Table 1Empirical evidence on the relationship between audit quality and auditor tenure.. Audit quality Auditor tenure Carcello and Nagy 2004 ≤3 years Ns − Carey and Simnett 2006 1021 Pub
Trang 1R E V I S T A D E C O N T A B I L I D A D
S P A N I S H A C C O U N T I N G R E V I E W
w w w e l s e v i e r e s / r c s a r
Auditor tenure and audit quality in Spanish state-owned foundations
Belén González-Díaza,∗, Roberto García-Fernándezb, Antonio López-Díazb
a Universidad de Oviedo, Departamento de Contabilidad, Facultad de Comercio, Turismo y CC Sociales, Laboral-Ciudad de la Cultura, Gijón, Asturias, Spain
b Universidad de Oviedo, Departamento de Contabilidad, Oviedo, Asturias, Spain
a r t i c l e i n f o
Article history:
Received 24 June 2013
Accepted 3 April 2014
Available online xxx
JEL classification:
C350
H830
L310
M420
Keywords:
Auditor tenure
Audit quality
Non-profit organizations
Foundations
Logistic regression
a b s t r a c t
Thispaperaimstoanalyzetheimpactofauditortenureonauditquality.Theresearchismotivatedby theabsenceofconsensusinpublishedworks,andbythescarcityofstudiescarriedoutonnon-profit organizations.Usingasampleof254auditscarriedoutbetween2003and2010onSpanishstate-owned foundations,wefindthat,althoughfoundationauditqualitydecreasesastenurelengthincreases,this qualitylossdoesnotbecomeapparentuntilthesixthyearofthefoundation–auditorrelationship,after
aninitialfiveyearsofimprovementinquality.Theempiricalevidenceisimportantforregulatorsand financialstatementusers,giventhatitsuggeststheneedfortheintroductionoftenure-reducingmeasures which,atthesametime,alsoensureaminimumtenureperiod
©2013ASEPUC.PublishedbyElsevierEspaña,S.L.Allrightsreserved
Permanencia del auditor y calidad de la auditoría en las fundaciones públicas estatales
Códigos JEL:
C350
H830
L310
M420
Palabras clave:
Permanencia del auditor
Calidad de la auditoría
Entidades no lucrativas
Fundaciones
Regresión logística
r e s u m e n
Estetrabajoanalizaelimpactodelapermanenciadelauditorsobrelacalidaddelaauditoríaenlas entidadesnolucrativas.Estainvestigaciónestámotivadaporlaausenciadeconsensoenlaliteraturasobre estacuestiónylaescasezdeestudiosrealizadosenelsectordelasentidadesnolucrativas.Utilizando unamuestrade254auditoríasllevadasacaboparaelperíodo2003–2010sobrefundacionespúblicas estatales,observamosque,sibienlacalidaddelaauditoríadelasfundacionesdisminuyeamedidaque
lapermanenciadelauditoraumenta,estapérdidadecalidadnosemanifiestahastaelsextoa ˜nodela relaciónfundación–auditor,yaqueenloscincoprimerosa ˜noslacalidadaumenta.Laevidenciaempírica
deestarelacióntieneimportantesimplicacionesparaloslegisladoresylosusuariosyaquesugierela necesidaddeintroducirmedidasquelimitenladuracióndelamismay,almismotiempo,asegurenuna duraciónmínima
©2013ASEPUC.PublicadoporElsevierEspaña,S.L.Todoslosderechosreservados
∗ Corresponding author.
E-mail address: bgonzal@uniovi.es (B González-Díaz).
http://dx.doi.org/10.1016/j.rcsar.2014.04.001
1138-4891/© 2013 ASEPUC Published by Elsevier España, S.L All rights reserved.
Trang 2deContabilidad–SpanishAccountingReview(2014).http://dx.doi.org/10.1016/j.rcsar.2014.04.001
ARTICLE IN PRESS
G Model
RCSAR-32; No of Pages 12
2 B González-Díaz et al / Revista de Contabilidad – Spanish Accounting Review xxx (xx) (2014) xxx–xxx
1 Introduction
Over thelast few decadesthe relationship betweenauditor
tenureandauditqualityhasbeenconstantlydebated.Evenwhen
priorresearch hasbeenwidespreadand not completely
defini-tive(Knechel&Vanstraelen,2007,113),mostofithasinvolved
inthefor-profitsector,specificallypubliclytradedcorporations
However,therehasbeenlittleresearchdoneintotheeffectofthe
auditoronauditqualityinnon-profitorganizationsandthatwhich
hasbeendone,aswillbeshowninthefollowingsection,hascome
outofresearchintoothermatters
Thisstudyprovidesfreshempiricalevidencewhichaddstothe
debateasitexaminestheeffectofauditortenureonauditquality
inasinglesector–non-profitmaking–wherethereishardlyany
empiricalevidenceaboutthisrelationship
Thisanalysisisparticularlyrelevantatthepresenttimesince
governmentsarebeingforcedby theeconomic recessionto
re-structureandrationalizeapublicsectorwhichhasballoonedover
thelastfewdecadeswiththecreationofnon-profitorganizations
toundertakecertainpublicfunctions(Lohmann,2007).Such
orga-nizationsmustconvincethegeneralpublicthattheirpoliciesand
systemsaretherightonestoguaranteeappropriatemanagement
oftheresourcesprovidedbytaxpayersforthemtocarryoutthe
activitiesforwhichtheyweresetup(Greenlee,Fischer,Gordon,&
Keating,2007)
Anauditis aninstrumentwhich inspiresconfidencein both
external users of financial data concerningthese organizations
(beneficiaries,publicbodies,donors)andinternalusers–mainly
financialdirectors(Bellostas,Brusca,&Moneva,2006).Thisis
espe-ciallytruewhentheauditopinionisunqualified
Thekindofauditopiniongivennotonlysuggeststhatthe
orga-nizationiscomplyingwithaccountingregulationsandisconcerned
aboutitsfinancialmanagement;italsobecomesamajorfactorin
identifyingorpreventingfraudulentactivity(Bell&Zimmerman,
2007)
Audit report opinions are, nevertheless, affected by
differ-entfactorswhich have been dealtwithin a number ofpapers
(González-Díaz, García, &López, 2013; Gosman, 1973; Ireland,
2003;Keasey,Watson,&Wynarczyk,1988;Krishnan,Krishnan,&
Stephens,1996).Onefactorisauditortenure
Theaimofthispaperistoanalyzehowtenureaffectsaudit
qual-ityinfoundations.Qualityisdefinedfromtheviewpointofexternal
usersofthefinancialstatementsasthelikelihoodthatanauditor
willsubmitaqualifiedopinion.Asampleof254auditscarriedout
between2003and2010wasused,representing46different
foun-dations.Severallogisticregressionmodelswerecalculatedto
mea-sureindifferentwaystheeffectofauditortenureonauditquality
Theresultsshowthat,althoughauditqualitydiminishesasthe
periodof auditortenureincreases,this lossofquality doesnot
becomeapparentuntilthesixthyearofthefoundation–auditor
relationship.Infactitimprovesoverthefirstfiveyears.The
empir-icalevidenceofthisrelationshipisimportantforpublicauditing
giventhatithighlights theneedfor theintroductionof
tenure-reducingmeasureswhich,atthesametime,alsoensureaminimum
tenureperiod
Theremainderofthepaperisorganizedasfollows:Thefirst
sec-tionreviewsrelevantliteratureregardingauditqualityandauditor
tenure.Thesecondsectionexplainsregulationsconcerning
foun-dationauditsandthehypothesisbehindtheresearch.Thethird
sectionoutlines thestudy’s methodology Resultsfollow in the
fourthsectionand,finally,conclusions
2 Prior research
DeAngelo(1981)definesauditqualityastheprobabilityofan
auditordiscoveringerrorsinaclient’sfinancesandthenbringing
theseerrorstolightintheauditreport.Auditquality,therefore, dependsonauditorcompetenceandindependence
Competenceisassociatedwithanauditor’sprofessionalskills andindependencemaybereal(theauditor’sunbiasedor objec-tiveattitude),orjustappear tobeso(different userperception
ofindependence).Someliteratureonthesubjectofauditquality considersthattheauditorisabletoseparatebothfeatures,while otherauthors assumethem to belinked.Thus, if anauditor is competent,themorelikelytheyaretobeindependent(Richard,
2006)
Giventhat itisdifficulttocomeupwithaproxy whichcan assessboth auditorcompetenceand independenceatthesame time (Vanstraelen, 2000, 420), and that the cost of measuring thequalityoftheauditor’sworkishighlysignificant,consumers developsubrogatesforauditquality(proxies)whichmaybe corre-latedtoquality(DeAngelo,1981).Carcello,Hermanson,andHuss (1995)pointoutsomeofthesewhichhave beenusedbyother authors: litigation against law firms, auditor selection, auditor changesandfirmsize,natureofauditors’opinions,pricingofaudit servicesanduserperception
Also,overthelastfewdecadesresearchershaveanalyzed deter-miningfactorsinauditqualityaswellastheeffectofauditortenure
onquality,withanumberofstudiesdevotedprincipallytothe lat-terandundertakenintheprivatesector–Table1summarizessome
oftheworkspublishedinthisregard
Specializedliteratureonthesubjecthaspointedoutthelack
ofconsensusconcerningtheauditquality–auditortenurerelation (Vanstraelen,2000)becauseauditortenurecanhaveapositiveor negativeimpactonthetwomaindeterminantsofauditquality: auditorcompetenceandauditorindependence
Geiger and Raghunandan (2002), Myers, Myers, and Omer (2003),GhoshandMoon(2005),KnechelandVanstraelen(2007) andJackson,Moldrich,andRoebuck(2008)haveshownthataudit quality improves with auditor tenure However, Levinthal and Fichman(1988)andDeisandGiroux(1992)showjusttheopposite
Bymeasuringtenureasthenumberofyearsanauditorhasaudited
acompany,thesestudiesconsidertheauditquality–auditortenure relationtobelinear
Ruiz,Gómez,andCarrera(2006)suggestthatdivergencesfrom
anempiricalpointofviewmaybeduetothefactthataudit qual-itydoesnotvaryinalinearwayoverthedurationofthecontract andthatitmaychangedependingonthedurationofclient/auditor relationship
Longauditortenuremayincreasecompetencebecausethe audi-tor’sclient-specificknowledgeincreasesovertheyears(St.Pierre andAnderson,1984).Thiswillallowthemtoimprovethequalityof theirauditingbutitcouldalsoreducetheirdegreeofindependence
inthesensethatalongauditor–clientrelationshipmaymakethe auditorfinanciallyreliantontheclient(Ruizetal.,2006)andbring aboutsuchacloserelationship(Whittington,Grout,&Jewitt,1995) thatunbiasedassessmentiscompromised(Shockley,1981)bylack
ofbothinnovationandproceduralrigour(Schockley,1982) Johnson,Khurana,andReynolds(2002)claimthatspecificclient knowledgegainedovertheyearscouldentailareductioninauditor effort,yetthisdoesnotnecessarilyinvolveathreattothequality
ofthework
Short auditor tenure could negatively affect competence becauseauditors’clientknowledgeislessoverthefirstfewyears andtheyneedtimetogetusedtotheirclients’activityand account-ingprocedures(Carcello&Nagy,2004).Theirindependencecould alsobecompromisedsincetheyneedtokeepnewclientsinorder
torecovertheirinitialclient-specificinvestment,whichcannotbe transferredtoothercontracts(Ruizetal.,2006)
Industryspecialization,however,canimprovebothcompetence andindependence,leadingtohigherauditqualitysinceauditors knowtheirsectorfarbetterthannon-specialistfirmsandcanaudit
Trang 3Table 1
Empirical evidence on the relationship between audit quality and auditor tenure.
Audit quality Auditor tenure
Carcello and Nagy
(2004)
≤3 years
Ns
− Carey and Simnett
(2006)
1021 Public companies listed
on the Australian Stock Exchange
1995 Issuing a going-concern audit
opinion
>7 years
≤2 years
− Ns Abnormal working capital
accruals
>7 years
≤2 years
Ns Ns Extent of earnings
management: just misses breakeven
>7 years
≤2 years
−
− Extent of earnings
management: just beats breakeven
>7 years
≤2 years
Ns Ns Deis and Giroux
(1992)
232 Quality Control Review (QCRs) on CPA audits of Texas Independent School District (ISD)
1983–1988 Natural log of the weighted
quality metric based on the QCRs letters of findings
Number of years the auditor has audited the ISD
−
Geiger and
Raghunandan
(2002)
117 Public company bankruptcies
1996–1998 Issuing a going-concern audit
opinion prior to bankruptcy
Natural log of auditor tenure
in years
+
Ghosh and Moon
(2005)
Traded firms (35,826 firm-years)
1990–2000 Earnings response coefficients
(ERCs)
Duration of the auditor–client relationship in years
+
opinion
Length of the auditor–client relationship
in years
+
Discretionary accruals Length of the
auditor–client relationship in years
Ns
Johnson et al (2002) US corporations (11,148
firm-year observations)
1986–1995 Absolute value of unexpected
accruals
≥9 years 2–3 years
Ns + The persistence of the accrual
components of earnings
≥9 years 2–3 years
Ns + Knechel and
Vanstraelen (2007)
309 private Belgian companies bankruptcies
1992–1996 Issuing a going-concern
audit opinion
Duration of the auditor–client relationship in years
Ns
309 private Belgian companies non-bankruptcies
1992–1996 Issuing a going-concern
audit opinion
Duration of the auditor–client relationship in years
Ns
Krishnan and
Shauer (2000)
164 voluntary health and welfare organizations
N.a The entity’s compliance with
eight GAAP reporting requirements
Levinthal and
Fichman (1988)
opinion
The number of years from the beginning of an auditor–client relationship to its ending
−
Lim and Tan (2010) Non-financial firms audited
by Big N auditors (12,786 firm-years)
2000–2005 Modified accrual quality
measure by McNichols 2002
Median tenure in the sample
−
≥9 years
≤3 years
Ns + Lowensohn et al.
(2007)
241 surveys to finance officials from counties, local municipalities and special districts within the State
of Florida
Audits for the fiscal year ending September 30, 2002
Perceived quality of audit Natural logarithm
of the audit firm’s tenure as the government’s auditor
+
Myers et al (2003) All firms-years with sufficient
data on the 2001 Compustat annual industrial
(42,302 firm-years
consecutive years that the firm has retained the auditor
+
Trang 4deContabilidad–SpanishAccountingReview(2014).http://dx.doi.org/10.1016/j.rcsar.2014.04.001
ARTICLE IN PRESS
G Model
RCSAR-32; No of Pages 12
4 B González-Díaz et al / Revista de Contabilidad – Spanish Accounting Review xxx (xx) (2014) xxx–xxx
Table 1 (Continued)
Audit quality Auditor tenure
Monterrey and
Sánchez-Segura
(2007)
136 non-financial companies listed on the Spanish Stock Exchange (396 firm-years)
Ruiz et al (2006) 377 companies listed on the
Spanish Stock Exchange
1990–2000 Issuing a going-concern audit
opinion
≥8 years
≤3 years
Ns + Vanstraelen (2000) 796 financially companies 1992–1996 Issuing a qualified audit
opinion
Length of the auditor–client relationship
in years
−
Ns: not significant; Na: not available.
newclientsmoreeasily.Also,thehigherafirm’sreputationinits
sector,themoreindependenceitseemstohave(Lim&Tan,2010)
Vanstraelen(2000),Johnson etal (2002),Carcelloand Nagy
(2004),CareyandSimnett(2006),Ruizetal.(2006),andMonterrey
andSánchez-Segura(2007)acknowledgethenon-linearnatureof
therelationbyattemptingtoshowthatthedurationofanauditor’s
contractcaninfluenceauditquality
Vanstraelen(2000)inparticularpointsoutthatauditor
perfor-manceisdifferentinthefirsttwoandinthefinalyearsoftenure.She
claimsthatthelikelihoodofissuingaqualifiedopinionisgreaterin
thefinalyearthaninthefirsttwobecausetheauditorknowsthat
theircontractwillnotberenewedandthatthecurrentauditwill
bethelast
Johnsonetal.(2002),lookingatacontextinwhichauditor
rota-tionisnotcompulsory,provideevidencethatfinancialreporting
qualityisassociatedwithshortaudit-firmtenures.Ruizetal.(2006)
confirmthefactthatauditqualitygrowsoverthefirstfewyears
butthendoesnotdropawayinlongtenures.CarcelloandNagy
(2004,55)pointoutthat“fraudulentfinancialreportingismore
likelytooccurinthefirstthreeyearsoftheauditor–client
rela-tionship”althoughtheyfailtoshowthatthishappenswithlonger
tenures
Monterrey and Sánchez-Segura (2007) show that when the
auditor–clientrelationshiphaslastedover5yearsthebenefitsof
suchfamiliaritybecomeobvious,astheauditorhasgottoknow
theclientwell.Thestudysuggeststhatrotationregulationsshould
beaimedatensuringlongtenure,asconstantchangesdonotbring
aboutgreateraccountingquality
Ontheotherhand,CareyandSimnett(2006,674)claimthat
thelongertheauditpartnertenurethelowerthequalityofthe
audit,whenqualityismeasuredintermsoftheauditor’spropensity
toissueagoing-concernauditopinionandjustmeeting(missing)
earningsbenchmarks
Byanalyzingthesepreviousstudiesitisclearthattheir
empiri-calresultsarenotconclusive.Ewelt-Knauer,Gold,andPott(2012,
2013, 35) suggest that the positive or negative effects of the
client/auditorassociationdependonwhatresearchmethodisused
aswellastheproxychosentomeasureauditquality.Theyalso
pointoutthat,byanalyzingstakeholders,“regulatorstakeastance
infavourofrotation,arguingthatrotationprovidesanopportunityto
overcomeproblemscausedby(excessive)tenure.Attheotherextreme,
auditfirmsarecriticalandpointtoalossofknowledgeandexpertise
potentiallycausedbyrotation.Theviewsofauditclientsand
share-holdersoverallappeartoberelativelymixed”
The audit quality–auditor tenure relation has barely been
researchedinthepublicandnon-profitsectors.Inthepublic
sec-torDeisandGiroux(1992)lookatwhatdeterminesauditquality
inindependentschooldistrictsinTexasandLowensohn,Johnson,
Elder,andDavies(2007)studytherelationbetweenaudit
qual-ityperceivedby241Floridalocalgovernmentfinancedirectors
andcertainaudit/auditorattributes.Theresultsofthesestudies arecontradictorywiththequality–tenureresultsbeingnegativein theformerandpositiveinthelatter
Inthenon-profitsectorKrishnanandShauer(2000)examine therelationbetweenqualityauditingandauditortenure,although themainaimoftheirstudyistherelationbetweenauditorsize andtheauditqualityof164voluntaryhealthandwelfare organi-zationsinsouth-easternPennsylvaniaand southernNewJersey Theirresultsshowzeroimpactonauditquality
Otherworksconcerningthenon-profitsectorincludeastudyof auditortenurebeingadeterminingfactor,notinauditqualitybut
inaudit-firmfees(Ellis&Booker,2011;Vermeer, Raghunandan,
&Forgione,2009)orinformationoninternalcontrolweaknesses (López&Peters,2010).Inthiscontext,theaimofthisstudyisto provideevidencefromthenon-profitsectorthat mayassistthe debateaboutwhetherornottoestablishtenurelimits
3 Auditing of state-owned foundations
Foundationsareorganizationswhichcollaboratewiththestate
inordertoachieve aimsofgeneralinterest(IGAE,2010).Tobe
ofgeneralinterest,oneorbothofthefollowingpre-requisitesare necessary:
• Theyshould becreated through a direct or indirectmajority contributionfromtheGeneralStateAdministration,itspublic organismsorotherentitiesbelongingtothepublicsector
• Morethan50percentoftheirfixedassetsshouldcomprisegoods
orrightswhichhavebeenprovidedbyortransferredfromthe above-mentionedpublicentities
ExternalauditsoftheseorganizationsareregulatedbytheLaw 50/2002onFoundations,whichcameintoeffecton1January2003, andbytheRoyalDecree1337/2005,whichencompasses regula-tionsreferringtostate-ownedfoundations.TheGeneralBudgetary Act(GBA)47/2003alsoincludesregulationsgoverningfoundations TheFoundationsActdesignatesthe“IntervenciónGeneraldel Estado”(IGAE)(GeneralStateComptroller)astheauditorof foun-dations.TheIGAEisrequiredtoaudittheseorganizationsovertwo consecutiveyearswhenatleast2ofthefollowingsituationsarise
2yearsinarow:
• Assetsamounttoover2,400,000Euros
• Annualrevenuesamounttoover2,400,000Euros
• Thefoundationhasover50employees
Establishing a minimum related to assets, state subsidies, incomeandcostsorthenumberofemployeesisageneralcriterion appliedinmostcountriestodecidewhetherornotanon-profit organizationshouldbeauditedornot(Kitching,2009;Tate,2007)
Trang 5accord-ingtothenatureofthefoundation(EFC,2011)
FoundationswhichcomeintotheIGAEcompulsory-audit
cat-egoryareconsideredlarge.Smallandmedium-sizedfoundations
arenotalllegallyrequiredtobeauditedandyetsomevoluntarily
requestprivatefirmaudit
Upuntilnowtypesoffoundationauditorshavebeenanalyzed,
but not thelengthof time their relationshiplasts TheIGAE is
requiredtocarryoutannualexternalauditsonlargefoundations,
whichentailsthepublicauditorhavinganindefiniterelationship
withthefoundationunlessiteitherceasestoexistorisnolonger
considered“large”.Asfarasprivateauditorsareconcerned,
Span-ish auditlaw states that, when audits are compulsory, private
auditorscannotbecontractedbyanysingleorganizationforless
thanthreeyearsormorethannine.Whenauditsarevoluntary,
however,as isthecase in this study,theserestrictionsare not
applied,whichmeansthatafoundationthatchoosestobeaudited
annuallycanmaintainanindefiniterelationshipwithanauditor
Itis,therefore,anenvironmentwithoutmandatoryauditor
rota-tion
The“TribunaldeCuentas”,asthesupremeauditingandfinancial
managementbodyoftheSpanishstateandofthepublicsector,may
undertakecontrolactivityinstate-ownedfoundations Sinceits
beginningsin1982,ithasbeeninvolvedinscrutinizingand
check-ingpublicaccounts,includingthestatefoundationsector,where
externalauditscarriedoutbytheIGAEfigureprominently
Also,the“TribunaldeCuentas”annualprogrammemayinclude
foundationauditing.Between2003and2010thisinstitutionissued
9reportsontheauditingofstate-ownedfoundations.Thereports
deal with a variety of auditing objectives, from financial and
legalauditstomanagementefficiencyandcompliancewithhiring
requirements
The scarcity of academic studies on the non-profit
sec-tor prevents us from reachingany firm conclusions about the
auditquality–auditortenurerelation.However,diverseempirical
evidenceintheprivatesector,plusthelegalframeworkfor
foun-dationsinSpain,leadsustothefollowinghypothesis
Hypothesis. Thereisanassociationbetweenauditqualityand
auditortenure
4 Data and methodology
4.1 Sampleselection
The main data analyzed in this study was obtained from
theInventoryofState/PublicSectorOrganizations(INVESPE)–the
mainsourceandasoftwareapplicationwhichhasbeenprepared
andupdatedbytheIGAE–aswellastheDeclarationofthe
Gen-eralStatementofStateAccounts,andthefinancialreportsonstate
foundationsfrom2003to2010.Thestudystartedin2003,when
theFoundationsLaw(whichestablisheswhoexactlyshouldaudit
foundations)cameintoeffect.Itconcludedin2010,theyearwith
thelatestavailabledata.Prior totheFoundationsLaw,Spanish
legislationdidnotspecifyeitherwhoshouldauditfoundationsor
whichfinancialreportingregulationsshouldbeapplied(González,
García,&López,2011)
Ofthe 256audits carried out over this period, 2have been
excludedasoneofthemcontainsanadverseopinionandtheother
isadisclaimer.Thefinaldatasetusedforthisstudyincludes254
auditswithpositiveorqualifiedopinions
4.2 Methodology
Theeffectofauditortenureonauditqualityinfoundationswas
examinedviatheestimationoflogisticregressionmodelsinwhich
auditquality(audqual)isthedependentvariable.The indepen-dentvariableisauditortenure(audtenure)andcontrolvariables are:typeofauditor(auditor),size(size),previousyear’sopinion (prevyear'sopin),thefoundations’revenueexceedsitsexpenses (surplus),thesector(department)andtheyear(year)
4.2.1 Dependentvariable Audqualisadummyvariablewhichtakesthevalues1and0 dependingonwhetherthereportisqualified(1)orunqualified(0) Mostofthestudieswhichwereexamineduseonlyoneproxy
tomeasureauditquality,exceptJohnsonetal.(2002)andJackson
etal.(2008),whichusetwo,andCareyandSimnett(2006)which usesfour
Inthisstudyauditqualityhasbeenmeasuredfromthepoint
ofviewofexternalusersoffinancialstatementsasthelikelihood that anauditor willissuea qualified opinion –the proxyused
byLevinthalandFichman(1988)andVanstraelen(2000).When
anauditorissuesanuncleanauditreportitmeanstheyareable
to objectivelyassess business resultsand resist client pressure
toissueacleanopinion(DeFond,Raghunandan,&Subramanyam,
2002,1248–1249).Thissuggeststhatthereisapositivecorrelation betweentheissuingofaqualifiedopinionandthelevelofauditor competenceandindependence
Moreover,thisproxy waschosen duetothefactthataccess
tofinancialinformationfromtheseorganizationsisverydifficult giventhattheyfailtocomplywitharticle136oftheGeneral Bud-getaryActwhichrequirestheirannualaccountstobepublishedin theOfficialGazette.Moreover,notallofthosewhohavepublished theirannualaccountsintheGazettehaveincludedtheauditreport,
akeydocumentforidentifyingandclassifyingqualifiedopinions thatcanbeusedasauditqualityproxies.Itishoped,nevertheless, thatthenewLaw,19/2013ontransparency,informationaccess andgoodgovernment,whichrequiresfoundationstopublishtheir annualaccountsandauditreports,willreinforcetheircommitment
topublishintheGazette
4.2.2 Independentvariable Tenureismeasuredintwoways:asacontinuousorasadummy variable.Intheformercase, tenureiscalculatedasthenumber
ofconsecutiveyearsafoundationhasbeenauditedbythesame auditor(Ellis&Booker,2011;Geiger&Raghunandan,2002;Ghosh
&Moon,2005;Gul,Jaggi,&Krishnan,2007;Lowensohnetal.,2007; Myersetal.,2003).Theyear1999waschosenasthestartingpoint becausethefoundationsfinancialreportsbecameavailablethen Usingthisproxyentailsregardingtheauditquality–auditortenure relationaslinear;inotherwords,thelongerthetenurethegreater
orlessertheauditquality
Forthedummyvariablethreemeasurements,obtainedby cal-culatingtenurequartiles(Q1=2;Q2=4;Q3=6)foreverysample, were used The three measurements, therefore, are defined as tenure≤2years,tenure≤4yearsandtenure≥6years
Itisusualtobearinmindtheregulationseachcountryhas con-cerningthelengthofauditortenureinordertodefinethetenure variablewhenauditqualitydoesnotvaryinalinearwayforthe durationofthecontract.Inthissense,theUnitedStatesandsome
EUcountriesregulatetenure(Vanstraelen,2000)bysetting min-imumandmaximumperiods,whichsomestudieshave usedas
areferencefordefiningshort,mediumandlongtenure(Carcello
&Nagy,2004;Gunnyetal.,2007;Ruizetal.,2006),orsimplyas
ameansofspecifyingminimumtenure,attheendofwhichthe clientandauditormayputanendtotheirrelationship(Knechel& Vanstraelen,2007)
Thestudyhasbeenunabletoconsidercurrenttenurelegislation
inSpainbecause,aswaspointedout,thereisnoupperorlower tenurelimitforauditorsofstatefoundations.Thecriteriausedto definedummyvariableswastocategorizethecontinuoustenure
Trang 6deContabilidad–SpanishAccountingReview(2014).http://dx.doi.org/10.1016/j.rcsar.2014.04.001
ARTICLE IN PRESS
G Model
RCSAR-32; No of Pages 12
6 B González-Díaz et al / Revista de Contabilidad – Spanish Accounting Review xxx (xx) (2014) xxx–xxx
variableasadummyvariable,accordingtothelimitsestablishedby
thequartilesasproposedbyMiján(2002,392)inordertodetect
theindependenteffectsthat mayexist ateachleveland,atthe
sametime,bearinmindobservationallocation.LimandTan(2010),
whencategorizingauditortenure,usepercentilesasthemedian
andFitzgerald,Thompson,andOmer(2012)classifyauditortenure
intoshort(1–2years),medium(3–5years),andlong(6ormore
years)
4.2.3 Controlvariables
Thecontrolvariablesarethosewhichhavebeenidentifiedin
previousstudiesasbeingpossibledeterminingfactorsinquality
audits:auditor,size,opinionfromthepreviousyear’sreport,
sur-plus,sectorandyear
Externalauditsofthefoundationsarecarriedoutbyprivate
auditorsorbytheIGAE,dependingoncertainconditions.There
isadummyvariable(auditor)whichtakesavalueof1whenthe
auditorsareprivateand0whenitistheIGAE.Therearealso
differ-encesofopinionconcerningtheinfluenceofauditortypeonaudit
quality.Johnsonetal.(2002),Ruizetal.(2006)andLimandTan
(2010)donotfactor inaudittype whenanalysingtenure
influ-enceonquality.YettheresearchofVanstraelen(2000),Knechel
andVanstraelen(2007)andMonterreyandSánchez-Segura(2007)
doesstudyitsimpactwhilstmakingadistinctionbetweenthe“Big
4”auditfirmsandtherest.Theresultsarebynomeansuniform;
Vanstraelen(2000)andKnechelandVanstraelen(2007)findno
empiricalevidencethatauditortypeaffectsquality;Monterreyand
Sánchez-Segura(2007),however,concludethatqualityimproves
iftheauditorisalargefirmanditstenureislessthanfiveyears
Thisresearchexplorestheeffectofauditortypeonauditquality
inacontextwhereauditortypedependsonthesizeoftheaudited
organization,distinguishingbetweenpublicandprivatetypesof
auditorratherthanbetweenlargefirmsandtherest
Inthissense,existingliteraturepointsoutrelevantdifferences
betweenpublicandprivateauditors.Jakubowski(1995)carriedout
astudywhichattemptedtoascertainwhetherthetypeofauditor
–publicorprivate–mightormightnotaffectthenumberof
con-trolweaknessesdetectedinlocalgovernments.Thestudyrevealed
thatstateauditorsdiscoveredmoreweaknessesthandidprivate
firms.Theauthorconsidersthatstateandprivateauditorsviewthe
auditprocessdifferently.Whereastheformerareunderno
pres-surefromtheirclients(taxpayers),thelattermayloseclients(local
governments)ifthenumberofqualificationspublishedis
partic-ularlyhigh.Thiswouldhelptoexplainwhystateauditorsreport
moreweaknessesthanbothlargeandsmallprivatefirms
In a report commissioned by the GAO (Government
AccountabilityOffice,2007)thePresident’sCouncil onIntegrity
and Efficiency reviewed a sample of 208audits undertaken by
stateandprivateauditorsonrelevantstateandlocalgovernments
andnon-profitorganizationsin2003.Thereportconcludedthatan
auditwasunacceptablewhendeficiencieswere“soseriousthat
theauditors’ opinionon atleast onemajor programme cannot
bereliedupon;e.g noevidenceof internalcontrol testing and
compliancetestingfor allormostcompliancerequirementsfor
oneormoremajorprogrammes,unreportedauditfindings,and
atleastoneincorrectlyidentifiedmajorprogram”–30.29percent
oftheauditsreviewedwerethusratedandallofthesehadbeen
conductedbyprivateauditors
Thisresult may beexplained by Jakubowski’s thesis, but it
could also be true that private auditors fail to devote enough
resourcesfordetectingweaknessesininternalcontrolsystemsin
theseorganizations.In fact,Lópezand Peters(2010)obtaineda
resultwhichchangedtheempiricalevidenceofpreviousstudies
Theyresearchedtherelationshipbetweenopiniontypeandauditor
typebylookingatasampleofauditreportspertainingtoasetofUS
citiesandcountiesovertheperiod2004–2006.Theyreachedthe
conclusionthatthelikelihoodofdetectinginternalcontrolissues increasesiftheauditorsareprivate.Theyarguethat,intheUSA, privatefirmsboosted resourcesfor assessinginternalcontrolin theauditstheycarriedout
On the other hand, Dehkordi and Makarem’s study (2011) publisheda year aftertheone by López and Peters and which examines theeffects of auditor type and size onaudit quality
inIran,concludesthatfinancialstatementsauditedbythe pub-lic“AuditOrganization”containfewerdiscretionaryaccrualsthan statementsauditedbyprivatefirms.Thissuggeststhatwhenthe public auditor doesthe audit the quality is higher Bearing in mind that, with the exception of the López and Peters’ study
in 2010,allotherworkis unanimous, it istobe expectedthat thesignofthedependentvariable-typeofauditorrelationwillbe positive
Sizeisdefinedasthenaturallogoftotalassets(size).Some stud-iesshowthatthebiggeranorganization,thelowertheauditquality, sincethecomplexity ofitsoperations demandsspecificauditor knowledge(Krishnan&Shauer,2000;O’Keefe,King,&Gaver,1994) Therefore,apositivesignispredicted
Previousyear’sopinion(prevyear'sopin)isdefinedasadummy variablewhichtakesthevalueof1whentheopinioninthe pre-viousyear’sreportwasqualifiedand0intheoppositecase.This variablehasbeenincludedinthestudysince thereisempirical proofthatshowsthatthesameauditopinionisrepeatedovertime, increasingtheprobabilityofreceivingaqualifiedopinion ifthe previousyear’sopinionwasalsoqualified(Ireland,2003;Keasey
etal.,1988)
The auditfirm mayhave an incentive to repeatedly qualify auditreportsif,havingqualified thefirstone,thecompanyhas not decidedwhether itsserviceswill still berequired (Ireland,
2003).Also,thetypeofqualificationmaycausethequalified opin-ion to be maintainedover several years since a company that receivesaqualifiedopinionbroughtaboutbyuncertaintythe pre-viousyearisverylikely togetthesame opinioninthecurrent yeargiventhatuncertaintymayextendbeyondoneyear(González
et al.,2011; Monroe&Teh,1993).Therefore, a positive signis predicted
Thisstudyalsoincludesanindicatorofwhetherthe founda-tion’srevenuesexceeditsexpenses(surplus).Thisisexpressedas
adummyvariablewithavalueof1whenitsincomeislessthan itsrunningcosts(losses)and0intheoppositecase.Itisdefined
asadummyratherthanacontinuousvariablebasedonwork car-riedoutbyotherauthorsconcerningthenon-profitsector(Keating, Fischer,Gordon,&Greenlee,2005;Petrovits,Shakespeare,&Shih,
2011)
Non-profit organizations may have incentives to minimize theirprofits.Ontheonehand,donorsandregulatingbodiesmay gettheimpressionthat profitcontradictsthefundamental pur-poseof theorganization(Trussel,2003).A studycarriedoutby Calabrese(2011) showsthat future contributionsof donorsare negativelyaffectedwhenwealthlevelsaredeemedexcessive;on theotherhand,asprofitshouldbereinvestedintheorganization andnotsharedoutamongpartnersandemployees,itispossible thatcertainillicitpracticessuchasunwarrantedexpenseclaims couldtakeplace(Greenleeetal.,2007).Itistobeexpectedthat thecoefficientsignrelatedtothesurplusvariablewillbe posi-tive
Thevariableconcerningtherelationshipwithparticular gov-ernmentdepartments(department)isalsoadummyoneandtakes thevalueof1whenthefoundationbelongstotheTaxandFinance Department,and0intheoppositecase
Empirical evidence obtained from private sector studies (Bamber,Bamber,&Schoderbeck,1993;Maletta&Wright,1996) considersthattherearereasonstoassumethatthesectorinwhich
acompanyoperatesisavariablewhichcanexplainopiniontype
Trang 7greaterinternalcontrolandthattherequirementsforprogrammes
relatedtocertaintypesofsubsidyororganizationarestricterin
somesectorsthaninothers,whichentailvariationsinquality.The
variablesignisnotpredictedaspreviousstudieswhereithasbeen
includedhaveuseddiverseclassificationswhicharedifferentfrom
thatproposedinthisstudy(Keatingetal.,2005)
Statefoundationsmaybelinkedaccordingtosectorsbymeans
of their “Protectorate”, an umbrella organization whose aim is
toguaranteelegalandfoundationpurposecompliance.The
Pro-tectorate provides support, initiatives and legal, financial and
accountingadvice.ItcomesundertheauspicesofStateServices
throughdifferentgovernmentdepartmentswhoseresponsibilities
aremoredirectlyrelatedtothefoundationpurposes(
González-Díazetal.,2013)
Thedecisiontomakethisadummyvariablewastakenbecause
23.62percentofthe254selectedfoundationsreporttotheTax
andFinanceDepartmentandtheresttootherdepartments.Also,
amoredetailedbreakdowncouldcausestatisticalinference
prob-lemsbroughtaboutbythesmallnessofaparticulardepartment
(Sánchez&Sierra,2001)
Finally,theyear2003isincludedasacontrolvariable(year)
sincethisiswhentheFoundationsLaw,requiringtheIGAEtoaudit
largefoundations,cameintoforce.Thisvariabletakesthevalueof
1whenthefoundationwasauditedin2003and0intheopposite
case
4.2.4 Modelspecification
Theuseof thedependentaudqualvariableanda significant
numberofdummyindependentandcontrolvariables hasgiven
risetologisticregressionbeingusedastheanalysismethodology
(Keaseyetal.,1988)inwhichtheregressioncoefficientsestimate
theimpactoftheindependentvariableontheprobabilitythatthe
typeofopinionwillbequalified.Apositivesignforthecoefficient
meansthatavariableincreasestheprobabilityofaqualified
opin-ion;anegativesignindicatesthereverse
Themodelcanbeexpressedinthisway:
AUDQUAL=ˇ0+ˇ1AUDTENURE+ˇ2AUDITOR+ˇ3SIZE
+ˇ4PREVYEAR'SOPIN+ˇ5SURPLUS+ˇ6DEPARTMENT+ˇ7YEAR
AsummaryofallvariablesisincludedinTable2
Table 2
Description of all variables.
audqual = 1, if the audit opinion is qualified; 0, otherwise
tenure = The number of consecutive years that the
foundation is audited by the same auditor tenure ≤ 2 = 1 if the foundation is audited by the same
auditor for 2 consecutive years or less; 0, more than 2 consecutive years
tenure ≤ 4 = 1 if the foundation is audited by the same
auditor for 4 consecutive years or less; 0, more than 4 consecutive years
tenure ≥ 6 = 1 if the foundation is audited by the same
auditor for 6 or more consecutive years; 0, less than 6 consecutive years
auditor = 1, if the foundation is audited privately; 0 if the
auditor is the IGAE size = The natural log of total assets
surplus = 1, the foundation’s revenues do not exceed its
expenses; 0, otherwise prev year's opin = 1, if the previous year’s opinion is qualified;
0, otherwise department = 1, if the foundation belongs to Tax and Finance
Minister; 0, otherwise year = 1, if the foundation is audited in 2003;
0, otherwise
5 Results and analysis
Table 3 shows the descriptive statistics of the sample The tenurevariableprovidesameanvalueslightlyhigherthanthree and a median of two for qualified reports and greater values, for both themean and the median in unqualified reports The descriptivestatisticsofthetenurevariables,measuredasadummy variable,showthatthegreaterthetimeoftheauditor–foundation relationship, the lesser the likelihood of qualified reports Also, 18.97 percent of IGAE reports were qualified, whereas the percentage drops to 15 percent when the auditors were private
In addition, Table 3 describes the remaining variables used
in the study It can be seen that audit quality could in some waybelinkedtothepreviousyear’sopinionsince60percentof the foundationsthat got a qualified opinion had got thesame the previous year,as did 91.39 percent of those whose previ-ousyear’sopinion hadbeenunqualified.Yetitdoesnotappear thatrevenueearningsgreaterthanexpensesaffectqualitygiven that the percentage of qualified reports is very similar in the caseofbothpositive(17.76percent)andnegative(17.65percent) results
As for foundations’ relationships with government depart-ments, the descriptive data highlights the fact that those organizationsthatreceivefewerqualifiedreportsarethoseunder theauspicesofthedepartmentsofEducation,Scienceand Inno-vation(10.42percent),Environment(8.33percent),andTaxand Finance(8.33percent).AttheoppositeextremeareJustice(60 per-cent),HealthandConsumerAffairs(50percent)andPublicWorks (30.77percent)
Thedescriptivestatisticsoftheyearvariableshowsthatthe per-centageofqualifiedopinionsdecreasessteadilyfromthebeginning
oftheperiodstudieduntiltheendgiventhat30.43percentof foun-dationsreceivedaqualifiedopinionin2003andonly9.3percent
in2010
Soastodeterminepossiblerelationsbetweenexplanatory vari-ablesandopiniontype,aunivariatestudywascarriedoutwhereby observationsaredividedintothosewithaqualifiedandthosewith
anunqualifiedopinion,theaimbeingtodetectanymajor differ-encesamongthem(Table4)
TheresultsoftheKolmogorov–SmirnovTestsuggestthatthe explanatoryvariablesdonotfollowanormalpattern,exceptthe sizeone,whichalsoshowshomoscedasticity
Therefore,theStudent’st-testwasappliedtothesizevariable, thenon-parametricMann–Whitneytesttothevariabletenureand Pearson’schi-squaretesttothedummiesvariables
Itcanbeseenthatauditortenure(tenure,tenure≤2,tenure≤4 andtenure≥6)isdifferentaccordingtowhetherthereportswere cleanornot.Andalsoseveralcontrolvariables(prevyear'sopin and department) behave differently according to opinion type
Lastly,Table5showsthecorrelationmatrixbetweenthe differ-entvariables.Onlyoneofthecorrelationexceeds0.40(variables sizeandauditor)suggesting thatthereare nomulticollinearity problemsinthedata(Vermeeretal.,2009)
5.2 Multivariateanalysis
Table6showstheresultsofthefourestimatedlogistic regres-sionmodelsfortheperiod2003–2010,whichdifferonlyin the
Trang 8Table 3
Descriptive Statistics.
tenure ≤ 2
tenure ≤ 4
tenure ≥ 6
auditor
surplus
prev year's opin
department
year
Trang 9Table 4
Univariate analysis.
Variable Qualified opinion report mean
rank (sum of ranks)
Unqualified opinion report mean rank (sum of ranks)
Mann–Whitney Test (sig bilateral)
* Significant at 5%.
Table 5
Correlation matrix.
Variable tenure tenure ≤ 2 tenure ≤ 4 tenure ≥ 6 auditor size prev year's opin surplus department year
tenure ≤ 2 −.710 ** 1
* Significant at 5%.
** Significant at 1%.
waysthevariableaudtenureis defined.Theresultsare
consis-tentwiththoseobtainedintheunivariateanalysis,apartfromthe
yearvariable,whichisnowstatisticallysignificant
Theresults provideevidence for themain variable of
inter-est for this study Auditor tenure is significant whether there
is a linear relation (tenure) or not (tenure≤2; tenure≤4;
tenure≥6) withthe variable AUDQUAL The results for model
1 suggest that the longer theauditor tenure is, thelesser the
likelihood of receivinga qualified report,and sothe lowerthe
quality of the audit.Previously quoted studies (Deis &Giroux,
1992;LevinthalandFichman,1988;Vanstraelen,2000)coincide
whenstatingthatlong-termauditor–clientrelationships
signifi-cantlyincreasethelikelihood ofauditorsissuing anunqualified
opinion
Resultsobtainedwhendummyvariablesareused(models2–4)
tomeasuretenureindicatethatthisvariableshowsapositivesign
andisstatisticallysignificantforauditor–foundationrelationships
offewerthan5years(models 2and3)andanegative,
statisti-callysignificantsignforrelationshipsover5years(model4).In
otherwords,whentheauditorhasbeenauditingafoundationfor
atleast6 years thelikelihoodof itgettinga qualified reportis
reduced
Overall, the empirical results for Spanish state-owned
foundationsshow that while audit quality decreasesas tenure
increases,thisqualitylossdoesnotbecomeapparentuntilthesixth
yearof the auditor–clientrelationship as auditquality actually
increasesoverthefirstfiveyears
Asfarasthecontrolvariablesareconcerned,itshouldbepointed outthatthevariableprevyear'sopinseemstoplayanimportant part;thesignispositiveandstatisticallyvariableandsuggeststhat foundationswhichhavereceivedaqualifiedreportoneyeartend
toreceivethesamethefollowingyear
The variable department shows a negative, statistically sig-nificantsign, suggesting that belongingtothe Taxand Finance Departmentreducesthelikelihoodofreceivingaqualifiedreport ThisresultisinaccordancewiththatofKeatingetal.(2005)as
itshowsthatthesectorafoundationbelongsto,inthiscaseits departmentalrelationship,canaffectauditreportopinion Thevariableyear’ssignispositiveandstatisticallysignificantfor
2003.Thisdemonstratestherepercussionsofthechangein Foun-dationLawregulationswhichcameintoeffectin2003andwhich substantiallymodifiedthepreviousregulationsconcerningauditor typeandmadeauditingoflargefoundationscompulsory.2003is whenalargenumberoffoundationswereauditedforthefirsttime
bytheIGAEinsteadofbyprivatefirms
Finally,thestudyhighlightsthefactthatthevariablesauditor, sizeandsurplusdonotappeartoinfluenceaudqual.
Therobustnessoftheresultswastestedbyusingalternative definitionsforsomevariables.Specifically,thevariabletenurewas substitutedwiththenaturallogofthenumberofconsecutiveyears thefoundationhasbeenauditedbythesameauditor;surplusis definedascontinuousandsizeisdefinedinthreeadditionalways –total assetsineuros,revenuesineurosandthenaturallogof revenues.Ineachcasetheresultsdonotchange
Trang 10deContabilidad–SpanishAccountingReview(2014).http://dx.doi.org/10.1016/j.rcsar.2014.04.001
ARTICLE IN PRESS
G Model
RCSAR-32; No of Pages 12
10 B González-Díaz et al / Revista de Contabilidad – Spanish Accounting Review xxx (xx) (2014) xxx–xxx
Table 6
Logistic regression results.
Dependent variable: AUDQUAL (n = 254)
Coefficient estimate Standard error Wald Statistic Significance Coefficient estimate Standard error Wald Statistic Significance
tenure ≤ 4
tenure ≥ 6
Coefficient estimate Standard error Wald statistic Significance Coefficient estimate Standard error Wald statistic Significance tenure
tenure ≤ 2
Notes: The model has a high explanatory power, with a highly significant chi-squared test Another way to assess the performance of the maximum likelihood model is to measure the percentage of correct observations and compare it to the classification rate that would be obtained by chance [the baseline rate, which is equal to a 2 + (1 − a) 2 , where a is the proportion of audit reports which have received a qualified opinion (17.72%) in the sample] This model predicts the likelihood of getting a qualified opinion better than a random model would, with a classification improvement that ranges from 20.55 percent to 22.25 percent, which is close to the improvement of 25 percent suggested by Hair, Anderson, Tatham, and Black (1995).
The specificity (its capacity to correctly predict reports with an unqualified opinion) of the model is very good to excellent, while its sensitivity (its capacity to correctly predict reports with a qualified opinion) is good At the same time, the global capacity to correctly classify the cases, measured using the Hosmer and Lemeshow Test, ranges from 85.4 percent to 86.6 percent Pseudo R 2 measures (Cox & Snell and Nagelkerke) confirm that the model has very good explanatory power.
* Significant at 10%.
** Significant at 5%.
*** Significant at 1%.
6 Summary and conclusions
Thestudyexaminestheauditortenurerelationshipwithaudit
quality,thelatterbeingconsideredfromthepointofviewof
exter-nalusers
Usingasampleof254auditscarriedoutonSpanishstate-owned
foundationsbetween2003and2010,theresultsrevealthatsucha
connectiondoesexist.Inotherwords,itshowsthatalong
relation-shipbetweenafoundationanditsauditorincreasesthelikelihood
of the auditor issuing a clean report Auditor performance is,
however,differentin thefirstfew yearsas theprobabilityof a
qualifiedreportincreases.Inotherwords,auditquality,measured
asthelikelihoodthatanauditorwillsubmitaqualifiedopinion, increases over the first five years of the relationship and then decreases
Multivariateanalysisresultsareconsistentwithunivariateones andmayberegardedasgenerallyrobusttoalternative specifica-tionsandsensitivityanalysis
The study contributes to the literature on the relationship betweenauditortenureandauditqualityinanenvironmentwhere thereisnomandatoryauditorrotationandinasector,non-profit making,whereempiricalresearchisverylimited