion 2) o ha pu pose, he esul s o which a e shown in Table 11. The model co ec ly classified 61.7% o cases. Table 11 Logis ic eg ession esul s (equa ion 2) ITVEi =α0+α1meanCARj −1+α2LIQi +α3SOLVi +α4DEBTi +α5ROIi +α6SIZEi +α7 a GDPj +α8 o MKTj +ei Table 11. Logis ic eg ession esul s (equa ion 2) ITVEi = α 0 + α 1mean Ň CARj -1 Ň + α 2LIQi + α 3SOLVi + α 4DEBTi + α 5ROIi + α 6SIZEi + α 7 a GD α 8 o MKTj +ei B S anda d e o Wald d Sig. mean | CAR -1 | 47.868 19.345 6.123 1 0.013 LIQ 0.072 0.158 0.209 1 0.648 SOLV 0.113 0.195 0.333 1 0.564 DEBT -0.121 0.075 2.574 1 0.109 ROI 3.118 1.546 4.064 1 0.044 SIZE 0.015 0.171 0.008 1 0.930 a GDP -0.054 0.026 4.195 1 0.041 o MKT -0.008 0.004 3.088 1 0.079 Cons an -1.146 1.430 0.642 1 0.423 ITVEi is he dependen a iable ep esen ing he incen i e o i m i in pe iod o manage ea upwa d o mee analys s’ o ecas s. I s alue will be 1 o i ms wi h he incen i e o manage ea and 0 o he wise. See compu a ion in Annex 2. mean Ň CARj -1 Ň is he explana o y a iable quan i ying he a e age ewa d (o penal y) in ma k pe iod -1 o mee ing (o ailing) o mee ea nings o ecas s. I is de ined as he mean o he a b alues o cumula i e abno mal e u n in a -1+1 window a ound he ea nings announcemen d a i ms lis ed in index j in pe iod -1. See compu a ion in Annex 2. LIQi is he liquidi y a io o i m i in pe iod , which we de ine as he quo ien be ween he c u asse s and cu en liabili ies (da a om DATASTREAM). SOLVi is he sol ency a io o i m i in pe iod , which we de ine as he quo ien be ween h e asse s and o al liabili ies (da a om DATASTREAM). DEBTi is he deb a io o i m i in pe iod , which we de ine as he quo ien be ween he lia b and equi y (da a om DATASTREAM). ROIi is he e u n-on-in es men o i m i in pe iod , which we de ine as he quo ien be we e ope a ing p o i s and o al asse s (da a om DATASTREAM). SIZEi measu es he size o i m i in pe iod , and we de ine i using he asse loga i hm (da a DATASTREAM). a GDPj is he pe cen age a ia ion o G oss Domes ic P oduc (GDP) be ween yea and -1 coun y in which ma ke j is loca ed (da a om Wo ld Bank). o MKTj is a p oxy o he liquidi y o ma ke j in pe iod , which is de ined as he quo ien be he ade olume and he ma ke capi aliza ion a he end o yea (da a om DATASTREAM) . ITVEi is he dependen a iable ep esen ing he incen i e o fi m i in pe iod o manage ea nings upwa d o mee analys s o ecas s. I s alue will be 1 o fi ms wi h he incen i e o manage ea nings and 0 o he wise. See compu a ion in Annex 2. meanCARj −1is he explana o y a iable quan i ying he a e age ewa d (o pen- al y) in ma ke j in pe iod -1 o mee ing (o ailing) o mee ea nings o ecas s. I is defined as he mean o he absolu e alues o cumula i e abno mal e u n in a -1+1 window a ound he ea nings announcemen da e o fi ms lis ed in index j in pe iod -1. See compu a ion in Annex 2. LIQi is he liquidi y a io o fi m i in pe iod , which we define as he quo ien be ween he cu en asse s and cu en liabili ies (da a om DATASTREAM). SOLVi is he sol ency a io o fi m i in pe iod , which we define as he quo ien be ween he o al asse s and o al liabili ies (da a om DATASTREAM). DEBTi is he deb a io o fi m i in pe iod , which we define as he quo ien be ween he liabili ies and equi y (da a om DATASTREAM). ROIi is he e u n-on-in es men o fi m i in pe iod , which we define as he quo ien be ween he ope a ing p ofi s and o al asse s (da a om DATASTREAM). SIZEi measu es he size o fi m i in pe iod , and we define i using he asse loga i hm (da a om DATASTREAM). a GDPj is he pe cen age a ia ion o G oss Domes ic P oduc (GDP) be ween yea and -1 in he coun y in which ma ke j is loca ed (da a om Wo ld Bank). o MKTj is a p oxy o he liquidi y o ma ke j in pe iod , which is defined as he quo ien be ween he ade olume and he ma ke capi aliza ion a he end o yea (da a om DATASTREAM). The ele an a iable o he p oposed objec i e is mean|CAR|. As we can obse e, i is significan a he 5% le el; in o he wo ds, he impac o mee ing ea nings o e- cas s on he cumula i e abno mal e u n o sha es explains fi ms’ incen i e o manage ea nings o mee analys s’ o e- cas s. As men ioned abo e, p e ious s udies (Ba o e al., 2002;Edmonds e al., 2018;Skinne & Sloan, 2002) ha e examined he posi i e (nega i e) ma ke eac ion when com- panies mee ( ail o mee ) ea nings o ecas s (i.e. he CAR a ound an ea nings announcemen ). Likewise, p e ious s ud- ies (e.g. Callao & Ja ne, 2018;Dechow e al., 2000;Ma - sumo o, 2002;Payne & Robb, 2000) ha e p o en ha he e is a ela ionship be ween ea nings managemen and ea nings o ecas s, and o he in es iga ions (Caneghem, 2002;Ga cía Osma e al., 2005;Niskanen & Keloha ju, 2000) iden ified analys s’ o ecas s as possible incen i es o ea nings man- agemen . The p esen pape p o ides e idence ha he ela ionship be ween ea nings managemen and mee ing ea nings o e- cas s (ITVE) depends on he magni ude o he ma ke eac- ion when companies mee (o ail o mee ) ea nings o e- cas s (CAR a ound ea nings announcemen s). Since we find significan di e ences in he magni ude o he ma ke eac- ion, he esul s sugges ha he incen i e o manage ea n- ings in o de o mee o ecas s depends on he ma ke in S. Callao Gas ón, J.I. Ja ne Ja ne /Re is a de Con abilidad Spanish Accoun ing Re iew 24 (1)(2021) 75-89 83 which a company is lis ed. The posi i e a iable coe ficien means ha he la ge he ma ke ’s ewa d (penal y) o mee ing ( ailing o mee ) ana- lys s’ o ecas s, he mo e likely i is ha companies will pe - cei e an incen i e o manage ea nings. This is he case o fi ms lis ed in he FTSE and DAX indexes. The fi ms lis ed in he Hang Seng and IBEX indexes a e hose ha pe cei e his incen i e o be leas s ong. ROI is he only financial a iable ha is significan (a he 5% le el). This sugges s ha analys s a e mo e demanding wi h he mos p ofi able companies when hey o ecas hei ea nings and he e o e hese fi ms a e mo e likely o manip- ula e hei accoun ing figu es o a ain hese high- e u n a - ge s. The o he financial a ios a e no significan , which means ha he company’s financial posi ion is no e y el- e an o ea nings managemen pe o med o mee ea nings o ecas s, unlike when companies manage ea nings o o he easons. Two o he con ol a iables a e significan : a GDP (a he 5% le el) and o MKT (a he 10% le el). Ou expec a ion o a GDP, which was based on he wo k o Chih e al. (2007) and Shen and Chih (2005), was me . The nega i e coe fi- cien indica es ha he mo e he GDP dec eases and he eco- nomy wo sens, he g ea e he likelihood ha companies will pe cei e an incen i e o manage ea nings in o de o mee o ecas s. A nega i e coe ficien was also ound o o MKT, which indica es ha when he ma ke is mo e liquid, he e is a lowe p obabili y ha companies will pe cei e an incen i e o manage ea nings. Since some s udies sugges a posi i e e- la ionship in his ega d and o he s sugges he opposi e, we did no p edic ed he sign o he coe ficien . Howe e , ou esul s align wi h hose o Ascioglu e al. (2012). 6. Sensi i i y analysis The esul s o ou s udy could be biased by he model used o ob ain DA, he alue o which was used o de e mine he alue o he dependen a iable, ITVE, in eg ession (2). Fo his eason, we ca ied ou a sensi i i y analysis, epea ing ou s udy using La cke and Richa dson’s (2004) model and equa ion (3) o es ima e acc uals: TAi Ai −1 =α1 1 Ai −1 +α2 (ΔSALEi −ΔRECi ) Ai −1 +α3 PPEi Ai −1 +α4B Mi +α5 CFO i Ai −1 +ei , (3) whe e: TAi ep esen s he o al acc uals o fi m *i* in pe iod * *, which we e calcula ed based on he di e ence be ween ac- ual ea nings (AE) and cash flow om ope a ions (CFO): *TA i =*AE* i * - CFO* i *. ΔSALEi ep esen s he change in sales o fi m *i* in pe iod * * compa ed o * *-1. ΔRECi ep esen s he change in ecei ables o fi m *i* in pe iod * * compa ed o * *-1. PPEi ep esen s he p ope y, plan s and equipmen o fi m *i* in pe iod * *. B Mi ep esen s he book o ma ke a io o fi m *i* in pe iod * *. CFO i ep esen s he cash flow om ope a ions o fi m *i* in pe iod * *. Ai −1 ep esen s he o al asse s o fi m *i* in pe iod * *-1, which we used as a defla o o p e en he e oscedas ici y p oblems. ei is he e o e m o fi m *i* in pe iod * *. Fo his eg ession, da a we e ex ac ed om Da as eam. A e es ima ing he pa ame e s o equa ion (3), we used hese alues o p edic he o al acc uals du ing he pe iod o analysis (2006–2015) and o calcula e he p edic ion e o using equa ion (4): DAi Ai −1 =TAi Ai −1 −(a1 1 Ai −1 +a2 (ΔSALEi −ΔRECi ) Ai −1 +a3 PPEi Ai −1 +a4B Mi +a5 CFO i Ai −1 ) (4) whe e DAi ep esen s he disc e iona y acc uals o fi m i in pe iod and a1,a2,a3,a4and a5a e he es ima ed alues o pa ame e s α1−α5. The desc ip i e s a is ics o DA, which we e es ima ed based on La cke and Richa dson’s (2004) model, a e p esen- ed in Table 12. Table 12 Desc ip i e s a is ics o DA (disc e iona y acc uals) om La cke & Richa dson (2004) model TAi Ai −1 =α1 1 Ai −1 +α2 (ΔSALEi −ΔRECi ) Ai −1 +α3 PPEi Ai −1 +α4B Mi +α5 CFO i Ai −1 +ei Table 12. Desc ip i e s a is ics o DA (disc e iona y acc uals) om La cke and Richha dson (2004) model L W LW LW LW LW LW LW LWLW LWLW LW H $ &)2 %W0 $ 33( $ 5(&6$/( $$ 7$  ''   1 54 1 3 1 2 1 1 1 )( 1 DDDDD Va iable Minimum Maximum Mean S anda d de ia ion DA (BOVESPA) -0.1457 0.0138 0.0011 0.0451 DA (DAX) -0.0853 0.0853 0.0005 0.0262 DA (DOW JONES) -0.0632 0.0581 0.0011 0.0187 DA (FTSE) -0.1058 0.1009 0.0003 0.0309 DA (HANG SENG) -0.0978 0.1158 -0.0021 0.0334 DA (IBEX) -0.0852 0.0788 -0.0003 0.0246 DA (INDEX): Disc e iona y acc uals ob ained om he es ima ion o La cke and Richa dson model o he i ms lis ed in each INDEX ) 1 ( 1 54 1 3 1 2 1 1 11   ''  LW LW LW LW LW LW LWLW LWLW LW LW LW $ &)2 D%W0D $ 33( D $ 5(&6$/( D $ D $ 7$ $ '$ DAi is he disc e iona y acc uals o i m i in pe iod a1 , a2 , a3 , a4 and a5 a e he es ima ed alues o pa ame e s α 1 o α 5. TAi is he o al acc uals o i m i in pe iod , which has been calcula ed using he d i be ween ac ual ea nings (AE) and cash low om ope a ions (CFO): TAi = AE i - CFO i ∆ SALEi is he change in sales o i m i in pe iod compa ed wi h -1. ∆ RECi is he change in ecei ables o i m i in pe iod compa ed wi h -1. PPEi is p ope y, plan and equipmen o i m i in pe iod . B Mi is he book o ma ke a io o i m i in pe iod . CFOi is he cash low om ope a ions o i m i in pe iod . Ai -1 is he o al asse s igu e o i m i in pe iod -1 and we ha e used i as a de la o o he e oscedas ici y p oblems. Da a we e ex ac ed om DATASTREAM. DA (INDEX): Disc e iona y acc uals ob ained om he es ima ion o La cke and Richa dson (2004) model o he fi ms lis ed in each INDEX DAi Ai −1 =TAi Ai −1 −a1 1 Ai −1 +a2 (ΔSALEi −ΔRECi ) Ai −1 +a3 PPEi Ai −1 +a4B Mi +a5 CFO i Ai −1 DAi is he disc e iona y acc uals o fi m i in pe iod a1,a2,a3,a4and a5a e he es ima ed alues o pa ame e s α1 o α5. TAi is he o al acc uals o fi m i in pe iod , which has been calcula ed using he di e ence be ween ac ual ea nings (AE) and cash flow om ope a ions (CFO): TAi =AEi −CFO i . SALEi is he change in sales o fi m i in pe iod compa ed wi h -1. RECi is he change in ecei ables o fi m i in pe iod compa ed wi h -1. PPEi is p ope y, plan and equipmen o fi m i in pe iod . B Mi is he book o ma ke a io o fi m i in pe iod . CFO i is he cash flow om ope a ions o fi m i in pe iod . Taking in o accoun he defini ion o ITVE, we alloca ed al- ues o 1 and 0 depending on he es ima ions ob ained wi h La cke and Richa dson’s (2004) model. A alue o 1 was assigned o fi ms o which NDEi <EFi and DAi >0, and a alue o 0 was assigned o he wise. Using he new alues o ITVE a iable (ITVEi ) ob ained wi h La cke and Richa d- son’s (2004) model, we pe o med eg ession (2) again and ob ained he esul s p esen ed in Table 13. As shown in he able, he esul s a e no e y di e en om hose p esen ed in Sec ion 5, al hough he model co - ec ly classifies only 59.5% o he cases, which is lowe han he 61.7% co ec ly iden ified using he o he model. As abo e, he explana o y a iable, mean|CAR|, was significan a he 5% le el and he coe ficien had he expec ed sign. Wi h espec o he con ol a iables, ROI and a GDP we e significan a he 10% le el and had posi i e and nega i e