conservatism – a measurement maze ho yew kee yuan yi ... · conservatism – a measurement maze...

55
Conservatism – A Measurement Maze Ho Yew Kee [email protected] Yuan Yi [email protected] Department of Accounting NUS Business National University of Singapore 15 Kent Ridge Drive Singapore 119245

Upload: phamdiep

Post on 22-May-2018

219 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

Conservatism – A Measurement Maze

Ho Yew Kee [email protected]

Yuan Yi

[email protected]

Department of Accounting NUS Business

National University of Singapore 15 Kent Ridge Drive

Singapore 119245

Page 2: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

1

Conservatism – A Measurement Maze

Abstract

Conservatismhasbeenabedrockofaccountingasitgovernstheprudentpreparationoffinancial statements to prevent over‐optimism in financial reporting bymanagement.Howeverconservatismislikesunlightwhereonecanfeeltheeffectsandwarmthofitbuthasdifficultyquantifyingormeasuringit.Thepurposeofthisstudyistoinvestigatethe robustnessof accounting conservatismmeasures in their applications todifferentindustriesandconservatismimpactassessmentstudies.Currentconservatismresearchoftenadoptsdifferentconservatismmeasurementmodelsarbitrarilyandappliesthemtocross‐industrysampleswithoutcontrolling for industrydifferences. Therearealsosignificantdisagreements among findingsof conservatism impact assessment studies.Thisstudy,throughempiricalassessmentofanextensivesampleof43,434firm‐yearsover a 23‐year period spanning 1988‐2010, aims to provide evidence to show thesignificant inconsistencies among five popular conservatism measurement models.Using Penman and Zhang’s (2002) earnings persistence regression and differentunconditional conservatism measurement models, the conclusion reached is thatPenmanandZhang’smodeldoesnotholdforallindustries.Therefore,thisstudyraisesconcerns about the arbitrary application of conservatism measurement models incurrentresearchandthereliabilityofresultsproducedbysuchstudies.

Keywords:Conservatism,accruals,measurement,earningspersistence

Page 3: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

2

Conservatism – A Measurement Maze

1 Introduction

Conservatismisoneofthefundamentalprinciplesinaccounting.Scholarshave

suggestedthatitsinfluenceonaccountingconventionisbothhistoricalandentrenched

(Watts,2003a,2003b;Basu,1997,Sterling,1970).Thewideapplicationofconservative

accounting practices before the 1900s shows that they were generally accepted as

positivepracticesthatimproveaccountinginformationquality.1

However, in recent years, scholars and practitioners have begun to doubt the

unquestioned application of accounting conservatism, especially when there is no

conclusive research evidence to substantiate the benefits of conservative accounting

practices.Theregulators’standonconservatismhasalsobeguntowaver,andinrecent

yearshasbecomelesssupportive.Forinstance,FinancialAccountingStandardsBoard

(FASB)hasreviseditspositiononthequalityoffinancialreportingbystating:

“Understatingassetsoroverstating liabilities inoneperiod frequently leads to

overstating financial performance in later periods—a result that cannot be

describedasprudentorneutral.”

FASB8,September2010

Although FASB’s position has shifted to support neutral financial reporting

instead of conservative financial reporting, nonetheless conservatism is still an

1This is consistent with Robert Sterling’s observation in 1970 that in the presence of “effervescentoptimismoftheentrepreneur…universaltendencytoovervaluetheenterprise”,theaccountantsetoutin “solidarityand stability… to combatoverstatement, heproposedunderstatement, perhapswith thehopeofstrikingabalance”.(Sterling,1970,p.256)

Page 4: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

3

entrenched concept. Some scholars have argued that accounting conservatism is

beneficial in reducing potential regulatory, litigation, contracting costs, and taxation

expenses(Watts,2003a).Others,however,claimthatconservatismisnotdesirableas

itdecreasesthequalityofaccounting informationor increasestheabilitiesof firmsto

manage their earnings (Jackson and Liu, 2010). The controversy surrounding

conservatismpoints to theneed to study the impacts of conservatismboth in theory

and empirically. This would then facilitate evaluation of the costs and benefits of

accountingconservatismparticularlyinitsroleinaccountinginformationprovision.In

1993, Ross Watts raised the awareness of research in conservatism and proposed

various research agendas2. This led to a boom in research on conservatism. Many

studiesonthemeasuresandimpactassessmentofaccountingconservatismhavebeen

published since. Despite extensive studies conducted in the past two decades, the

findings regarding the pros and cons of accounting conservatism aremore anecdotal

than conclusive. There is also limited applicability of these research findings in

accounting regulations. Therefore the justification for excluding conservatism from

accountingstandards(FASB,2010)sofar issolelyqualitativewithoutstrongresearch

backing. Inaddition,anecdotalevidenceshows inconsistencies inresultsproducedby

studiesadoptingdifferentconservatismmeasures(Chandra,2011)3.Thus,quantitative

evidence may be the missing link in the formulation of accounting policies and

regulationswithrespecttoaccountingconservatism.

Before scholars can assess the impact of conservatism, there needs to be a

commonly recognized definition of accounting conservatism and methodology to

measure accounting conservatism. Currently, there is neither a single authoritative

2AProposalforResearchonConservatismattheAmericanAccountingAssociation(AAA)Convention.3Infact,Wanget.Al.(2009)documentconvergentvalidityissuesamongstthefivemeasures.

Page 5: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

4

definitionofconservatismnoraunanimouslyacceptedconservatismmeasure.Wanget

al. (2009) report that 94% of published papers adopt one or more of the five

conservatism measures or their adaptations. The choice and application of these

conservatismmeasurementmodels,however, isoftenarbitrary. It ispossiblethatthe

different conservatism measurements are highly correlated and proxies for the

underlying accounting conservatism of a firm. However, this is only a conjecture.

Hence,theresearchresultsproducedusingdifferentconservatismmeasuresmaynotbe

comparable or reconcilable and may often lead to inconsistent results. This study

raises,andseekstoprovideananswertothequestion“areconservatismmeasurestoo

variedtobeconsistent?”. Ourstudyservestoraiseawarenessof theseverityofsuch

inconsistencies, and to spur research interest in thisareaso that futureconservatism

research will consider the importance of the conservatism measures in terms of

reliability,consistencyandconclusiveness.

This study extends the body of literature of conservatism research by

systematically assessing the robustness of different conservatism measures in their

applicationtodifferentindustriesandonstudiesofaccountingconservatism.Through

empirical analysis of extensive data comprising 43,434 firm‐years over an extended

period of 1988‐2010, tests conducted in this study provide evidence to reveal the

presence,extent,andimpactofinconsistenciesamongconservatismmeasures.

This study is organized as follows: Section 2 reviewsmajor papers and prior

research conducted in the area of accounting conservatism. Section3 articulates the

various hypotheses developed in this study while Section 4 provides the details of

researchdesign,methodology,sampledataandresults. FinallySection5presentsthe

implicationsofthisstudyandareasforfutureresearch.

Page 6: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

5

2. LiteratureReview

One of the primary reasons for the interest in accounting conservatism is its

impact on quality of earnings which is also referred to as the informativeness of

earnings.4The informativenessofearnings isoftenmeasuredby theabilityofcurrent

earnings to provide indications of future earnings (Penman and Zhang, 2002)5. A

general conclusion of the literature is that there are disagreements as to whether

conservatism has negative effects on the persistence and predictability of earnings

(Ruch and Taylor, 2011). For example, Penman and Zhang (2002) show that

unconditionalconservatismcreates‘hiddenreserves’whichincreasebiasesanderrorin

currentreportedearnings,andhencereducestheabilityof investorstopredictfuture

earnings. Adoptingthenegativeaccruals(NA)modelbyGivolyandHayn(2000),Kim

and Kross (2005), however, obtain resultswhich suggest that increasing the level of

conservatismincreasesthepredictabilityoffutureoperatingcashflows.

Thefindingsontheeffectsofconservatismoninformationqualityorasymmetry

aremixedandvaried.Somestudiesfindthatconservatismspurscompaniestoimprove

the quality of their accounting information (Fan and Zhang, 2012). Others, however,

find that conservatism leads to biases and errors in accounting reports, thereby

reducing reliability of such reports (Nishitani, 2010). However, when the news is

extremely negative, unconditional conservatism actually correlates positively with

errorsinforecast.GiglerandHemmer(2001)arguethatconservativeaccountingleads

tomoretimelyvoluntarydisclosureswhileArtiachandClarkson(2011)commentthat

themajorbenefitofconservatismisinitssignalingeffect,whichcreatesanimpression

4Informativenessofearningsismostcriticalincontractingandalargenumberofstudiesarepublishedinthiscontext(JacksonandLiu,2010;Bettyetal.,2008;Chenetal.,2007).Anotherstrainofthoughtisthatconservatismcontributestoinformationasymmetry(Liu,2010;LaFondandWatts,2008)5Expressed in another way, the informativeness of earnings is critical for valuation purposes as itprovidesinformationonthecashflowofafirm(BalachandranandMohanram,2011;Bandyopadhyayetal.,2010;Huietal.,2009;Chenetal.,2007;Zhang,2000).

Page 7: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

6

ofbetterquality information,ratherthanactually improvingthequalityofaccounting

information.

RecentstudiesbySohn(2012),Laraetal.(2011),Bandyopadhyayetal.(2010),

JacksonandLiu(2010),KimandPevzner(2010),DichevandTang(2008),LaFondand

Watts (2008) amongst others continue to generate conflicting findings on the

usefulness of conservatism in financial reporting and it seems that the jury is still

deliberating. Inaddition, there is currentlynostudyconducted to testandcompares

therobustnessofdifferentmeasurementmodels.Thereisalsolimitedefforttoaddress

theinconsistencyandinconclusivenessofresearchfindingsspecificallywithreference

to the use of different conservatism measurement methods and their impact on

differentindustries.

A reviewof the literature on accounting conservatismhas revealed twopertinent

issuescontributingtotheinconsistentfindings.Theyare:

a. Definitionsofaccountingconservatismarevagueandarbitrary,and

b. There is an absence of a single, comprehensive, and authoritative model for

quantificationandmeasurementofaccountingconservatism.

The next sub‐section will briefly discuss the definitions of conservatism and the

majormeasurementmethodsusedincurrentresearch.

AuthoritativeDefinitionofConservatism

Accountingconservatismcanbedefinedvaguelyas‘exerciseofprudence’or‘the

exercise of caution’ (Givoly and Hayn, 2000). A popular adage describes accounting

conservatism as “anticipate no profits, but anticipate all losses” (Watts, 1993). The

abovedefinitions,whilevagueandall encompassing,arenothelpful in settingup the

theoretical framework for research in accounting conservatism. For instance,

Page 8: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

7

conservatism in accounting needs to be clearly distinguished from management or

businessconservatism,whichoftenreferstoprudentrisktakingbehaviorforoperating

orbusinessdecisions. Accountingconservatismisaboutprudenceinthepresentation

of accounting information and statement of accounting numbers, as for example,

recognitionandquantificationofimpairmentchargesintheincomestatementduetoa

prolongedandsubstantialdownwardadjustmentinthevaluationofanassetwhichhas

already happened. The economic or trigger event may or in fact may not have

happened. For instance, provision for doubtful debts pertains to judgment on the

collectability of debts where the uncollectibility of the debt has not happened. The

quantum of impairment charge or provision of doubtful debts will invoke different

degreesofaccountingconservatismasintendedbytheprepareroffinancialstatements.

The GAAP’s definition of conservatism principle is more precise in clarifying

prudence to be a choice to state a lower income and net assetwhenever the choices

exist.Thismeansalowerlevelofrevenueorahigherlevelofexpensestoberecognized

whenever there is a choice or judgement to be made. However, it describes

conservatism in a relative sense, in comparing the effects of two ormore acceptable

accountingchoices.Itdoesnotprovideasolutiontothemeasurementofconservatism

as often the benchmark is not defined or is elusive or a matter of judgment. Some

scholars fill in themeasurement gap by proposing the benchmark formeasurement.

One of themost recently accepteddefinitions is “systematic undervaluationof equity

relative to economic value” (Watts, 2003a; Givoly et al., 2007). This definition is

supportedandfurtherclarifiedbyFelthamandOhlson(1995)tobethedownwardbias

ofbookvalueofafirm’sequityascomparedtoitsmarketvalue,withmarketvalueas

the benchmark measure of the true economic value of a firm’s equity. Under this

Page 9: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

8

definition,anyformofearningsreportthatresultsinthebookvalueoftheentitybeing

lowerthanitsmarketvaluewouldbeconsideredconservative.

Thedefinitionofaccountingconservatismisalsomademorecomplexbyadding

atimedimensiontoaccountingconservatism.Accountingpracticesprescriberulesand

guidance not only to determine the amount to be recognized but also the timing of

recognition.Anumberoftextbookwritersincorporatedthetimedimensionintotheir

conservatism definition, mainly stating that accounting conservatism is the slower

recognition of income as compared to recognition of expenses (Wolk et al., 1989;

Davidsonetal.,1985).6Alternatively,accountingconservatismistheunder‐reportingof

earnings as compared to cash flows from operations over a defined period of time

(Givolyetal.,2007).Thiscarriesthenotionthattherewillbeconvergencebetweencash

flowsfromoperationsandnetincomeoverareasonablylongperiodoftime.

Asaresultofthecomplexityandthelackofaconceptuallywell‐developedand

authoritativedefinition,manyresearchershavechosentoprovidetheirownversionsof

thedefinitionofconservatismtojustifydifferentchoicesofconservatismmeasuresand

proxies. Basu defines conservatism as recognition of ‘bad news’ in a more timely

manner than ‘good news’ in reported earnings (Basu, 1997).7Feltham and Ohlson

(1995), and Beaver and Ryan (2000) choose to define conservatism as persistently

reporting net assets at amounts lower than their market value. Penman and Zhang

(2002) focus on the biases in the net operating assets level due to accumulation of

layers of reserve (“hidden reserve”) in their definition of accounting conservatism.

Currently, there is no single authoritative definition that is applied across accounting

6ThistimedimensionofconservatismwasoperationalizedbyBasu(1997)andhenceforthbecamethemostpopularmeasurementofconservatism.7Basu’s interpretation implies that conservatism can be represented by the “systematic differencesbetweenbadandgoodnewsperiodsinthetimelinessandpersistenceofearnings”.Thisiscurrentlythemostpopularrepresentationofconservatism(Wangetal.2009).

Page 10: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

9

research. Without such agreement and consensus, comparing findings of different

conservatismstudiesisoftenlikecomparingappleswithoranges.

MeasurementsofConservatism

An immediate result of the lack of an authoritative definition for accounting

conservatism is the proliferation of measurement methods. Various definitions of

conservatismemphasizedifferentaspectsofconservatisminaccountingpracticesand

hence, lead to confusions over the applicability and biases of different methods of

measurementofconservatism.

Onesuchconfusionstemsfromthelackofacleardistinctionandunderstanding

ofconditionalandunconditionalconservatism.Conditionalconservatism,asdefinedby

Basu(1997),relatesmarketreactionstoaparticulargoodorbadaccountingearningor

cash flow news. It is event‐based and involves an external trigger of circumstances,

namely, share price returns and thus the term “conditional”. 8 Unconditional

conservatism, on the other hand, does not involve such a trigger (Ruch and Taylor,

2011).Itisoftenimplicitintheinitialrecognitionofrevenuesandexpenses,andassets

andliabilities(BeaverandRyan,2005),suchasimmediateexpensingofResearchand

Development (R&D) costs. These two types of conservatism are largely studied

separately, producing a spectrum of conservatism measures that are essentially

measuring different purported characteristics of conservative accounting practices.

Theirdistinctionhasnotbeenwellunderstooduntil recentlywhereBeaverandRyan

(2005) model the linkage between these two measures through the effect of

understatingthenetassets.

8However,inrecenttimes,therearenumerouscriticismsaboutthevalidityconstructofBasu’smeasureanddifferentmodificationsweremadetoaddresstheconstruct issue(Balletal.,2011;PatatoukasandThomas,2011;Dietrichetal.,2007;Givolyetal.,2007).

Page 11: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

10

Table 1 summarizes the five commonly used measurement models9in the

currentstateofresearch(Wangetal.,2009).Basu’s(1997)asymmetrictimeliness(AT)

conditional conservatism measure is most popular (69%) amongst the 52 papers

surveyedbyWangetal(2009)followedbythemarket‐to‐book(MTB)model(25%).

[Table1abouthere]

There have been attempts to reconcile unconditional and conditional

conservatismmeasures(Ryan,2006). Dietrichetal.,(2007)findthattheATmodel is

associatedpositivelywiththeMTBmeasureoverlongperiodandnegativelywithMTB

measure over short period. It is also postulated that conditional conservatism and

unconditionalconservatismarenotcompletelyindependentofeachother(Ryan,2006).

Therefore, reconciling the two typesof conservatism requires amodelmore complex

thanasimplesummationorsubtractionofconditionalandunconditionalconservatism

measures.Theasset‐basedmodelproposedbyBeaverandRyan(2005)tomergethese

two typesofmeasures isyet tobe tested for its effectiveness.Themore fundamental

questionsare: When isconditionalconservatismappropriate foruse inconservatism

impact assessment study as against unconditional conservatism? Are conditional and

unconditional conservatismmeasuring the same underlying accounting conservatism

practicesoffirms?

All the above measurement methods are subjected to different criticisms of

which the three main criticisms are: a) choice of proxies problem: b) confounding

problemandc) incompletenessproblem.Forthechoiceofproxiesproblem,there isa

lackofaconsistentconceptualframeworkinthechoiceofproxiesusedtoconstructthe

9SeeAppendixAforformulaandthevariablesusedincomputingthesefiveconservatismmeasures.

Page 12: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

11

conservatism measure. For example, Basu (1997) uses earnings while Ball and

Shivakumar(2005)useoperatingcash flow.There isnocompelling theory tosuggest

that both measures are complementary measures of accounting conservatism. Even

though earnings and operating cash flow are highly related, these proxies may not

measure thesamephenomenonwhich theconservatismmeasure is trying tocapture.

Theconfoundingproblemistargetedattheuseofmarketsharepricesorannualstock

returnswhichsubjecttheconservatismmeasurementtomarketwidefactorsorother

intervening events which are price sensitive10. These confounding events may have

nothingtodowithaccountingconservatism,forexample,theMTBmodel11.Finally,for

theincompletenessproblem,thisistheclassicomittedvariableproblemwherethereis

no assurance that the relevant factors which measure accounting conservatism are

adequatelycapturedbytheproxiesusedinthemeasurementmethod.12Forexample,in

the HR model by Penman and Zhang (2002), impairment losses and provision for

doubtfuldebtswhicharehugeaccountingconservatismplaygroundsareomittedfrom

themeasurement.Figure1providesadiagrammaticsummaryof thevariouspossible

factorswhichcontributetotheconservativeaccountingpracticesofafirmwhichmay

notbecurrentlycapturedbyallthemeasurementmodels.Therefore,itwouldseemthat

each measurement has its unique failings and there is currently no consensus as to

whichmeasurementmodelisdominantorconceptuallymostrobust.

Duetothedifferentchoiceofproxiesandtheabovethreeproblems,itispossible

that the results invarious studies couldbedrivenby themeasurementmethodused.

10See Manuel and Manuel (2011); Patatoukas and Thomas (2011); Gotti (2008); Roychowdhury andWatts(2007);Givolyetal.(2007).11MTBratioisarguedtobebiasedupwardinmeasuringthelevelofconservatisminafirmastheeffectof economic rent in depressing book value is not recognized and separated from the effects ofconservatism(RoychowdhuryandWatts,2007).12For example, Ball etc al. (2011) address this omitted variable problem in Basu’s measure byintroducingafixed‐effectintotheregression.

Page 13: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

12

Thisisnotincludingthepossibleconfoundingindustryeffectwhereadifferentindustry

may have a different degree of accounting conservatism because of its unique

accounting practice. For example the degree of conservatism asmeasured by theHR

modelwhichincludesR&Dwillbeverydifferentforfirmsinindustrieswithlittleorno

R&DversusthosewhichareR&Dintensive.

In response to these limitations, some researchers have chosen to applymore

thanonemeasurementmodelasrobustnesstests.Wangetal(2009)reportthat40%or

21outof52studiesreviewed,usemorethanonemeasurementmodel.Inaddition,the

choiceofmeasurementmodel isalsoveryarbitrary. Forexample,nopaperreviewed

justifiesthechoiceofconditionalconservatismmeasureoverunconditionalmeasure.It

is also noted that researchers do not use an industry adjusted conservatism

measurementmodel. Insteadmostwould simply use an industry dummy variable in

theirregressiontoaddressthepossibleindustryeffects.Consideringthedifferencesin

the nature of accountingmethods for different industries, it is possible that different

conservatismmeasuresaresuitablefordifferentindustries.However,currently,there

is no study which tests the applicability of these measurement models to different

industriesorhowtocontrolfortheindustryeffectsinamorerobustmanner.

3 HypothesesDevelopment

Thissectionfocusesondevelopingthehypothesesthatconservatismmeasures

arenotrobustintheirapplications. ThefivemeasuresinTable1arechosenbecause

they are the most widely applied conservatism measures in published papers on

accounting conservatism. The robustness of these measurement models will have

implicationsonthereliabilityoftherecentconservatismimpactassessmentstudies.

Page 14: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

13

Outofthefiveconservatismmeasures,ATandACCFmodelsmeasurethelevelof

conditional conservatism, whereas HR, MTB, and NA models measure the level of

unconditional conservatism in firms. Since the relationship between conditional and

unconditionalconservatismisstillunderinvestigation,comparisonsbetweenthesetwo

groupsofconservatismmeasureswillhavetobetreatedwithcare.

IndustryConfoundingEffects

Different industries may have different unique accounting practices which

render inter‐industry comparison of conservatism measurements difficult.

Conceptually,differentindustrieshaveuniquecharacteristicsthatarelikelytoinfluence

the type and extent of conservatism in their accounting practices. In terms of

conditional conservatism, different industries could bemeasuredunfairly as they are

sensitive to different kinds of news that impact the firms over different lengths of

investmenthorizon. Forexample,oilandgas firmsaresensitive togeneraleconomic

newsthatoftenimpactstheearningsperformanceofoilandgasfirmsoverlongterm,

whilehightechnologyconsumergoodsaremoresensitivetofirmspecificnewsthatare

morelikelytohaveshorttermimpactsonthefirms’earnings.Hence,itisimportantto

verify whether industry effects have significant implications for the different

conservatismmeasurementmodels.13

Thedifferenceinthedurationofinvestmentandearningspersistenceislikelyto

affect the extent of the incremental timeliness of recognizing ‘bad news’ over ‘good

news’ under the conditional conservatism measures. Different unconditional

conservatismmeasuresmaybeappropriateforoneindustrybutnotforothersbecause

13However, a majority of empirical conservatism impact assessment research applies conservatismmeasures arbitrarily to cross‐industry data. Chandra (2011) focusing on the technology sector is anexception. A general control for industry effect is to use industry dummy variables in the regressionanalysis.

Page 15: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

14

of the relative importance of their income statement and balance sheet items. For

example, firms from high technology industry which are light on assets because of

immediate expensing of R&D costs and heavy importance of human resource capital

could be measured wrongly when compared with capital intensive industries like

utilities that have a significant amount of long term assets. Different conservatism

measureswillcapturedifferentaspectsofhightechnologyfirms.Hence,itispostulated

that:

H1: Different conservatism measurement models will rank different

industriesdifferently.

ImpactAssessmentStudyUsingDifferentConservatismMeasures

Measurementofconservatismisanintegralandimportantpartofconservatism

impactassessmentstudies.Empiricalresearchinthisareaoftenregressconservatism

scoreagainstproxiesforearningsqualityandhencedifferentconservatismscoresmay

impacttheresultsofsuchregressionsdifferently.Only15%ofthepublishedstudies(8

of52)usedthreeormoreconservatismmeasuresforrobustnesstestingand60%ofthe

studies(31of52)usesonlyonemeasurementmodel (Wang,etal.,2009). Ifdifferent

measurement models produce different results, the reliability of studies using one

measurementmodelwillbecomequestionable.

We use Penman and Zhang’s (2002) study on the relationship between

accounting conservatism and earnings persistence as the benchmark conservatism

impactstudy. Ourstudydoesnotexamine thesoundnessof theproposedtheoretical

framework used by Penman and Zhang. Rather, it investigates whether the same

conclusionbetweenaccountingconservatismandearningspersistencyasdocumented

by Penman and Zhang (2002) can be established using different unconditional

Page 16: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

15

conservatism measurement models. We replicate Penman and Zhang (2002) and

substitute their unconditional conservatism Q‐Score with the MTB and NA

unconditional conservatism scores.14The robustness of the different conservatism

scoreswillbetested.Thefollowingstatesthenullhypothesisforourstudy:

H2: Differentunconditionalconservatismmeasureswillproducesimilarresults

asfoundinPenmanandZhang’s(2002)earningspersistencestudy.

4. DataandResults

Data

DataforthisstudyaretakenfromCOMPUSTATAnnualIndustrialandResearch

files including non‐survivors and holding period return from CRSP monthly

stock/security files fromtheperiod1975 to2010.The testofhypothesisH2requires

the replication of Penman and Zhang’s study. To be consistent with their research

methodology,acomparativepoolofsamplesfrom1975to1997similartoPenmanand

Zhang’sisused.Thisisaugmentedbyadditionalfirm‐yearstakenfrom1998to2010.

TheinitialqueryonCOMPUSTATyielded96,085firm‐yearsacrossallindustries

from1975to2010.The96,085firm‐yearswerethensortedaccordingtotheir2‐digit

SIC, and the top ten industries with the most number of available firm‐years were

chosen foranalysis.Variouswholesaleand retail trading companiesunder the2‐digit

SIC50to59werecombinedtoformasinglewholesaleandretailindustry.Wejustified

this grouping by the similarities in their accounts such as near zero R&D and high

inventory turnover. This resulted ina totalof65,786 firm‐years (68.5%of the initial

sample)from1975to2010.Furtherfilteringrequiredremovaloffirm‐yearswhereany

14WedidnotincludetheconditionalscoreintheanalysisbecauseoftheconceptualdifferencebetweentheconditionalandunconditionalconservatismmeasuressinceWangetal.(2009)documentsignificantseparationbetweenthesetwotypesofconservatismmeasures.

Page 17: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

16

of the fiveconservatismscorescouldnotbecomputedmeaningfully.The finalsample

consistedof43,434firm‐yearsfortenindustries.Thefilteringprocess,thetenselected

industries,theirSICandthenumberoffirm‐yearsarefoundinTable2.Eachfirmyear‐

datacontains27differentvariablesusedtocomputethefiveconservatismscoresand

core‐RNOA.Thevariablesandtheformulaforcomputingthefiveconservatismscores

arefoundinAppendixA.

[Table2abouthere]

While the original sample includes firm‐years from 1975 to 2010, the final

sample consists only of firm‐years from1988onwards because of the extensive data

requirementsfortheconstructionoftheconservatismscores. Outofthe43,434firm‐

years in the final sample,19,931 firm‐years (45.9%of thesample)arebetween1988

and1997,replicatingthedataperiodPenmanandZhangusedintheirstudy15andthe

remaining23,503firm‐years(54.1%ofthesample)arefrom1998and2010.

Results

FollowingPenmanandZhang’s(2002)dataapproach,firm‐yeardataweretaken

from1975‐1997.Anextendedperiodofdatafrom1988‐2010wasanalyzedtofurther

test the robustness of the test results across a longer period of time, as well as to

provideanupdatetotheresultsofPenmanandZhang(2002)byincludingmorerecent

firms’ information. Table 3 summarizes the descriptive statistics of nine selected

variables16forthetwotimeperiodsofanalysis,1988‐1997(10years)and1988‐2010

(23years).

15PenmanandZhang(2002)useasampleof29,796firm‐yearsfortheirfinalregression.16Thereare27variablesusedinthisstudy.Forbrevityinreporting,theseninekeyvariablesareselected.Thedescriptivestatisticsofthe27variablesareavailableuponrequest.

Page 18: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

17

[Table3abouthere]

FromTable3, it is evident that firm‐years frombothperiodsof analysis show

significant positive skewness in most of the major accounts. This means that data

representation isbiasedtowards larger firms. Thedegreeofskewnessremains fairly

consistentacrossbothperiodsofanalysis,showingthattheircompositionsoffirmsof

differentsizesarelikelytobeconsistent.However,themaximumvalueofeachselected

account for the period 1988‐2010 is much higher than the maximum value of the

similar accounts for theperiod1988‐1997. This led to a significantdifference in the

means of the key variables between 1988‐1997 and 1988‐2010. Market value, total

asset (AT),andexpenseaccounts(DP,XAD,andXRD)nearlydoubled in theextended

periodshowingthatfirmshavegrownsignificantlyinsizeandassetbasesubsequentto

1997. Thishasasignificantimplicationforaccountingconservatismasgrowingfirms

oftenaccumulatehiddenreserveswhenincreasedassetsareexpensedinsteadofbeing

capitalizedandamortizedoverfutureyears(PenmanandZhang,2002). LIFOreserve

remainsrelativelysimilar(6.600intheearlierperiodversus7.051inthefinalsample)

probablyduetothepopularityofFIFOinventoryaccountinginrecentyears,asoutside

theUS,LIFOisoftennotallowed.Onthebasisofthevariationsofaccountsovertime,

assetbasedconservatismmeasures,especiallyHR,arelikelytoobserveanincreasein

conservatism since the growth in firm assets is likely to increase the size of hidden

reserves. Therefore the changes in level of conservatism over the yearsmay not be

measuredconsistentlyacrossdifferentmeasurementmodels.Hencetheconsistencyof

conservatismmeasurement models when applied to different industries was further

analyzedoveranextendedperiodof1988to2010.

Page 19: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

18

Table 4 provides the breakdown of selected accounts of firms in Table 3

accordingtoindustry17. Foreachindustry,eachaccountmeaniscomparedtothefull

samplemeanandpresentedasapercentage.Eachindustryhasclearlydifferentlevels

of accounts. For example, Oil & Gas (13), Business Services (73) have typically low

inventory balances (almost 20% of overall sample average) while Transportation

Equipment (37) and Trade (50‐59) have on average very high balances (295% and

163%of overall samplemean respectively). On the other hand, Trade (50-59) has

nearzeroR&Dexpense(1%ofoverallsamplemean)whileTransportationEquipment

(37)incurshighR&Dexpense(273%ofoverallsamplemean).Differentmeasurement

models consisting of different combinations of accounts are hence likely tomeasure

conservatism of these companies differently. In order to test the significance of the

differences in accounts between industries, ANOVA was performed to test if inter‐

industryvarianceineachvariableissignificantlyhigherthanintra‐industryvariancein

the variables. The results show that for all 9 variables in Table 4, inter‐industry

varianceissignificantlyhigherthanintra‐industryvariance.

[Table4abouthere]

The results show a strong support for the grouping of firms according to

industries, as the variables vary significantly from industry to industry. The

conservatismmeasurementsassociatedwiththeseaccountsarealsolikelytovaryfrom

industrytoindustry.Asnosingleconservatismmeasureincludesallrelevantvariables,

their suitability and comprehensiveness in measuring a firm’s level of accounting

conservatismislikelytovaryfromindustrytoindustry.

17Duetolimitationofspace,thenumbersfor1988‐1997arenotpresentedbutareavailableonrequest.

Page 20: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

19

ResultsoftheRobustnessofConservatismScoresatIndustryLevel

To investigate H1, accounting conservatismmeasurementmodels were tested

for their consistency in their application to different industries. Conditional

conservatismscores,asymmetrictimeliness(AT)andasymmetricaccrualstocashflow

(AACF) were computed according to the regression model described in Appendix A.

Unconditional conservatism scores, including market‐to‐book ratio of equity (MTB),

hiddenreserves(HR),andnegativeaccruals(NA),werecomputedforeachfirm‐year.

TheregressionresultsfortheATandAACFmodelsaresummarizedinTable5.

Table 5 shows significant difference between timeliness of recognition of good news

andbadnewsforamajorityof industries. Regressioncoefficientsobtainedunderthe

AT model (ranges from 0.109 to 0.700) show less variation between industries as

comparedtotheAACFmodel(rangesfrom‐0.985to1.9688).Infact,theresultsofthe

AT model are relatively stable between industries and across time18. For the AACF

model, for theperiod1988‐1997(PanelC), fouroutof ten industries’ (SIC20,28,38,

48) β1 is not significant at 10% level of significance. This weakens the model’s

applicabilityontheindustrylevel. Inaddition,forthesameperiod,twooutoftheten

industries(SIC37and50‐59)aresignificantlynegative,suggestingthatfirmsrecognize

badnewslesstimelyascomparedtogoodnews,whichisinconsistentwiththeresults

obtainedbytheATmodel. TheresultsinPanelDfortheperiodof1988‐2010forthe

AACFmodel are notmuch better. This shows that between conditional conservatism

models,themeasurementresultsdonotalwaysagree.

[Table5abouthere]

18TheexceptionisSIC48fortheperiod1988‐2000whichshowsinsignificantresults.

Page 21: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

20

In order to investigate the extent of inconsistency between conservatism

measurementmodels in their application to different industries, the industries were

rankedbasedonthescoreobtainedusingeachconservatismmeasurementmodel.Raw

conservatismscoresandrankingofindustriesforboth1988‐1997and1988‐2010are

summarized inTable6. The scoresandrankswere furtheranalyzed to investigate if

there was any agreement between conservatism models in their measurement of

industryconservatism.

[Table6abouthere]

Table6showsthattherankingsofconservatismmeasuresforthe10industries

are very different. For both periods, there is no single industry that is ranked

consistentlyacrossallfiveconservatismmeasures. Someofthedifferencesinranking

canbe stark. For example, inPanelA, Communications (48) is rankedas the second

leastconservativebyATscorebutitisrankedasthemostconservativebyNAandMTB

scores. From thedescriptive statistics (Table5), it canbe seen thatCommunications

(48) has extremely high accruals, which is 834% of full sample mean, and high

depreciation,whichis788%offullsamplemean.Henceitsaccrualbalanceislikelyto

causeanupwardbias in itsconservatismscoremeasureusingnegativeaccrualsonly.

Evidently,ahighdepreciationlevelindicatesahighlikelihoodofadeflatedbookvalue

through accelerated depreciation, and hence leads to an upward biased MTB

conservatismscore.ThemostconsistentlyrankedindustryistheOil&Gas(13)asitis

rankedasmostconservativebyall threemeasures:ATscore,HRscoreandNAscore.

PanelBprovidesanevenmoreconfusingrankingasonlyChemical(28)isrankedbyHR

andMTBasthemostconservative.Theaveragerankoftheindustriesrangedfrom2.6

to7.6(2.4to8.2)inPanelA(B).ThestandarddeviationsforPanelA(B)rangedfrom

Page 22: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

21

1.3to3.6(0.8to4.0). Figure2providesapictorialsummaryoftherankingoftheten

industriesaccordingtoeachofthefiveconservatismscores.Thereisnoconsistencyin

thepatternsofrankingforthesetenindustriesinFigure2.

[Figure2abouthere]

Overall, there is significant disagreement in conservatism score and ranking

among the five conservatism measurement models. Ideally, when all conservatism

measuresareconsistent,Figure2shouldshowhorizontallinesthatareparalleltoeach

other. 19 This means that if a particular industry has high or low accounting

conservatism, its rankwouldbe consistentlyhighor lowacrossall fivemeasurement

models. Howeverthisisnotthecase.Therefore,thereisstrongsupportforH1. Both

conditionalmeasuresandunconditionalmeasuresranktheindustriesverydifferently.

There is no clear relationship identified amongst the conditional or unconditional

conservatismmeasures.Thenextsectionpresentstheresultsofclassificationtests.

ClassificationOverlapforUnconditionalandConditionalMeasures

To conduct classification tests, each industry is classified under high

conservatism (top 3 in conservatism score) or low conservatism (bottom 3 in

conservatism score) according to each conservatism model. If more than one

conservatismmodel classifiesaparticular industry in thesamecategory,highor low,

thisisconsideredanoverlapinclassification. Thegreatertheoverlap,thegreaterthe

agreementamongstdifferentconservatismmodels.Figures3and4presenttheresults

19The same inconsistencies are observedwhenwe compared the conditional conservatism scores ATwithAACFortheunconditionalconservatismscores(HR,NAandMTB)separately.Figuresareavailableonrequestfromauthors.

Page 23: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

22

of classification for the periods 1988‐1997 and 1988‐2010 for the conditional and

unconditionalconservatismmeasuresrespectively.

[Figures3and4abouthere]

FromFigure3PanelA,fortheperiod1988‐1997,forhighconservatism,twoof

the three industries (SIC 35 and 36) are classified similarly. However for the low

conservatismclassification,onlyoneindustry(SIC50)isclassifiedassuch.ForPanelB,

fortheperiod1988‐2010,onlyoneindustryisclassifiedconsistentlyforbothhigh(SIC

73)and low(SIC50)conservatismcategories.TheATmodelof classification forhigh

conservatismistotallydifferentbetweenthetwoperiods,namely,SIC20,35and36for

1988‐1997, and SIC 13, 37 and 73 for the period 1988‐2010. However the low

conservatismclassificationisrelativelystableforbothperiodswiththesameindustry

classified as low conservatism in both measures in both time periods (SIC 50).

Therefore,atbest,theconditionalconservatismmeasuresofATandAACFgiveasimilar

classification for two of the three industries (67% classification consistency), and at

worstoneinthreeindustries(33%).

From Figure 4 Panel A, for the period of 1988‐1997, there is no consistent

rankingbyallthethreeunconditionalconservatismmeasuresforanyindustryasahigh

conservatism industry. In fact, atbest, forany twounconditionalmeasures,onlyone

industry isclassifiedsimilarly.Forexample,usingNAandMTB,onlyCommunications

(48)isclassifiedsimilarlyashighconservatism.Forthelowconservatismclassification,

only the Oil & Gas (13) is classified as low conservatism in all three unconditional

measures. The same is observed for the period 1988 to 2010. However, a unique

differenceisthatforbothMTBandHRmeasures,thethreeindustriesareconsistently

Page 24: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

23

classifiedaslowconservatism(SIC13,37and50).TheVenndiagramsinbothPanelsA

and B suggest that consistency in classification is limited except for the low

conservatismfortheOil&GasindustryandbetweenMTBandHRmeasures.

Theclassificationoverlapsamongstall fiveconservatismmeasurementmodels

arealsoexaminedandtheresultsaresummarizedinTable7. Thereisnounanimous

agreementontheclassificationofanyindustryforallfivemeasures,namely,thecolumn

with“5overlaps”isanullset.InPanelsAandB,theclassificationofOilandGas(13),

andTrade(50) iscomparativelythemostconsistent for theperiod1988‐1997as low

conservatism by four of the fivemeasures. In general,most industries are classified

similarly only by two of the five measures. Panels C & D provide an almost similar

conclusion except where four measures classify Business Services (73) as high

conservatismandTrade(50)aslowconservatismsimilarly.

[Table7abouthere]

Apartfromclassificationtestconductedatindustrylevel,classificationtestwas

alsoperformedat firm‐year level forunconditional conservatismmeasures forwhich

firm‐specificscoresareavailable.Therationaleforconductingtheadditionalfirm‐year

analysis is tounderstandthedegreeofoverlap intheclassificationof individual firm‐

years. Figure 5 summarizes the results obtained for the classification overlap of all

firm‐yearsinthefinalsample.FromPanelsAandB,fortheperiod1988‐1997,forhigh

(low)conservatism,only9.8%(12.7%)ofthesampleissimilarlyclassifiedbythethree

measures.FromPanelsCandD,fortheperiod1988‐2010,forhigh(low)conservatism,

only10.2%(11.9%)of the sample is similarly classified.The results suggest that less

Page 25: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

24

than13%ofthefirm‐yearsareclassifiedsimilarlybythesethreemeasuresashighor

lowconservatism.

[Figure5abouthere]

Table8summarizestheclassificationoverlapoffirm‐yearsineachindustry.Oil

&Gas(13)hasthemostconsistentclassificationforhighandlowconservatismforboth

periods, ranging from 24% to 29%. For the rest of the industries, the consistency of

classificationrangesfromalowof4%forCommunications(48)forhighconservatism

fortheperiod1988‐1997andamaximumof15%forChemicals(28),Electronics(36),

Transportation Equipment (37) and Communications (48) for low conservatism. The

levelof consistency isvery lowas theoverall consistencyclassificationranges froma

minimumof1.2%(4%*30%)toamaximumofmerely8.7%(29%*30%)ofthetotal

firm‐years.

[Table8abouthere]

Overall, the classification overlap among the three unconditional conservatism

measures is consistently lowat below10% forbothhigh and lowconservatism.This

suggests that when conservatism measures are applied to firm‐year data across

industries,onlyoneoutofeverytenfirm‐yearsinvestigatedismeasuredandclassified

consistentlyacross threeunconditionalconservatismmeasures. This findingseverely

weakens the reliabilityofusingonlyoneconservatismmeasure todeterminea firm’s

level of unconditional conservatism. However, using more than one conservatism

measure may lead to conflicting results. In Penman and Zhang (2002), firms are

classifiedintohighconservatism,moderateconservatismandlowconservatismbased

ontheHRscore.Theportfoliosoffirmsclassifiedundereachcategoryarelikelytobe

Page 26: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

25

differentifadifferentmodelischosentocomputetheconservatismmeasure.Willthe

useofadifferentconservatismmeasureresultinadifferentconclusionfortheirstudy?

Thisquestionisaddressedinthenextsub‐section.

AnalysisofResultsofRobustnessTestBasedonPenmanandZhang’sRegression

In the current study, therewas sufficientdata to calculateQ‐Scores for36,839

firm‐yearsfortheperiod1988‐2009(22years).Thefrequencycurveofcore‐RNOAand

Q‐Scorevariablesshowedthattherewereafewbutveryextremevalueoutliersonboth

ends.Extremevalueoutlierswereremovedtoobtainreliableregressionresults.Since

the frequency curvesof variablesdisplayedvery thin tails, dataoutsideone standard

deviationforeachvariablewereremoved.Thefilteredsamplecontained33,943firm‐

years, whichwas 92% of the final sample. Table 9 summarizes the key statistics of

Q_HRscoreof the filteredsample,andthesevaluesarepresentednext to theQ‐Score

statisticscomputedbyPenmanandZhang.Althoughdatawereretrievedfromthesame

source(CRSPandCOMPUSTAT),andcomputationofQ‐Scoreinthisstudyfollowedthe

methodologydescribedbyPenmanandZhang,differencesindatawereexpecteddueto

thedifferenceinthedataperiodsandtheextensivedatarequirementsinconstructing

therestoftheconservatismmeasures.

[Table9abouthere]

Earningspersistence regressionswereperformedusingQ_HRaccording to the

equationusedbyPenmanandZhangasthecontrolfordatavariation.Ifthedataused

inthisstudywerenotsignificantlydifferent fromthatusedby PenmanandZhang, it

wasexpectedthatthesameresultsastheregressionsperformedbyPenmanandZhang

Page 27: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

26

wouldbeproduced.ToreplicatePenmanandZhang’sstudy,22annualcrosssectional

regressionswereperformedusingcomputedQ_HRforthefilteredsample. Coefficient

forQ_HRwasobtainedforeachyear. One‐samplet‐testwasthenperformedtotestif

themeanofthe22coefficientswassignificantlydifferentfromzero.AsshowninTable

10,theresultsofthet‐testforQ_HRcoefficients inthisstudyaresignificantat1%,as

werethoseobtainedbyPenmanandZhang.

[Inserttable10abouthere]

The significance of Q_HR coefficient obtained in this study also suggests that

PenmanandZhang’smodel isrobustacrossperiodsandstillappliestoamorerecent

timeperiodalthoughthemagnitudeofthecoefficientsissmallerinourstudy.However,

whenthesametestswereperformedforQ_MTBandQ_NA, theresultsobtainedwere

notsignificant.Hence,furtherregressionswereperformedforeachindustrytofurther

investigatetherelationshipbetweendifferentQ‐ScoresandnextyearCore‐RNOA.

Table 11 summarizes the results of the three regressions for each Q‐Score by

runningasinglepooledregressionforallfirm‐yearsforthesample. Theresultsshow

thatcoefficientsofQ_HRandQ_MTBaresignificantat1%.Q_NAmodelshowsaslightly

weakerresultbutstill significantat5%. Theoverall resultsproducedbydifferentQ‐

Scoresarefairlyconsistent.However,wenextexaminedtheregressionsperformedat

industrylevel.

[InsertTable11abouthere]

StackedregressionswereperformedforeachoftheQ‐Scores.Dummyvariables

wereassigned to industries fornineoutof the ten industries. Business (73)was the

Page 28: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

27

baseindustryanditsregressioncoefficientwassubsumedintheconstantterm.TheQ‐

Score coefficientswereobtained foreach industry and summarized inTable12. The

resultswerevery inconsistentacross industries. ForQ_HR,although itscoefficient in

earnings persistence regression was shown to be significant in both annual cross‐

sectionalregressionandsinglepooledregression,itwasnotsignificantforthreeoutof

nineindustries(37,38and48).Inaddition,twoofthenineindustries(13and50)had

significantlynegativecoefficients.ForQ_MTB,onlytwooutofnine industries(20and

50)coefficientsweresignificant. Finally,forQ_NA,allcoefficientswerenotsignificant

evenat5%levelofsignificance.Suchinconsistentresultsraisetwoissues.First,using

different Q‐Scores leads to different conclusions about the relationship between

conservatismandone‐year‐ahead RNOA. AdoptionofQ_HRscore leads toapositive

conclusionwhile adoption of Q_NA score leads to a negative conclusion. The results

obtainedusingQ_MTBaremixed. Secondly, the regression results arenot consistent

acrossindustries.ThismeansthattheabilityofQ‐scoretoimprovethepredictabilityof

nextyearCore‐RNOAmayvaryfromindustrytoindustry.

The conservatism measures are not robust in their application to different

industries and to Penman and Zhang’s earnings persistence regression. The results

obtainedinthisstudyshowthatconditionalandunconditionalconservatismmeasures

score and rank different industries very inconsistently and hence provide strong

support for H1. For the second part of the study, although results from the single

pooled regression produce significant results for all three Q‐Scores, results obtained

using annual cross‐sectional regressions and industry stacked regressions show that

therearesignificantinconsistenciesintheresults. Hence,onthebasisofthefindings,

thisstudyrejectshypothesisH2.

Page 29: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

28

5. Conclusion

As discussed, there are three main problems with current research on

conservatism. First, definitions of accounting conservatism are vague and arbitrarily

interpreted. Secondly, there is no authoritative model for quantification and

measurement of accounting conservatism. Thirdly,measurementmodels are chosen

andappliedarbitrarilyinempiricalresearchtoassesstheimpactofconservatism.The

above‐mentioned issues have led to significant disagreements in current research

findings. Our study seeks to document the potential inconsistencies among different

conservatismmeasurementmodelsandtheirimpactonconservatismimpactstudies.

Thisstudyhenceconductedtwoteststoinvestigatetherobustnessoffivemost

influentialconservatismmeasurementmodels.Thefirsttestedtherobustnessofthese

measurementmodelsintheirapplicationtodifferentindustries.Thesecondtestedthe

ability of different unconditional conservatism measurement models to produce

consistentresultsinPenmanandZhang’searningspersistenceregression.

The results suggest that both conditional conservatism measurement models

and unconditional conservatism measurement models rank the industries

inconsistently.Thisimpliesthatconservatismmeasurementsatindustrylevelbasedon

a single model are highly unreliable, and comparisons between different industries

measuredusingdifferentmodelscannotbereliablymade.

The results of the robustness test based on Penman and Zhang’s earnings

persistence regression show that results obtained using Q‐Score computed from

differentunconditionalconservatismmeasuresproduceveryinconsistentresults.This

suggests that different unconditional conservatism measurement models may be

measuring different conservatism phenomena. Therefore, the results produced by

conservatism impact assessment studies are shown tobedependent on the choiceof

Page 30: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

29

conservatism measure. The results obtained from industry‐based regressions are

significantlydifferentfromindustrytoindustry,too.Thisimpliesthatfirmsindifferent

industriesarenothomogenousintheirearningspersistenceresponsetoconservatism.

The relationship established between accounting conservatism and predictability of

coreRNOAinPenmanandZhang’sstudymaynotapplytoallindustries.

Since this studyhas shown that the choiceof conservatismmeasure canaffect

the results of the impact assessment study of PENMAN AND ZHANG significantly,

rampantdisagreementsincurrentresearchliteraturearelikelyaresultofinconsistent

application of conservatism measurement models. Since 94% of current research

adopts one or more of the five conservatism measurement models assessed in this

study, and amajority of studies (60%) rely on a singlemeasurementmodel, there is

sufficient evidence to raise concerns on the reliability and generalizability of results

obtained from these studies. These results may not be replicated using different

conservatismmeasures. These results also have limited applicability on the industry

levelduetoheterogeneousearningsresponsetoconservatismamongindustries.

AreasofInterestforFutureResearch

It has been established that studies on accounting conservatism produce

unreliable results if they rely onlyononeor two current conservatismmeasurement

models, especially when applied to cross‐industry samples. Hence, future research

should be conducted to explore, investigate, and propose ways to improve current

conservatism measurement methodology to ensure that reliable results can be

obtained.Thereareseveralinterestingareasthatscholarsmaylookinto.First,thereis

a need to refine the concept and definition of conservatism and to propose a

measurementmodelthatisconsistentwiththeconceptanddefinitionofconservatism

Page 31: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

30

so articulated. This is potentially aproject of immense scale, but if successful, itwill

result in an authoritative measurement for conservatism. This authoritative

conservatism measure will also have greater potential in practical application to

improveregulationpoliciesandmanagementdecision‐making.Secondly,sinceindustry

effects are documented in the conservatismmeasures in this study, industry specific

conservatismmeasuresshouldbederivedandresearchmayneedtobeconductedon

industry specific impact assessment of accounting conservatism. This is likely to

producemorereliableandmorespecificresults.

Page 32: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

31

ReferenceArtiach,T.andP.Clarkson.2011.Conservatism,disclosureandthecostofequitycapital.

AccountingandFinance51(1):2‐49.Balachandran, S. and P. Mohanram. 2011. Is the decline in the value relevance of

accounting driven by increased conservatism? Review of Accounting Studies16:272‐301.

Ball,R.andL.Shivakumar.2005.EarningsqualityinUKprivatefirms:comparativeloss

recognitiontimeliness.JournalofAccountingandEconomics39(1):83‐128.Bandyopadhyay,S.,C.Chen,A.HuangandR.Jha.2010.Accountingconservatismandthe

temporal trends in current earnings’ ability to predict future cash flows versusfuture earnings: evidence on the trade‐off between relevance and reliability.ContemporaryAccountingResearch27(2):413‐460.

Basu, S. 1997.The conservatismprinciple and theasymmetric timelinessof earnings.

JournalofAccountingandEconomics24(1):3‐37.Beatty,A.,Weber,J.,andJ.J.Yu.2008.ConservatismandDebt.JournalofAccountingand

Economics45(2‐3):154‐174.Beaver,W.H.andS.G.Ryan.2000.Biasesandlagsinbookvalueandtheireffectsonthe

ability of the Book‐to‐Market ratio to predict book return on equity. JournalofAccountingResearch38(1):127‐148.

Beaver, W.H. and S.G. Ryan. 2005. Conditional and unconditional conservatism:

Conceptsandmodeling.ReviewofAccountingStudies10(2‐3):269‐309.Chandra, U. 2011. Income conservatism in the U.S. technology sector. Accounting

Horizons25(2):285‐314.Chen, Q., T. Hemmer and Y. Zhang. 2007. On the relation between conservatism in

accounting standards and incentives for earnings management. Journal ofAccountingResearch45(3)(June):541‐565.

Davidson, S., Stickney, C.P., and R.L. Weil. 1985. Intermediate Accounting: Concepts

MethodsandUses,4thEdition.USA:DrydenPress.Dichev, I. and V. Tang. 2008. Matching and the changing properties of accounting

earningsoverthelast40years.TheAccountingReview83(6):1425‐1460.Dietrich,J.R.,K.A.MullerandE.J.Riedl.2007.Asymmetrictimelinesstestsofaccounting

conservatism.ReviewofAccountingStudies9:495‐521.Fan, Q. and Zhang X.J. 2012. Accounting conservatism, aggregation, and information

quality.ContemporaryAccountingResearch:29(1):38‐56.

Page 33: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

32

FinancialAccountingStandardsBoard.Conceptualframeworkforfinancialreporting–Chapter 1, The objective of general purpose financial reporting, and Chapter 3,Qualitative characteristics of useful financial information. Statementof FinancialAccountingConceptsNo.8,September2010.

Feltham, G.A., and J.A. Ohlson. 1995. Valuation and clean surplus accounting for

operatingand financial activities.ContemporaryAccountingResearch11(2):689‐731.

Gigler, F.B. and T. Hemmer. 2001. Conservatism, optimal disclosure policy, and the

timelinessoffinancialreports.TheAccountingReview76(4):471‐493.Gotti, G. 2008. Conditional conservatism in accounting: new measure and tests of

determinants. Conference Paper: American Accounting Association (AAA) 2008AnnualMeetingProceedingsSeries.

Givoly, D. and C. Hayn. 2000. The changing time‐series properties of earnings, cash

flowsandaccruals:Hasfinancialreportingbecomemoreconservative?JournalofAccounting&Economics29(3):287‐320.

Givoly, D., C. Hayn and A. Natarajan. 2007. Measuring reporting conservatism.

AccountingReview82(1):65‐106.Hui, K.W., Matsunaga, S. and D. Morse. 2009. The impact of conservatism on

management earnings forecasts.TheJournalofAccountingandEconomics 47(3):192‐207.

Jackson,S.andX.Liu.2010.Theallowanceforuncollectibleaccounts,conservatism,and

earningsmanagement.JournalofAccountingResearch48(3):565‐601.Kim,M. andW. Kross. 2005. The ability of earnings to predict future operating cash

flowshasbeenincreasing–notdecreasing.JournalofAccountingResearch43(5):753‐780.

Kim, B. H. and M. Pevzner. 2010. Conditional accounting conservatism and future

negative surprises: an empirical investigation. Journal ofAccountingandPublicPolicy29(4):311‐329.

LaFond,R.andR.L.Watts.2008.Theinformationroleofconservatism.TheAccounting

Review83(2):447‐478.Lara, J.M.G., B.G. Osma and F. Penalva. 2011. Information effects of conservatism in

accounting. Working Paper: Universidad Carlos III de Madrid, UniversidadAutonomadeMadrid,andUniversityofNavarra‐IESEBusinessSchool.

Li, J. 2010. Accounting conservatism, information uncertainty and analysts’ forecasts.

WorkingPaper:CarnegieMellonUniversity.

Page 34: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

33

Manuel,C., andN.N.Manuel.2011.Consistentestimationof conditional conservatism.WorkingPaper:UniversityofJaenandUniversidadCarlosIIIdeMadrid.

Nishitani,J.2010.TypeIandtypeIIerrorsofconservatism.ConferencePaper:Canadian

AcademicAccountingAssociation(CAAA)AnnualConferenceProceedingsSeries.Patatoukas, P.N., and J.K. Thomas. 2011. More evidence of bias in the differential

timeliness measure of conditional conservatism. The Accounting Review 86(5):1765‐1793.

Penman,S. andX.Zhang.2002.Accounting conservatism, thequalityof earnings, and

stockreturns.TheAccountingReview77(2):237‐264.Roychowdhury,S.andR.L.Watts.2007.Asymmetrictimelinessofearnings,market‐to‐

bookandconservatisminfinancialreporting.JournalofAccountingandEconomics44(2):2‐31.

Ruch, G.W., and G. Taylor. 2011. Accounting conservatism and its effects on financial

reportingquality:areviewoftheliterature.WorkingPaper,UniversityofAlabama.Ryan,S.G.1995.Amodelofaccrualmeasurementwithimplicationsfortheevolutionof

thebook‐to‐marketratio.JournalofAccountingResearch33(1):95‐112.Ryan, S.G. 2006. Identifying conditional conservatism. European Accounting Review.

15(4):511‐525.Sohn, B.C. 2012. Analyst forecast, accounting conservatism and the related valuation

implications.AccountingandFinance:52(1):311‐341Sterling, R. 1970. Theory of themeasurement of enterprise income. USA: Accounting

PublicationofScholarsBookCo.Wang,R.Z.,C.O.HogartaighandT.V.Zijl.2009.Measuresofaccountingconservatism:a

constructvalidityperspective.JournalofAccountingLiterature28:165‐203.Watts,R.L.1993.Aproposalforresearchonconservatism.ConferencePaper:American

AccountingAssociation(AAA)Convention,SanFrancisco,August1993.Watts, R.L. 2003a. Conservatism in accounting part I: Explanations and implications.

AccountingHorizons17(3):207‐222.Watts, R.L. 2003b. Conservatism in accounting part II: Evidence and research

opportunities.AccountingHorizons17(4):287‐301.Wolk, H. I., Francis, J.R. and M.G. Tearny. 1989. AccountingTheory:AConceptualand

InstitutionalApproach.USA:PWS‐KENTPublishingCompany.Zhang,X.J.2000.Conservativeaccountingandequityvaluation.JournalofAccounting&Economics29(1):125‐149.

Page 35: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

34

AppendixA:SummaryofFormulasandComputationVariables

Thisappendixpresentsasummaryoftheformulasusedforthecomputationoftheconservatismmeasuresandregressionsusedinthisstudy. ConservatismMeasures Formula Datavariablesneeded COMPUSTAT/

CRSPCode1 AT‐Score

RitiscalculatedbycompoundedCRSPmonthlyreturndata isadummyvariablewhereitis1ifthereturnisnegative.AT‐Score=

Basicearningspershare EPSPIFiscalyearclosingprice PRCC_FMonthlyReturn RET

2 AACF‐Score ∗CFO=Earningsbeforeextraordinaryitems‐accrualsAccruals=Δinventory+Δdebtors+Δothercurrentassets–Δcreditors–Δothercurrentliabilities–depreciation

isadummyvariablewhereitis1iftheCFOisnegative.AACF‐Score=

Earningsbeforeextraordinaryitems

IB

Inventory INVTDebtors RECTOthercurrentassets ACOCreditors APOthercurrentliabilities DLCDepreciation DP

3 HR‐Score

=reportedLIFOreserveforfirmiinfiscalyeart=CapitalizedR&Dexpense–amortization=Capitalizedadvertisingexpense–amortization=Totalassets–totalliabilities–preferenceequity(book)

HR‐Score=

LIFOreserve LIFRR&Dexpense XRDAdvertisingexpense XADTotalassets ATTotalliabilities LTPreferenceEquity PSTK

4 NA‐Score

∆ ∆ ∆ . ∆ ∆

^NA‐Score=(‐1)(NOACC)

NetIncome NIDepreciation DPCashflowfromOperations OANCFAccountReceivable RECTInventory INVTPrepaidExpense XPPAccountPayable APTaxPayable TXP

Page 36: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

35

5 MTB‐Score ∗

^MTB‐Score=

Numberofoutstandingsharesatfiscalyearend

CSHO

Fiscalyearclosingprice PRCC_FTotalasset ATTotalLiabilities LTPreferenceequity PSTK

RobustnessTestofConservatismMeasuresusingPenmanandZhang’s(2002)Methodology

Item Formula DataVariables Compustat/CRSPCode

1 Core‐RNOA

∗ 1

^CoreOperatingIncome=Operatingincomeafterdepreciation+InterestExpense^ =Totalassets–totalliabilities–preferenceequity(book)

Operatingincomeafterdepreciation

OIADP

Interestexpense XINTTotalassets ATTotalliabilities LTPreferenceequity(book)

PSTK

2 Q‐Score 0.5 0.5^ =ConservatismScoreforfirmiinfiscalyeart

3 Regression

Page 37: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

36

Table1:FiveCommonlyUsedMeasurementMethods

This table presents the frequency of the five conservatism measurement methodssurveyedbyWangetal(2009).

S/NMeasurement

Method ContributorFrequencyofUse(Wangetal.,2009)

TypeofConservatismMeasures

1Asymmetric

Timeliness(AT)Model

Basu(1997)36of52papers(69%)

ConditionalMeasure

2AsymmetricAccrualtoCash‐Flow(AACF)

Model

BallandShivakumar(2005)

7of52papers(13%)

ConditionalMeasure

3 HiddenReserve(HR)Model

PenmanandZhang(2002)

9of52papers(17%)

UnconditionalMeasure

4NegativeAccrualsMeasure(NAM)

ModelGivolyandHayn(2000)

10of52papers(19%)

UnconditionalMeasure

5 Market‐to‐Book(MTB)Model

FelthamandOhlson(1995)

Ryan(1995)BeaverandRyan

(2000)

13of52papers(25%)

UnconditionalMeasure

Page 38: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

37

Table2:SampleFilteringSteps

This table summarizes the sampling criteria that result in the final sample of 43,434firmyearsover23‐yearperiodof1988‐2010. Figures inparenthesisarepercentagesrelativetotheten‐industrysample.Thefinalsampleconsistsoffirm‐yearsfrom1988to2010 because of the data demands in computing the five conservatism scores eventhoughtheoriginalperiodofstudywasfrom1975to2010.PanelA:DerivationoftheFinalSample

DescriptionSampleSize(%)

Firm‐yearsintheten‐industrysample65,786(100.0%)

Less:firm‐yearswithnegativeorzerobookvalue(BV) 3,142(4.8%)

Less:firm‐yearswithnegativeorzeronetoperatingassets(NOA) 2,458(3.7%)

Firm‐yearsbeforescreeningforunavailableconservatismscore 60,186(91.5%)

Less:firm‐yearswithoutNA‐Score10,114(15.4%)

Less:firm‐yearswithoutreturnsdata 6,638(10.1%)

Firm‐yearswithcompleteinformationforallfiveconservatismmeasures(FinalSample)

43,434(66.0%)

Firm‐yearsfrom1988to199719,931(45.9%)

Firm‐yearsfrom1998to2010 23,503(54.1%)

FinalSamplecovering1988to2010 43,434(66.0%)

PanelB:DistributionoftheFinalSampleAmongsttheTenIndustries

NameofIndustry 2‐digitSIC

Firm‐years

OilandGas 13 2,139FoodandKindredProducts 20 1,916ChemicalsandAlliedProducts 28 5,752IndustrialandCommercialMachineryandComputerEquipment 35 4,672ElectronicsandotherElectricalEquipment 36 5,910TransportationEquipment 37 1,531InstrumentsandRelatedProducts 38 4,840Communications 48 2,081RetailandWholesaleTrade 50‐59 8,392BusinessServices 73 6,021FinalSample 43,434

Page 39: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

38

Table3:DescriptiveStatisticsfortheFinalSampleThistablepresentsthedescriptivestatisticsofthekeyvariablesforthetwoperiodsofanalysisadoptedinthisstudy,1988‐1997and1988‐2010.MKTdenotesmarketcapitalizationofthefirm,ATdenotestotalasset,INVTdenotesinventory,LIFRdenotesLIFOreserve, EPSPI denotes basic earnings per share, ACC denotes accruals, DP denotes depreciation, XAD denotes advertisingexpense,andXRDdenotesR&Dexpense.

PanelA‐1988‐1997DescriptiveStatisticsMKT AT INVT LIFR EPSPI ACC DP XAD XRD

NumbersofFirm‐Year 19931 19931 19931 19439 19931 19931 19931 6455 13961 Mean 1224.190 1042.223 137.725 6.600 0.416 ‐48.242 54.653 57.287 50.524Median 81.152 76.007 8.584 0.000 0.300 ‐1.211 3.077 2.459 2.397Std.Deviation 5649.324 4712.890 633.085 51.276 1.753 385.251 327.618 216.341 262.473Skewness 11.765 11.103 12.496 25.118 ‐0.646 ‐23.577 22.308 7.147 10.163Minimum 0.062 0.218 0.000 ‐88.000 ‐39.570 ‐18472.000 ‐4.385 0.000 0.000Maximum 164758.840 132864.000 17665.831 2123.000 28.020 3979.000 17287.000 3468.000 5522.258Percentiles 25 19.915 20.401 1.109 0.000 ‐0.180 ‐10.323 0.727 0.410 0.231

75 381.279 326.612 50.672 0.000 1.050 0.825 13.814 14.667 12.651

PanelB‐1988‐2010DescriptiveStatistics MarketValue AT INVT LIFR EPSPI ACC DP XAD XRD

NumbersofFirm‐Year 43434 43434 43434 42394 43434 43434 43434 15370 31162 Mean 2935.599 2342.280 215.505 7.051 0.432 ‐116.250 119.877 92.558 103.162Median 148.387 141.617 12.390 0.000 0.330 ‐2.911 5.653 3.488 4.452Std.Deviation 14141.116 11484.951 918.001 59.733 2.024 862.805 771.365 352.001 518.982Skewness 11.690 12.573 13.259 24.327 ‐0.689 ‐19.078 18.378 7.674 9.499Minimum 0.062 0.218 0.000 ‐196.100 ‐52.840 ‐47110.156 ‐4.385 .000 ‐.202Maximum 467092.880 324939.000 35180.000 3003.000 45.510 9417.000 33750.967 7937.000 12183.000Percentiles 25 32.427 33.135 1.261 0.000 ‐.240 ‐24.465 1.215 .543 .407

75 846.432 754.743 85.318 0.000 1.170 0.436 30.465 25.300 23.979

Page 40: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

39

*AlllevelsofaccountsareindollarvaluesareinmillionsofUSD.

Table4:DescriptiveStatisticsoftheFinalSampleandtheTen‐IndustrySub‐Samplefrom1988‐2010This table presents the descriptive statistics for nine main variables used in the computation of conservatism scores, earningspersistence scores and core‐RNOA for the extendedperiod1987‐2010.MKTdenotesmarket capitalismof the firm,ATdenotes totalasset, INVTdenotes inventory,LIFRdenotesLIFOreserve,EPSPIdenotesbasicearningspershare,ACCdenotesaccruals,DPdenotesdepreciation,XADdenotesadvertisingexpense,andXRDdenotesR&Dexpense.

FullSample Oil&Gas Food Chemicals Machinery Electronics Transport Instru. Comm. Trade Business2‐DigitsSICCode 13 20 28 35 36 37 38 48 50‐59 73SampleSize 43434 2139 1916 5752 4672 5910 1531 4840 2081 8392 6201

MKTMean 2,936 1,764 4,230 4,832 2,554 2,494 4,088 1,039 10,740 1,676 2,171%offullsamplemean 100% 60% 144% 165% 87% 85% 139% 35% 366% 57% 74%Median 148 137 280 216 148 114 219 85 852 139 110StandardDeviation 14,141 6,475 13,670 18,860 13,653 12,080 15,151 3,733 27,763 8,772 16,064 ATMean 2,342 1,621 2,817 2,465 1,998 1,726 6,033 783 12,714 1,504 1,138%offullsamplemean 100% 69% 120% 105% 85% 74% 258% 33% 543% 64% 49%Median 141.62 173.55 369.65 114.09 149.71 108.19 323.16 59.79 1125.48 220.02 81.87StandardDeviation 11,485 4,954 5,862 9,343 8,296 7,274 26,292 3,047 34,724 6,078 6,698

Page 41: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

40

Table4:DescriptiveStatisticsoftheFinalSampleandtheTen‐IndustrySub‐Samplefrom1988‐2010(cont.)

*AlllevelsofaccountsareinmillionsofUSD.

FullSample Oil&Gas Food Chemicals Machinery Electronics Transport Instru. Comm. Trade Business2‐DigitsSICCode 13 20 28 35 36 37 38 48 50‐59 73SampleSize 43434 2139 1916 5752 4672 5910 1531 4840 2081 8392 6201

INV Mean 216 47 332 235 246 199 636 94 154 351 41%offullsamplemean 100% 22% 154% 109% 114% 92% 295% 43% 71% 163% 19%Median 12.39 0.00 54.50 9.09 25.66 17.92 60.57 10.69 2.59 49.04 0.00StandardDeviation 918 210 733 735 1,005 802 2,112 354 467 1,285 461

LIFRMean 7.05 1.13 9.25 11.20 13.95 2.74 12.33 3.52 0.07 12.38 0.02%offullsamplemean 100% 16% 131% 159% 198% 39% 175% 50% 1% 176% 0%Median 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00StandardDeviation 59.73 11.54 33.64 61.83 134.71 19.46 55.09 40.43 1.02 58.64 0.61

EPS1Mean 0.43 0.43 1.06 0.33 0.48 0.30 1.04 0.45 0.39 0.59 0.05%offullsamplemean 100% 100% 245% 77% 110% 70% 241% 103% 91% 136% 13%Median 0.33 0.21 0.80 0.09 0.38 0.26 0.87 0.29 0.32 0.59 0.11StandardDeviation 2.02 2.20 1.95 1.89 2.23 1.65 2.43 1.46 3.58 1.93 1.84 ACCMean ‐116 ‐108 ‐97 ‐87 ‐74 ‐94 ‐230 ‐25 ‐970 ‐47 ‐55%offullsamplemean 100% 93% 84% 75% 64% 81% 197% 21% 834% 41% 47%Median ‐2.91 ‐10.96 ‐7.42 ‐1.77 ‐2.42 ‐2.40 ‐5.05 ‐0.66 ‐38.02 ‐3.48 ‐2.69StandardDeviation 863 375 296 477 432 569 1,697 148 3,079 319 406

Page 42: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

41

Table4:DescriptiveStatisticsoftheFinalSampleandtheTen‐IndustrySub‐Samplefrom1988‐2010(cont.)

*AlllevelsofaccountsareinmillionsofUSD.

FullSample Oil&Gas Food Chemicals Machinery Electronics T.Equip. Instru. Comm. Trade B.Services2‐DigitsSICCode 13 20 28 35 36 37 38 48 50‐59 73SampleSize 43434 2139 1916 5752 4672 5910 1531 4840 2081 8392 6201

DPMean 120 115 105 97 81 92 254 30 944 56 51%offullsamplemean 100% 96% 87% 81% 68% 76% 212% 25% 788% 46% 43%Median 5.65 13.62 17.04 3.56 5.64 4.48 11.81 2.10 42.27 8.14 3.60StandardDeviation 771 362 211 369 389 396 1,284 128 2,932 217 318

XADMean 93 2 228 181 44 69 274 22 245 67 31%offullsamplemean 100% 2% 246% 195% 48% 74% 296% 23% 265% 72% 34%Median 3.49 0.12 18.34 4.54 2.56 0.80 3.31 0.86 12.65 10.20 1.24StandardDeviation 352 9 425 593 151 339 822 104 654 178 158 XRDMean 103 68 34 196 113 141 282 40 217 1 92%offullsamplemean 100% 66% 33% 190% 109% 136% 273% 39% 210% 1% 90%Median 4.45 2.64 8.35 13.32 9.00 7.86 10.86 4.36 9.17 0.00 7.53StandardDeviation 519 162 51 764 455 611 969 158 626 30 521

Page 43: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

42

Table5:IndustryBasedRegressionResultsforConditionalConservatismModels

ThistablepresentstheresultsoftheATandAACFmultipleregressionsperformedforeach individual industry. β1is the slope coefficient of the interaction term betweendummyvariableandproxyfornews(returninthecaseofATandCFOforthecaseofAACF). It represents the incremental timeliness of recognition of bad news ascomparedtogoodnews,andhencetheconservatismscore.Fordetailsofcomputationandnotationsofregressionvariables,pleaserefertoTable1.

ATRegressionEquation: ∗

PanelA:ATRegressionResults(1988‐1997) PanelB:ATRegressionResults(1988‐2010)

Industry β1Std.Error t Sig.

Industry β1Std.Error t Sig.

13 0.250 0.237 1.053 0.293 13 0.384 0.139 2.759 0.006

20 0.700 0.213 3.286 0.001 20 0.351 0.100 3.497 0.000

28 0.455 0.082 5.554 0.000 28 0.354 0.043 8.267 0.000

35 0.462 0.090 5.138 0.000 35 0.361 0.051 7.104 0.000

36 0.505 0.102 4.941 0.000 36 0.319 0.057 5.616 0.000

37 0.397 0.080 4.972 0.000 37 0.594 0.089 6.684 0.000

38 0.295 0.039 7.633 0.000 38 0.281 0.040 7.091 0.000

48 0.227 0.044 5.159 0.000 48 0.109 0.142 0.762 0.446

50‐59 0.288 0.034 8.389 0.000 50‐59 0.270 0.029 9.342 0.000

73 0.426 0.058 7.304 0.000 73 0.441 0.045 9.836 0.000

AACFRegressionEquation: ∗ PanelC:AACFRegressionResults(1988‐1997) PanelD:AACFRegressionResults(1988‐2010)

Industry β1Std.Error

t Sig. Industry β1Std.Error

t Sig.

13 0.840 0.061 13.866 0.000 13 0.682 0.079 8.639 0.000

20 0.138 0.106 1.306 0.192 20 ‐0.985 0.061 ‐16.039 0.000

28 ‐0.164 0.102 ‐1.608 0.108 28 ‐0.105 0.046 ‐2.264 0.024

35 0.994 0.090 11.015 0.000 35 0.528 0.051 10.281 0.000

36 1.208 0.048 25.129 0.000 36 1.055 0.019 56.911 0.000

37 ‐0.910 0.123 ‐7.389 0.000 37 ‐0.304 0.34 ‐0.896 0.371

38 0.109 0.077 1.418 0.156 38 0.513 0.022 23.659 0.000

48 0.062 0.068 0.915 0.360 48 0.941 0.024 39.808 0.000

50‐59 ‐0.612 0.020 ‐31.046 0.000 50‐59 ‐0.390 0.013 ‐30.118 0.000

73 1.968 0.415 4.743 0.000 73 0.853 0.029 29.262 0.000

Page 44: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

43

Table6:SummaryofConservatismScoresforIndustries

This tablesummarizesall five rawconservatismscores foreach industry. Conditional conservatismscoresareslopecoefficientβ1of regressionperformedforeachindustry.Unconditionalconservatismscoresarethemeansofallfirm‐yearscores.Industriesarethenrankedbasedontherawscores.

PanelA:RawConservatismScoreandIndustryRanking,1988‐1997(1–leastconservative;10–mostconservative)

Industry AT_Score Ranking AACF_Score Ranking HR_Score Ranking NA_Score Ranking MTB_Score RankingAverageRanking

StdDevRanking

13 0.250 1 0.840 7 0.01 1 5.35 1 2.51 3 2.6 2.620 0.700 10 0.138 6 3.50 8 30.86 9 3.11 5 7.6 2.128 0.455 7 ‐0.164 3 4.57 10 10.73 3 4.85 9 6.4 3.335 0.462 8 0.994 8 3.58 9 10.81 4 3.40 6 7.0 2.036 0.505 9 1.208 9 1.16 5 12.69 6 2.85 4 6.6 2.337 0.397 5 ‐0.910 1 1.09 4 19.56 8 2.25 1 3.8 2.938 0.295 4 0.109 5 0.89 3 6.84 2 3.96 7 4.2 1.948 0.288 2 0.062 4 1.83 6 84.02 10 5.15 10 6.4 3.6

50‐59 0.288 3 ‐0.612 2 0.42 2 11.40 5 2.43 2 2.8 1.373 0.426 6 1.968 10 2.67 7 15.77 7 4.48 8 7.6 1.5

PanelB:RawConservatismScoreandIndustryRanking,1988‐2010(1–leastconservative;10–mostconservative)

Industry AT‐Score Ranking AACF‐Score Ranking HR‐Score Ranking NA‐Score Ranking MTB_Score RankingAverageRanking

StdDevRanking

13 0.384 8 0.682 7 0.01 1 28.27 3 2.45 3 4.4 3.020 0.351 5 ‐0.985 1 2.03 7 27.56 2 3.34 6 4.2 2.628 0.354 6 ‐0.105 4 4.65 10 33.44 6 4.88 10 7.2 2.735 0.361 7 0.528 6 2.43 8 42.45 8 3.09 5 6.8 1.336 0.319 4 1.055 10 1.30 6 31.53 5 2.93 4 5.8 2.537 0.594 10 ‐0.304 3 1.00 3 77.91 9 2.39 1 5.2 4.038 0.281 3 0.513 5 1.06 4 13.27 1 3.65 7 4.0 2.248 0.109 1 0.941 9 1.14 5 329.15 10 4.22 9 6.8 3.8

50‐59 0.270 2 ‐0.390 2 0.37 2 31.19 4 2.44 2 2.4 0.973 0.441 9 0.853 8 2.88 9 34.84 7 4.18 8 8.2 0.8

Page 45: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

44

Table7:ClassificationOverlapfortheFiveConservatismModelsatIndustryLevelThis table summarizes the classification overlap among all five unconditionalconservatismmodels.Topthreeindustrieswithhighestscoreundereachconservatismmeasureareclassifiedashighconservatism,whilethe lowestscoringthree industriesareclassifiedaslowconservatism.Thenumbersinthefirstrowofeachpanelindicatethe number of conservatism models that classify a particular industry under thecategorystatedinthepaneldescription. Listingunder0indicatesthattheindustryisnotclassifiedunderthatcategorybyanyofthefiveconservatismmeasures.

PanelA:HighConservatism1988‐1997

NumberofConservatismModelsAgreeinClassification

5 4 3 2 1 0Industry 20 28 37 13

35 36 3848 5073

PanelB:LowConservatism1988‐1997

NumberofConservatismModelsAgreeinClassification 5 4 3 2 1 0Industry 13 37 48 20

50 28 3538 36

73

PanelC:HighConservatism1988‐2010

NumberofConservatismModelsAgreeinClassification 5 4 3 2 1 0Industry 73 48 28 13 10

37 36 3835 50

PanelD:LowConservatism1988‐2010

NumberofConservatismModelsAgreeinClassification 5 4 3 2 1 0Industry 50 13 20 48 28

37 38 353673

Page 46: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

45

Table8:ClassificationOverlapamongUnconditionalConservatismMeasureswithinEachIndustryThis table summarizes the classification overlap at firm‐year level for individual industries for both 1988‐1997 and 1988‐2010. Top 30% firm‐years of eachindustryrankedundereachconservatismmeasureareclassifiedashighconservatism,whilethelowest30%firm‐yearsofeachindustryrankedareclassifiedaslowconservatism.Thefirm‐yearoverlapisthenumberoffirm‐yearsthatisclassifiedunderthesamecategorybyallthreeunconditionalconservatismmeasures.%Overlapiscomputedasfirm‐yearoverlap/30%oftotalfirm‐yearsclassifiedundertheindustry.Itisimportanttonotethatdifferentindustryportfolioshavedifferentsamplesizes,andhence%overlapratherthanfirm‐yearoverlapshouldbeusedforcomparisonbetweenindustries.

Industry

Total 1988‐1997 Total 1988‐2010Firm‐years

HighConservatism LowConservatismFirm‐years

HighConservatism LowConservatism

  Firm‐yearOverlap

%Overlap

Firm‐yearOverlap

%Overlap

  Firm‐yearOverlap

%Overlap

Firm‐yearOverlap

%Overlap

13 1045 91 29% 77 25% 2139 153 24% 178 28%

20 901 13 5% 14 5% 1916 34 6% 71 12%

28 2530 65 9% 110 14% 5752 178 10% 251 15%

35 2251 56 8% 74 11% 4672 131 9% 131 9%

36 2682 80 10% 122 15% 5910 180 10% 226 13%

37 707 15 7% 27 13% 1531 45 10% 71 15%

38 2278 80 12% 88 13% 4840 142 10% 169 12%

48 910 11 4% 42 15% 2081 23 4% 52 8%

50 4091 69 6% 109 9% 8392 163 6% 231 9%

73 2536 62 8% 91 12% 6201 152 8% 180 10%

Page 47: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

46

Table9:HRQ‐ScoreComparison

This tablesummarizes thestatisticsof filteredQ‐Score(Q_HRScore),andtheQ‐Scorecomputed andpublished inPenmanandZhang’spaper (2002). This study replicatesthatofPenmanandZhang(2002).However,dataperiodforthispartofanalysisdiffersfromthatofPenmanandZhang(2002).

Filtered

Q_HRScore

Penman&Zhang

Q_Score

(Table1)

NumberofFirm‐years 33,943 29,796

DataPeriod 1988‐2010 1975‐1997

Mean 0.350 0.099

Percentiles 95 1.858 0.219

75 0.139 0.059

50 ‐0.011 0.009

40 ‐0.032 0.000

25 ‐0.101 ‐0.010

10 ‐0.205 ‐0.046

5 ‐0.282 ‐0.075

Page 48: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

47

Table10:SummaryofResultsofAnnualCross‐SectionalRegressions

This table summarizes the regression results replicating Penman and Zhang (2002)annual cross‐sectional regressions for the period 1976‐1996. The t‐statistic iscalculated based on a one‐sample t‐test of annual coefficients obtained against thehypothesisthatthemeanofcoefficientsequalstozero.Forthisstudy,22annualcross‐sectional regressions were performed for each of the Q‐Scores for the period 1988‐2009. FollowingPenmanandZhang,threesimilarone‐samplet‐testswereperformedforeachoftheQ‐scorestestingthemeanof22coefficientsforeachcase.**Significantat1%,and*significantat5%.

EarningsPersistenceEquation:RNOAt+1=α0+α1RNOAt+α2Qt+et+1

PenmanandZhangQ_HR(Table3,PanelA) Mean

FirstQuartile Median

ThirdQuartile

Intercept 0.016** 0.009 0.018 0.022RNOAcoefficient 0.800** 0.782 0.813 0.846Qcoefficient 0.096** 0.054 0.103 0.136

Q_HR MeanFirst

Quartile MedianThirdQuartile

Intercept 0.029* 0.008 0.044 0.060RNOAcoefficient 0.284** 0.188 0.256 0.394Qcoefficient 0.027** 0.002 0.030 0.050

Q_MTB MeanFirst

Quartile MedianThirdQuartile

Intercept 0.036** 0.022 0.037 0.073RNOAcoefficient 0.263** 0.139 0.255 0.343Qcoefficient 0.018 0.007 0.018 0.035

Q_NA MeanFirst

Quartile MedianThirdQuartile

Intercept 0.028 0.019 0.040 0.081RNOAcoefficient 0.263 0.141 0.255 0.369Qcoefficient <0.000 <0.000 <0.000 <0.000

Page 49: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

48

Table11:SummaryofResultsofPooledRegressions

Thistablesummarizestheresultsoftheregressionforthefilteredsampleasawhole.IndividualregressionswererunforeachoftheQ‐ScoresusingthePenmanandZhang’searningspersistenceregressionequation:RNOAt+1=α0+α1RNOAt+α2Qt+et+1.**Significantat1%,and*significantat5%.

PanelA:OverallRegressionResultforQ_HR

Variable Coefficient t‐Statistic

C 0.0208** 2.6373

CRNOA_T 0.2491** 33.7165

Q_HR 0.0185** 6.4362

F‐statistic 574.7051**

PanelB:OverallRegressionResultforQ_MTB

Variable Coefficient t‐Statistic

C 0.0240** 3.0414

CRNOA_T 0.2368** 33.4219

Q_MTB 0.0118** 3.3847

F‐statistic 559.2320**

PanelC:OverallRegressionResultforQ_NA

Variable Coefficient t‐Statistic

C 0.0256** 3.2547

CRNOA_T 0.2350** 33.2323

Q_NA 0.0002* 2.2110

F‐statistic 555.8412**

Page 50: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

49

Table12:SummaryofResultsofIndustryStackedRegression

TheindustrycoefficientswereobtainedbyrunningastackedregressionusingthePenmanandZhang’searningspersistenceregressionequation: RNOAt+1 = α0 + α1RNOAt + α2Qt + et+1.Pooling was performed by assigning dummy variables to nine out of ten selectedindustries.Business(73)wasthebaseindustryanditsregressioncoefficientsweresubsumedintheconstantterm.AseparatestackedregressionwasperformedforeachofthethreeQ‐Scores.**Significantat1%,and*significantat5%.

PanelA:IndustryRegressionResultfor

Q_HR

PanelB:IndustryRegressionResultforQ_MTB

PanelC:IndustryRegressionResultforQ_NA

Variable Coefficient t‐Statistic Variable Coefficient t‐Statistic

Variable Coefficient t‐Statistic

Constant 0.0219** 2.7824 Constant 0.0235** 2.9748 Constant 0.0257 3.2654

CRNOA_T 0.2501** 33.4517 CRNOA_T 0.2335** 32.862 CRNOA_T 0.2350** 33.2289

Q_HR*D_13 ‐2.5372** ‐3.0362 Q_MTB*D_13 ‐0.0039 ‐0.0688 Q_NA*D_13 0.0002 0.2867

Q_HR*D_20 0.0398* 2.2216 Q_MTB*D_20 0.1027** 5.6679 Q_NA*D_20 0.0002 0.7643

Q_HR*D_28 0.0131** 2.7056 Q_MTB*D_28 ‐0.0132 ‐1.8114 Q_NA*D_28 0.0002 0.8958

Q_HR*D_35 0.0288** 2.9398 Q_MTB*D_35 ‐0.0174 ‐1.4938 Q_NA*D_35 0.0002 0.8218

Q_HR*D_36 0.0677** 8.5352 Q_MTB*D_36 0.0182 1.8271 Q_NA*D_36 0.0001 0.6199

Q_HR*D_37 0.0033 0.1307 Q_MTB*D_37 0.0070 0.2621 Q_NA*D_37 0.0002 0.5981

Q_HR*D_38 0.0237 1.7581 Q_MTB*D_38 0.0140 1.4343 Q_NA*D_38 0.0001 0.4153

Q_HR*D_48 0.0002 0.0113 Q_MTB*D_48 0.0027 0.2065 Q_NA*D_48 0.0002 0.8353

Q_HR*D_50 ‐0.0393* ‐2.4151 Q_MTB*D_50 0.0364** 3.2794 Q_NA*D_50 0.0003 1.2764

F‐statistic 121.8645** F‐statistic 116.1929** F‐statistic 111.1905**

Page 51: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

50

Figure1:DigrammaticRepresentationoftheFactorswhichContributetotheConservatismofaCompany

Thisfiguresummarizesthemajorcontributionstotheconservatismofafirm.

↑Expenses

Delay recognition of revenue: Revenue recognition methods which have time

dimension: cost recovery, percentage of completion method

More stringent recognition and measurement criteria

↓Revenue

=

↓Earnings

Accelerated recognition of expenses: Write-offs Inventory (lower-of-cost or market value) Accounts Receivable Tangible assets

o Depreciation methods used o Impairment charges

Intangible assets o Amortization methods used o Impairment charges

Provisions (creation of liability) Provision for warranty Accrued expenses (leaves, staff benefits

etc) Contingent liabilities

Capitalization versus expense Future benefits generating expenses

(R&D, advertising, brand names, patent expenses, interest on construction

Net Effect is that annual earnings and cash flows become out of synchronization.

Page 52: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

51

Figure2:ConservatismScoreRankingforIndustries

Thisfigureshowstheconservatismrankcomputedforeachindustryforeachofthefiveconservatismmeasurementmodelsfortheperiod1988‐1997and1988‐2010.They‐axis indicates the rankwhereas the x‐axis indicates themeasurementmodel used tocompute the score. Each series shows the ranks of one industry under eachconservatismmodel.Thelegendindicatesthecodeoftheindustryrepresentedbyeachseries. Industries are ranked from 1‐10 with 1 being the least conservative and 10beingthemostconservative.

PanelA:ConservatismScoreRankingforIndustriesforthePeriod1988‐1997

PanelB:ConservatismScoreRankingforIndustriesforthePeriod1988‐2010

1

2

3

4

5

6

7

8

9

10

ATScore

AACFScore

HRScore

NAScore

MTBScore

13

20

28

35

36

37

38

48

50‐59

73

1

2

3

4

5

6

7

8

9

10

ATScore

AACFScore

HRScore

NAScore

MTBScore

13

20

28

35

36

37

38

48

50‐59

73

Page 53: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

52

Figure3:ClassificationbyConditionalConservatismModelsatIndustryLevel

ThisfigureconsistsoffourVenndiagramsoftheclassificationresultsarrangedintotwopanels. Panel A (B) presents the results obtained for 1988‐1997 (1988‐2010). Topthreeindustriesrankedundereachconditionalconservatismmeasureareclassifiedashighconservatism,whilethelowestthreeindustriesareclassifiedaslowconservatism.Overlap in classification is expressed as the overlap between the circles in the Venndiagram.Onlyconditionalconservatismmodelsareincludedinthisanalysis.

PanelA:IndustryClassificationfortheyears1988‐1997

PanelB:IndustryClassificationfortheyears1988‐2010

HighConservatism LowConservatism

HighConservatism LowConservatism

AACF AACF

AACF AACF 

Page 54: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

53

Figure4:ClassificationbyUnconditionalConservatismModelsatIndustryLevel

ThisfigureconsistsoffourVenndiagramsoftheclassificationresultsarrangedintotwopanels. Panel A (B) presents the results obtained for 1988‐1997 (1988‐2010). Topthreeindustriesrankedundereachunconditionalconservatismmeasureareclassifiedas high conservatism, while the lowest three industries are classified as lowconservatism.Overlapinclassificationisexpressedastheoverlapbetweenthecirclesin the Venn diagram. Only unconditional conservatism models are included in thisanalysis.

PanelA:IndustryClassificationfortheyears1988‐1997

PanelB:IndustryClassificationfortheyears1988‐2010

HighConservatism LowConservatism

HighConservatism LowConservatism

Page 55: Conservatism – A Measurement Maze Ho Yew Kee Yuan Yi ... · Conservatism – A Measurement Maze Ho Yew Kee ... Conservatism – A Measurement Maze ... The next sub‐section will

54

Figure5:Firm‐yearClassificationUsingUnconditionalConservatismModelsThisfigureconsistsoffourVenndiagrams.DiagramsPanelsAandB(CandD)presenttheresultsobtainedfortheperiod1988‐1997(1988‐2010).Top30%firm‐years(13,030firm‐years)rankedundereachconservatismmeasureareclassifiedashighconservatism,whilethe lowest 30% firm‐years (13,030 firm‐years) are classified as low conservatism.Percentage in parenthesis indicates the percentage of overlap (number of overlap firm‐years/30%oftotalfirmyears).Onlyunconditionalconservatismmodelsareincludedinthisanalysis.

Periodof1988to1997

Periodof1988to2010

PanelC:HighConservatism PanelD:LowConservatism

PanelA:HighConservatism PanelB:LowConservatism