The Importance of Audit Firm Characteristics and the Drivers of …eprints.gla.ac.uk/790/1/Beattieaccbusres1995.pdf · 2011-08-15 · The Importance of Audit Firm Characteristics

Page 1

Acciiuiilii:g ami Business Research. Vol. 25. No. 100. pp. 227-2.19. 1995 227

The Importance of Audit FirmCharacteristics and the Drivers of AuditorChange in UK Listed CompaniesVivien Beattie and Stella Fearnley^

Abstract—This paper explores the importance of audit firm characteristics and the factors motivating auditor changebased on questionnaire responses from 210 listed UK companies (a response rate of 70%). Twenty-nine potentiallydesirable auditor characteristics are identified from the extant literature and their importance elicited. Exploratoryfactor analysis reduces these variables to eight uncorrelated underlying dimensions: reputation/quality; acceptabilityto third parties; value for money; ability to provide non-audit services; small audit firm; specialist industryknowledge; non-Big Six large audit firm; and geographical proximity. Insights into the nature of'the Big Six factor"emerge. Two thirds of companies had recently considered changing auditors; the main reasons cited being audit feelevel, dissatisfaction with audit quality and changes in top management. Of those companies that considered change,73% did not actually do so, the main reasons cited being fee reduction by the incumbent and avoidance of disruption.Thus audit fee levels are both a key precipitator of change and a key factor in retaining the status quo.

Introduction

Significant changes have occurred in the UK audit-ing environment during the last decade. The relax-ation of ethical guidelines concerning advertisingand solicitation, the introduction of auditor regis-tration and monitoring and the economic recessionhave led to a relative shift in the cost and demandparameters of the audit services market and hencein competitive pressures. Observed behaviouralchanges by market participants include tendering,aggressive fee renegotiation and opinion shoppingby auditees, and low-balling (i,e, predatory pricing),response to auditees' opinion shopping behaviourand merger activity by auditors.

Although it is difficult to make causal inferences,these changes have undoubtedly had an impact onthe structure of the audit services market. Recentevidence concerning the UK listed company marketindicates an increase in both seller concentrationand the instability of client-auditor relationships(Beattie and Fearnley, 1994), Their study reportsmeasures of seller concentration for the period1987-1991 using a database of 2,079 fully listedand USM domestic companies. The increase in

*Vivien Beattie is a senior lecturer in accountancy in theDepartment of Accountancy and Finance, University of Stirlingand an academic fellow of the Institute of Chartered Accountantsin England and Wales, Stella Fearnley is a senior lecturer inaccounting in the Department of Accounting and ManagementScience, University of Portsmouth. The authors gratefullyacknowledge the financial support of the ICAEW. This paperhas also greatly benefited from the helpful comments of twoanonymous referees and the editor. Correspondence should beaddressed to Dr Beattie, Department of Accountancy andFinance, University of Stirling, Stirling, FK9 4LA.

the eight-finn concentration ratio from 0,64 to0,79 was partitioned into three categories: a 0,09increase attributable to audit firm mergers, a0,07 increase due to voluntary client-auditorrealignments and a countervailing decrease of 0,01due to the impact of market entrants,' Thus bothaudit firm mergers and voluntary realignmentswere significant causes of the rise in concentration.At the 20-firm level, however, the overall increasefrom 0,83 to 0,9 resulted primarily from voluntaryrealignments,

Beattie and Fearnley (1994) also report thata total of 341 companies (16,4%) voluntarilychanged their auditors at least once during1987-1991, Of these changes, 40% were classed asintra-audit tier changes and 60% as inter-audit tierchanges, with the tiers being defined as the topeight firms, firms 9-20 and those below the top 20,Given the overall increase in seller concentration,it is not surprising that the top tier gained thelargest number of clients.

This macro-level description of change withinthe market for audit services can, however, offerlimited insight into the competitive pressures andprocesses that give rise to the observed change.The present study therefore builds on existingresearch by examining the importance of audit firmcharacteristics and the factors motivating auditorchange at the micro-level.

The extant auditor choice literature encompassesboth the auditor change decision and the auditorselection decision. It is generally agreed that nocomprehensive, well-specified theory of either

'The concentration ratios quoted are based on number ofclients.

Page 2

228 ACCOUNTING AND BUSINESS RESEARCH

auditor change or selection currently exists,although empirical and theoretical research duringthe last 25 years has provided a theoretical frame-work within which such issues can be addressed.

To date, there has been no systematic empiricalstudy of the auditor choice process in the UK,To our knowledge, existing empirical studies relateexclusively to the US, with the exception of anAustralian study by Craswell (1988), Given thedifferent environments (especially regulatory envir-onments) that exist across countries, it is importantto conduct auditor choice studies in differentcountries, and perform cross-country comparisonsin order to gain new insights (Wallace, 1987), Thepurpose of this study is to contribute to the auditorchoice literature by examining the influences onthe auditor choice decision in a country where thisdecision has not been studied and following aperiod of rapid and significant environmental up-heaval within the auditing profession (see above),A postal questionnaire survey of listed companiesis used to elicit information concerning the import-ance of audit firm characteristics and factors thatinfiuence the auditor change decision, A randomsample of 300 companies was surveyed, resultingin 210 usable replies, which represents a 70%response rate,

A factor analysis of 29 potentially desirable auditfirm characteristics reveals a number of commonunderlying dimensions of importance to companies.The extent to which companies consider changingauditors, the main reasons for this and the reasonsfor not subsequently changing auditors are alsoexplored.

The remainder of this paper is structured asfollows. Section two reviews the extant auditorchoice literature, which forms the basis of thecontent of the research instrument. Section threeoutlines the sample and data collection proceduresemployed. The results, together with discussion,are presented in the fourth section. The finalsection summarises and concludes.

Auditor choice literatureIn conceptualising the auditor choice process it isimportant to recognise that auditor choice emergesfrom the client's characteristics, potential auditors'characteristics and the auditing environment, Asignificant change in one (or more) of these threeareas is required for a client to decide to changetheir auditor, since the costs of switching arematerial,- The auditor change process is usefullyseparated into two stages, as suggested by Francisand Wilson (1988, p, 668), since the reasons fordisplacement of the former auditor might be un-related to the specific choice criteria used in selecting

-DeAngelo {1981b, p. 188) cites the low rate at which firmschange auditors as evidence of significant switching costs.

the new auditor. Companies first decide to changeauditors and then make a reselection. Auditordisplacement may be motivated by a change incompany circumstances (i,e, by factors unconnectedwith the current audit firm's performance) such asa change in top management or by specific problemsand disagreements. The reasons for change are,therefore, not necessarily related to generic auditfirm characteristics and also not necessarily involvedin the choice of a new auditor. Although thereexists a common set of factors underlying changeand choice decisions, both decisions also haveunique factors.

The demand for audit services in free andregulated markets was examined by Wallace(1980), Three separate but interlinked sourcesof demand are identified: agency (or stewardship)demand, information demand and insurance de-mand. Agency theory, as developed by Jensen andMeckling (1976), indicates that the bonding role ofan audit can reduce agency costs arising from theself-interested behaviour of agents. The existenceof differential agency costs across clients and overtime therefore results in a heterogeneous demandfor audit services, characterised by DeAngelo(1981a, b) as different levels of audit quality.

The information demand for audit, originallyexplored by Dopuch and Simunic (1980 and 1982),is closely related to the agency demand, since it alsoarises from information asymmetries. The selectionof credible auditors not only signals management'shonesty and quality to all interested parties, butalso reduces agency costs via the monitoringfunction. The arguments of both DeAngelo andDopuch and Simunic are often referred to as the'product differentiation hypothesis'. More recentanalytical studies have employed the information/signalling framework to explore the implications ofauditor choice (Bar-Yosef and Livnat, 1984; Titmanand Trueman, 1986; and Dye, 1991), The insurancedimension of an audit forms the basis of the thirdsource of demand. It is argued that the audit servesto indemnify investors and creditors against finan-cial losses via the auditor's professional liabilityexposure.

Each of these three sources of audit demandgenerates a rank ordering of auditors, DeAngelo(1981b) argues that auditor size serves as asurrogate for audit quality, since larger firms havereduced incentives to lower audit quality opportun-istically in order to retain any single client, Dopuchand Simunic (1980 and 1982) infer that credibilityis associated with an auditor's reputation or brandname, based on the observed dominance of large

Mt has since been recognised that industry-specific reputation isan important alternative basis for rating audit firm credibility(Simunic and Stein, 1986, p. 71).

''Both credibility and reputation refer to perceived auditquality, since audit quality per se is unobservable (or at leastextremely costly to evaluate).

Page 3

AUTUMN 1995 229

audit firms in the market for publicly-held com-pany auditors.'-'' Finally, a number of writers arguethat larger audit firms have 'deeper pockets' thansmaller firms due to their higher level of insurance.

This consideration of demand-side factors sug-gests a number of client characteristics that willaffect the type of audit firm selected, in terms of itsquality and credibility. These characteristics centreon agency-related variables, such as size, level ofgearing and management share ownership. Thesefactors will determine the general size, class and/orindustry specialisation of the audit firm selected.The specific audit firm chosen is also influenced bysupply-side factors, i.e. auditor characteristics.In particular, audit firm specialisation in audittechnologies can be expected to yield economies ofboth scale and scope.'

Additional auditor characteristics of relevancerelate to the audit team, rather than the firmgenerally, and reflect the quality of workingrelationships. Other relevant client or auditorcharacteristics include the existence of significantforeign operations (Eichenseher, 1985), the exist-ence of interlocking directorates between clientcompanies (Davison et al., 1984), the associationbetween client and auditor structures (Kaplan et al.,1990) and the geographical proximity of the auditor(Stokes, 1992). Based on these arguments, client-auditor alignments can be viewed as the minimumcost match between client needs and auditorservices. A significant change in either the clientor auditor's characteristics can therefore induce arealignment if the other party is either unable orunwilling to accommodate it.

Early empirical studies collected management'sdeclared reasons for actual auditor changes usingopen-ended questionnaires (Burton and Roberts,1967; Carpenter and Strawser, 1971; Bedingfleldand Loeb, 1974; and Eichenseher and Shields,1983). The most frequently cited reasons for changewere changes in top management, need for addi-tional services/dissatisfaction with services oflered,audit fee too high, need for new financing, takeoverby another company, poor working relationshipsand technical disagreements.

Statistical studies have examined the associationbetween either (i) specific client or auditor char-acteristics and observed auditor choices, or (ii)changes in these characteristics and observedauditor changes. Firth and Smith (1992) find thatthe selection of a Big Eight audit firm (a brandname proxy for audit quality) is associated withagency cost variables and the need for signalling.The propensity to change auditors has been foundto be positively related to disagreements (DeAngelo,

^Scale economies permit firms to offer competitive fees whilescope economies, which result from knowledge spillovers, facili-tate the provision of sophisticated audit services in addition to arange of non-audit services.

1982), financial distress (Schwartz and Menon,1985), initial public offerings (Menon and Williams,1991), the early and late stages of the auditor-clientrelationship (Levinthal and Fichman, 1988), andreceipt of a qualified audit opinion (Chow andRice, 1982; Craswell, 1988; and Citron and TaflHer,1992) (although this latter finding is not supportedby Schwartz and Menon, 1985).'' The propensity tochange auditors has, however, been found not to berelated to the level of non-audit services purchased(De Berg et al., 1991).

In more recent multivariate studies, Williams(1988) finds that the propensity to change auditorsis negatively related to the incumbent's industryspecialisation and tenure and positively related tothe receipt of adverse media publicity by the clientcompany. Using panel data, Lindahl (1992) findsthat client company financial distress, size andmergers, and the market share and size of theauditor are all associated with auditor changes.Haskins and Williams (1990) focus on intra-BigEight auditor changes, finding that the mostimportant factors associated with such changes areclient financial distress, size and growth, togetherwith audit firm fee levels and industry dominance.

Johnson and Lys (1990) report that the directionof auditor change, in terms of relative audit firmsize (a proxy for differences in audit firm coststructures), can generally be attributed to changesin client characteristics that influence supplier costs(expansion, financing, profitability and auditrisk). Francis and Wilson (1988) find that, aftercontrolling for client size and growth, changesto/from a Big Eight firm are associated with agencycost variables. Healy and Lys (1986) find thatcompanies audited by a non-Big Eight firm aremore likely to retain a Big Eight acquirer followingan audit firm merger if they benefit from the BigEight firm's specialised services and/or reputation.

Studies have also examined several consequencesof auditor change. Companies that change auditorsfollowing receipt of a qualified opinion have beenshown in some studies to receive 'improved'opinions (Craswell, 1988), whereas other studiesshow no such shift (Chow and Rice, 1982 andSmith, 1986). DeBerg et al. (1991) find evidencethat the level of non-audit services purchaseddeclines following auditor change. Studies haveshown generally negative overall share price reac-tions to auditor changes, although in most casesthere are no systematic abnormal returns (Friedand Schiff, 1981; Nichols and Smith, 1983; Smith,1988; and Johnson and Lys, 1990).

'Krishnan (1992) distinguishes 'internal" opinion shopping,where incumbent auditors and companies negotiate to reach amutually acceptable opinion, from 'external" opinion shopping(a substitute behaviour), where companies search for moreaccommodating auditors. He finds that auditor changers, asfailed internal opinion shoppers, face stricter standards thannon-changers in the year prior to change.

Page 4

230 ACCOUNTING AND BUSINESS RESEARCH

There is, however, evidence that auditor changesassociated with disagreements, qualified auditopinions and changes away from Big Eight auditorsresult in statistically significant negative pricemovements, especially at high levels of manage-ment ownership (Smith, 1988; Eichenseher et al.,1989; and Albrecht and Lamy, 1992). With regardto audit fee cutting following change, Francis(1984) finds no evidence of this whereas Simon andFrancis (1988) and Ettredge and Greenberg (1992)find an average discount in the initial year of 24%and 25% respectively.'

MethodsSample Selection and Construction of Mailing List

The sample is drawn from the population ofdomestic officially listed and USM companies inthe UK and Ireland as at 30 April 1992. An alpha-betical listing of company names was obtainedfrom Extel Financial Ltd. This data set also con-tained addresses and the name of either the financedirector or (in the absence of this) the companysecretary. A systematic sample of 300 was selectedfrom this list, i.e. every seventh company.

Research Materials and Questionnaire Administra-tion Procedures

The questionnaire used closed-form questionsand contained six sections. The first section con-tained general questions, while the second sectionelicited the importance of 29 audit firm character-istics using a five-point Likert-type scale withverbal anchors. Questions in the second sectionare framed as a generic task, in that no specificreference is made to the company's current auditfirm. The third section identified those companieswhich had considered making a voluntary changein auditors during the past five years, elicited howseriously this change had been considered andasked the respondent to indicate which of 26reasons caused a change to be considered. Theremainder of the questionnaire forms part of aseparate paper. The draft questionnaire was pre-tested with the assistance of several senior businessexecutives and audit partners and the content,ordering and terminology was revised accordingly.

The reasons offered for considering an auditorchange resulted from an analysis of the declaredreasons for change cited in prior questionnairestudies and hypothesised variables from statisticalauditor change studies, with 26 distinct themesbeing identified. The 29 auditor characteristics

'Simon and Francis (1988) found that this fee discount in theinitial year persisted, at reduced levels, for approximately twoyears, and appeared to result partly from "low balling'. Ettredgeand Greenberg (1992) found that the size of fee discounts in theinitial year was associated with changes in audit quality andtechnological efficiency and the number of auditors bidding onthe engagement.

whose importance was rated by respondents werederived by extracting relevant themes from thepotential change drivers (rewording as necessary)and from a review of the auditor choice literature.Several additional characteristics were added duringpiloting.

The questionnaire was accompanied by anexplanatory covering letter which assured theconfidentiality of responses. A return envelope wasalso provided. The questionnaires were seriallynumbered to permit non-respondents to be followedup. A reminder letter was sent out after 10 days,with a second request (accompanied by a duplicatecopy of the questionnaire and the original coveringletter) being sent after a further 10 days.

ResultsResponse Rate and Tests for Bias

From the total sample of 300, 225 responseswere received. Fifteen responses were not usablefor a variety of reasons, leaving 210 usable replies,representing a response rate of 70%.** This ratecompares very favourably with those obtained inrecent UK studies covering similar populations ofcompanies and respondents.

A number of tests for response bias wereperformed. First, responding and non-respondingcompanies were compared on the basis of size(measured as total assets) using the Wilcoxon-Mann-Whitney non-parametric test. The hypoth-esis that the two groups have been drawn from thesame population was not rejected (a = 0.05). Thiscomparison was also applied to early (first 70)and late (last 70) responders, on the assumptionthat late responders are similar to non-responders(Oppenheim, 1966, p. 34). Again, the null hypoth-esis of no difference was not rejected (a = 0.05).'Finally, the companies were formed into four broadindustrial groupings based on Stock Exchangegroups (capital goods, consumer goods, other andfinancial). A chi-squared test indicated an associ-ation between industry and response/non-response(a =0.01). The percentage of responders washighest among the 'other' group (86%) and lowestamong the financial group (61%). Despite thelimitations of such tests (discussed by Wallaceand Mellor, 1988), given the high response rateobtained, we conclude that response bias is not aserious threat to the validity of our results. Table 1

"Five companies indicated that they did not wish to partici-pate due to lack of management time, four said that theircompany policy was not to participate in questionnaires of thisnature, three said that their auditor choice was in the hands ofa parent company or comparable to another company in thesample, two said that the addressee had left the company andone company was in voluntary liquidation.

'Further tests for response bias comparing early and lateresponders were not conducted since the questionnaire did notcontain suitable key summary variables.

Page 5

A U T U M N 1995 231

Table 1Analysis of Respondents by Company Size andIndustry Group

]

Size group

SmallMediumLarge

^otal assets(£m)

<2020-100

^100

No.

748052

206*

*Data on four newly listed companiesavailable.

Industry group

Capital goodsConsumer goodsOther (including

three oil and gascompanies)

Financial

Responserate

74%63%61%

86%

No.

704850

42210

%

35.938.825.3

100.0

were not

%

33.322.923.8

20.0100.0

provides an analysis of the respondents by size andindustry.

A further factor that can affect the validity ofquestionnaire responses is the suitability of indi-vidual respondents, who should be both knowl-edgeable about the relevant practices, and involvedin the relevant decisions at a senior level. Based ontitle, the respondents were: finance director (58%),company secretary (12%), financial controller (9%),director (6%), accountant (5%), chief accountant(3%) and other/not stated (7%). It is clear thatnearly all respondents are senior executives whowould be intimately involved in the auditor choicedecision.'" The risk of uninformed respondent biasis therefore minimal.

Audit Finn CharacteristicsSection two of the questionnaire asked respond-

ents to indicate the importance to their company ofeach of 29 audit firm characteristics, using a scaleof 1 to 5. The questions did not relate specifically tothe company's current audit firm, rather they wereset in a generic context. The results are summarisedin Table 2. Over the group as a whole, the threemost important characteristics to emerge were:(i) integrity of firm; (ii) technical competence offirm; and (iii) quality of working relationship withaudit partner(s). When each individual respondentwas asked to identify the three most importantcharacteristics to them, technical competence offirm service was most frequently cited over the threeslots (n = 86), followed by value for money of audit

"The questionnaire revealed that it is the finance directorwho is most frequently responsible for negotiating the audit feewith the auditors and who is most influential in the appointmentprocess (detailed results not reported here).

(n = 84) and quality of working relationship withaudit partner (n = 59) (see Table 3). Interestingly,Big Six audit firm was cited by 40 companies asa top three characteristic, despite the fact that itranked only 20th for the group of respondents asa whole.

Clearly there exist significant within-group vari-ations in preferences with respect to audit firmcharacteristics. We investigated the extent to whichcompany size or industry group could explain thesedifferences. First, we classified the 210 respondentsas small or large based on the median level oftotal assets. Using a two-tailed t-test, statisticallysignificant differences (at the 10% level) betweenthe two groups in terms of their responses to the29 characteristics in Table 2 emerged for 11 charac-teristics (see third-last column of Table 2). Ten ofthese characteristics were rated as of significantlymore importance by large companies: technicalcompetence of audit partner; ethical standardsof audit partner; technical competence of auditengagement staff; audit quality; willingness to offerguidance on accounting principles; quality of adviceto management; Big Six audit firm; acceptability tocompany's regulators; specialist knowledge of yourindustry; and existence of offices located close toprincipal accounting functions. One characteristicwas rated as of significantly less importance bylarge companies: low absolute level of audit fee.

Second, we classified the 210 respondents intofour broad industrial groupings based on StockExchange groups (capital goods, consumer goods,financial and other) and conducted a Kruskal-Wallis test to determine whether any statisticallysignificant differences existed among the fourgroups (Siegel and Castellan, 1988, pp. 206-212).Significant differences (at the 10% level) emergedfor seven of the 29 characteristics: quality of work-ing relationships with audit partner; willingness tooffer guidance on accounting principles; same auditfirm as other group companies; willingness toprovide detailed cost information; specialist knowl-edge of your industry; existence of offices locatedclose to principal accounting functions; and a localaudit firm (see second-last column of Table 2).The first four of these characteristics were mostimportant to capital goods companies, the fifthand seventh to financial companies and the sixth toconsumer goods companies.

Subsequent multiple comparison proceduresbased on Tukey's studentised range test (SAS, 1990,ch. 13) revealed, however, that in the case ofthe second, sixth and seventh characteristics listedabove, no pairwise comparisons were significant atthe 5% level." Capital goods companies rated the

"This lack of any significant pairwise differences is notunusual, since the Tukey test is conservative in that it controlsthe type 1 experimentwise error rate.

Page 6

^

^

-C) *... — O * ^ ry^O—"OfN r/\O OOO '' OOOO"?

A 5 c/:i i/i w:i i/i c/:i i/i c/5 § c/:i c/:i 5 vj g 1/31/3 p:; V3

a; A d>d><~> ZddddZ dZ d ZZZZZZZdZZdZ dZZdZ

a; A ZZd

•§[;;;<_!;« 00 i/:j

i; A zzz

00 00 [/5 1/5 !/3 1/5

zzzzzz1/3 g 1/5

Z d Z

ooooggoooooooriooi/igs

ZZddZZZZZZdd

00

d d

00o

ZZZZZZZddddd

§ • * ON ON5; (^ r^ r^ NO 00 00 00 — fO ON O

t^ O

p.

a:g

1..

inoin oo in^oo ON"d — d o — - d — • — • — (Nr<S

rnoOTt ONrnooooqo ^t^^ f*1 fN ^™ f 3 f ^ "™ ^ "

— — — fN

Odirt —fN fN m

O in ON •^ —• in ONfN Tj- Tj- fN 06 in NO fN

O__„ „__„___ —fN fN

ON00

00—'fNf^rnrOTt-^inNor^fNfNfNfNfNfNfNfN(N<N<N(N

1/3 1/3 1/3 1/3 gZZZZd

c/3 c/3 1/3 !/3 !/3

z z z z z

„ __ _ o o

fNj (-rj (^ ^ - \Qr^ r*~i c*~i Tt Tf

(N f* ON —• 00 00 ' ^ Tf (N

— — rN \O r^

r^m<7\ONCN^oor^'>o — 0000 Tj-oomr-oo

(^ ^- ^- r-j

r-. 00 0000 fN r O 0 0 C^ O^ r ^ ^™ r ' l CT\

(N — vO TT 00^ O O rn fN(N <N <N

C

aBoc

T3Cca

cca

oa.E

E

I• o3acca

t r-i in fN r-; NO^ ON Tt —" in .—•) in in in Tj- Tj-

^ ^ ON ^ ^ NO ON ON ON 0 0 ONs: _o— oooooo

fNfNfN fNfN<N<NfNfN

it:ca

oo <NO <

•c

<

es

-•I

-o3ca

Q.

g 2P

.5

I-5

1

Inte

gri

ty (

Tech

nic

al

Qu

ali

ty 0

E to-.&•£c« ca •—

a g g g a.•s •> " .t; •=- u a3 S

(u o ° ai S

mm

3 E

tner

au

dn

ce

rep al

part

ner(

Good r

eT

ech

nic

aV

alu

e fo

Eth

ical s

Tech

nic

aQ

uali

ty

CO 0 0

5 'S

" O <L>

2c

— 3S 00

ts >>2

IIIS o - c

g<^

NOofN

00c

•53

u

djsoca

jinONO — fN — — OfNONjinfNod-^inONfNinr-^inI'S'fNfNfNfNfNfN — — —

ON t ^ fN ^ ^ I ^ f^ NO NO 0 0 I—• ON 0 0ooo-~oooooooofNfNfNfNfN(NfNfNfN<N<NfN

M n

C c/,

•2 fe s2

52 W5 U- Cd « Crt Crt

flj ^^ ^^ ^^

c« x ; .

s l

ula

to

CO

13

' OJ . i :

°£-'••= 5 K

- | 1 | E E E-* r ^ r ^ r ^

D.D.D.c/3 ca

0 0Sc

E o

Os O^ ^^ 00 O\(N (N —. — —

4> crt r

u oo__c c caCO ' S C

3 §-2<£ O ^

§ 8 =^ 1 oca ca —c c ca

.2.2 §

e , ? ? l Ec ca CO V. c.2 OJ u 2 « pS :S .-H .E .r E

3 . - . - g 00 3O — :r 2 <u oox>X) = '

>ca

o&>>

11

fe«8SSo888c8-3-2g

u-| 2 E 2 2 2c^ 2 i i i E

S " ^E-ti S2 = = = "i = .2-^ o

njra-'rxraTOra TOO- —

c a x ) o S t r c j o c j . S P o o Q . x

E<<c/3^<<<pa<JoouJ

c >ca c/3

O. c caE cO O.

. . - O.CU -^ P o

ca •= " u•j ; (2 «J g

(_ C XJo ca - c^ I.. C CO

u o c a •-.

0 0 ^ p E

C f N . 5 —' « L^ C C^CO "H" c<, n,i> c ..Si ~~

ca .._ >N

CJ i3 ::>,„•a o r- -

-5.1 cSs

illlc/3 OJ ^ c/3

u -^ o u

.S2 " 3 . 2 2 . i £OJ

tj

e

<N • m NO 1^ 00 ON o — (

_ c o 0ca o ca caca c/3 ca

232

Page 7

AUTUMN 1995 233

Table 3The Three Most Important Auditor Characteristics

1.2,3.

4.5.

Auditor characteristic

Technical competence of firmValue for money of audit serviceQuality of working relationshipwith audit partner(s)Integrity of firmBig Six audit firm

Mostimportant

38'20'=18"

20'"26^

Individuai ranking frequency *2nd mostimportant

29'24-16"

20'8"=

lird mostimportant

19'40'25^

18"5"=

Totai"'

868459

5840

Group rank{from

Tabie 2)

263

120

Notes:*Superscripts indicate the rank of characteristics within each top three slot (thus, for example, 18 respondents

cited 'quality of working relationship with audit partner(s)' as the most important auditor characteristie andthis was the sixth most frequently cited sueh reason).

•Auditor characteristics are shown in decreasing frequency of total citations across the three top slots.

quality of working relationship with audit partnerand common group auditor of significantly moreimportance than financial companies, and willing-ness to provide detailed cost information of moreimportance than consumer goods companies.Financial companies rated specialist industryknowledge of more importance than the otherthree industrial groups.

Since 41% of the respondents had auditcommittees, we investigated whether this featureinfluenced the importance attributed to auditorcharacteristics. The final column of Table 2 reportsthe two-tailed significance level of a t-test betweenthe two groups. Significant differences (at the5% level) emerged for 13 of the 29 characteristics.With the exception of 'ability to provide additionalconsultancy services' and 'a regional audit firm',these characteristics were rated of more importanceby companies with an audit committee.

It is to be expected that many of the 29 auditfirm characteristics will be highly correlated. Anexploratory factor analysis was performed usingthe principal components method with varimaxrotation in an attempt to uncover the criticalunderlying dimensions. Eight initial factors wereextracted, based on the eigenvalue ^ 1 criterion(Kim and Mueller, 1978, p, 49). These factorsexplained 64% of the variance among the auditfirm characteristics.

Table 4 summarises these eight factors, providinga subjective factor label based on each factor's con-stituent auditor characteristics. Those constituentcharacteristics with factor loadings greater than10,501 (i,e, 25 out of 29) are also shown, togetherwith their factor loadings. This analysis reveals theimportance of auditors' acceptability to third partiesand ability to provide non-audit services (factorstwo and four respectively), which are auditorcharacteristics not rating highly in Table 2. It isinteresting to note that the auditor characteristicof being a Big Six firm does not emerge from the

factor analysis. In fact this characteristic appearsto be multidimensional, being positively associatedwith factors 1, 2, 6 and 8, and negatively associatedwith factor 5 (the relevant factor loadings are0,24, 0,28, 0,36, 0,37 and -0,29 respectively).This finding provides empirical support for thearguments that these firms enjoy a reputation thatmakes them more acceptable to third parties, thatthey are more likely to possess specialist industryknowledge and be located nearby. There areclearly, however, a significant number of com-panies which reject such firms in favour of eithersmall audit firms (factor 5) or larger, non-Big Sixaudit firms (factor 7). Nor does the eighth-rankedauditor characteristic 'technical competence ofaudit engagement staff' appear in the factoranalysis. This characteristic is two-dimensional,with positive loadings of 0,48 on factor 1 and 0,46on factor 3.

It is noticeable from Tables 2 and 4 that thereare several auditor characteristics that are notranked of high importance, but which neverthelessload on factors. In particular, factor 2 (accept-ability to third parties) comprises five relatedcharacteristics ranked 14, 17, 18, 19 and 21. Thisresult can be attributed to the existence of fivequestions which clearly relate to the same dimen-sion. While most respondents were concerned aboutthe auditor's acceptability to a specific third party,there were very few concerned about more thanone group (this is apparent from the high standarddeviation in Table 2). Thus the average importanceof each characteristic was low. In addition, factors4-8 include characteristics of middle and low im-portance. There is, in fact, a distinct drop betweenfactors 1-3 and 4-8 in the proportion of varianceexplained by these factors.

Auditor Change ConsiderationSection three of the questionnaire asked our

210 respondents to indicate how seriously they

Page 8

234 ACCOUNTING AND BUSINESS RESEARCH

•c

poi

E"en

•sen

cter

ito

r C

hara

is o

f A

udi

1

rank

ing

I

dito

r

1

Pri

rla

bees

crip

iion

s l« 5

-

of

firm

ity

o:ite

gr

c

uali

tyut

atio

n/q'

Rep

i

—

. . .

- p!o o

of f

irm

8C

E 'so 2

ical

(re

pu

c -o

l §H O

o o

u

t :

or p

£of

aud

it

8cw

ical

(qu

al;c

hnud

it

^ Soo o

;ner

lder

spa

rt's

ler

.'S >.

ca D.

^5

•§ o

1 st

a:ta

bil

thic

acc

ep

en

tie

CO

"2

ok«

of a

udit

ojp

tabi

lity

8<

- -

o o

enal en

s -

•o c

y-to

-;r

cha

•3 E

>. >-,

p" p"c c

o o

tabi

lta

bil

8 8< <

o

o

•I

g-p

o

Ity

t(ta

bil

a.

s

o

en

"o

reh

si&o

tabi

l

o.

8

o

co

1<2•E

1 co

stde

taile

c

O O

taff

C

ECO

so

3ca

•s- 8

atio

nshi

dit

serv

i

s to

pr

o'w

orki

njno

ney

o

gnes

:'il

lin

o

lati

t

nfor

ley:

cos

t i

o

<2

Val

uua

lit

alue

O>

ips

-ela

tion

shw

orki

ng i

and

-

o

en_U

•§•.P

PL

inti

o

i

idan

ce c

3SO

gnes

s to

off

e'il

lin

o

cO

•5£,

p w

itat

ions

hire

l;w

orki

ng

CM

o>

uali

t

o

ces

• >

Iit

iona

l t;

•o•oca

ide

3rov

o

bili

t'

c/1

8

ser

audi

vide

non

-lt

y to

pro

Abi

l

• > * '

-o

en

>

COon

suit

iona

l c

T3T3CO

ide

3rov

o

bili

t'

o

enUo>

ser

c

nti

3

8it

iona

l a

T3T3CO

ide

jrov

o

bili

t'

00

o

gI P

aud

onal

regi

^

E

11 au

dit f

iiSm

a

as

o o

u

indu

sif

clie

nt i

o

u

>

ser

so

nti

ccou

itio

nal

ao ^1) "Oso

litf

i:n

ov

ca en

B133C

B30|

U l

A'le

djs

try

kno'

iaii

st i

ndi

Iin

CO

ide

jrov

o

bili

t'

'^

o

g

3CO

co

nati

i

ocak«

inte

Six

-Big

non

^

E

•53ca

onal

inte

r)na

ti-B

ig S

ix (

Non

o

(11

anag

emi

Eo

advi

ce t

CMo>-.

uali

t

a

o

enPO

unct

(^

cc3o8

cao.

lnci

o.

se t

oat

ed c

lo

CJ

)f of

fices

8

)cis

te

M

prox

imit

'gr

aphi

cal

a

oo

Page 9

AUTUMN 1995 235

Table 5Stated Reasons for Consideration of a Change in Auditor

Reason *

1. Level of audit fee2. Dissatisfaction with audit quality (i.e. auditor's ability to detect problems)3. Changes in company's top management4. Company growth required increased technical capacity from audit firm5. Need for group auditor rationalisation6. High turnover of audit engagement staff7 = . Merger/takeover with/by another company7 = . Need for additional services9 = . Audit firm merger9 = . Use of inexperienced audit engagement staff

11. Need for Big Six audit firm12. Personality clashes with audit partner/staff13 = . Change in audit partner13 = . Poor working relationships with audit partner/staff15 = . Need for multinational audit firm with foreign oflices in same geographical

areas as client's operations15 = . Disagreements over accounting principles17 = . Influence of merchant bankers/underwriter17 = . Inaccessibility of audit partner19 = . Influence of actual or potential equity or loan providers19 = . Need for audit firm specialising (or not) in client's industry21. Need for national, rather than local, audit firm22. Disagreement with audit opinion23 = . Influence of regulators23 = . Need for audit firm with local domestic office25. Influence of company's day-to-day bankers26. Need for local, rather than national, audit firm

*Reasons are shown in decreasing frequency of citation.

Number ofcompanies

9246342524222020191918171515

1212101077654431

%(n = 139)

66332518171614141414131211II

997755443321

had considered changing their auditors in thepast five years, and the reasons for this (selectedfrom a list of 26 potential reasons). Only 34% hadnot considered changing their auditors at all, while24% had considered the matter very seriously,11% seriously, 10% fairly seriously and 21%casually. Table 5 provides a complete breakdownof all the stated reasons cited as contributing to theconsideration of a change in auditor. It should berecognised that a number of the reasons cited over-lap to some extent. For example, 'use of inexperi-enced audit engagement staff' (item ranked ninthequal) may contribute to an overall 'dissatisfactionwith audit quality' (item ranked second). The 139companies cited a total of 467 reasons, an averageof 3.3 contributory reasons each.

The two most common reasons cited related toproblems with the audit per se and were the levelof audit fee (66%) and dissatisfaction with auditquality (defined as the auditor's ability to detectproblems) (33%). Eighteen other reasons were,however, cited by a significant (i.e. ^5%) percent-age of companies. Of these reasons, five related toproblems with the quality of the relationship withthe audit team (high turnover of audit engagementstaff, use of inexperienced audit engagement staff.

personality clashes with audit partner/staff, poorworking relationships with audit partner/staff andinaccessibility of audit partner), eight related tostructural changes at the auditee, (top managementchanges, growth, auditor rationalisation, merger,and need for additional services. Big Six firm, multi-national firm or industry specialist), two related tochanges concerning the audit firm (audit firmmerger and change in audit partner), two related tothird party infiuences (merchant banker/underwriterand capital providers) and one concerned disagree-ments over accounting principles.

Of the 56 companies that had undergone an('« voluntary change in auditor due to an audit firmmerger (not shown in Table), 19 (33%) cited thisas a contributory reason for considering furtherchange. It may be noted, given the recent CadburyReport (1992) recommendations on the rotation ofaudit partners, that 11 % of respondents cited thisas a reason for considering an auditor change.Disagreements concerning accounting principleswere cited by 9% of companies as a reason forconsidering auditor change, with 4% citing dis-agreement with the audit opinion. This is suggestivethat auditor changes may in some cases arise from'opinion shopping'.

Page 10

236 ACCOUNTING AND BUSINESS RESEARCH

We again segmented the respondents into thosewith/without an audit committee. Based on chi-squared tests, there was no significant association(at the 5% level) between any of the top five statedreasons for considering change and the existence ofan audit committee.

Also using a chi-squared test, we investigatedwhether the stated reasons for consideration ofauditor change were associated with the seriousnesswith which change was considered. To avoid lowcell frequencies, adjacent categories were combinedto form a 'more serious' group (n = 73) and a 'lessserious' group (n = 66), Four reasons were citedsignificantly more frequently (at the 5% level) bythe more serious group: dissatisfaction with auditquality (p = 0,001); company growth (p = 0,002);auditor rationalisation (p = 0,048); and need forBig Six firm (p = 0.001), These issues relate prim-arily to structural changes at the auditee. Interest-ingly, however, the most frequently cited reasonoverall (level of audit fee) showed no statisticallysignificant difference.

When asked to rank the three most importantreasons for considering an auditor change, a similarpattern of responses to that in Table 5 emerged, ascan be seen from Table 6,

Clearly there will, in many cases, exist a relation-ship between the desirability of audit firm charac-teristics and change drivers, since if a companyis considering a change of auditor, then this ritayindicate that its current auditor is not satisfactorywith respect to a characteristic which is rated

important (although there can be other reasons,e,g, an auditor rationalisation policy). We there-fore investigated the consistency of the relationshipbetween the three most important change driversand the desirability of audit firm characteristics.We hypothesised that the 45 respondents who cited'level of audit fee' as the most important reasonfor considering auditor change would rate 'lowabsolute level of audit fee' as a more importantaudit firm characteristic than the other respondentswho had considered change. Based on a t-test, astatistically significant difference in the expecteddirection did exist between the two groups (t = 1,79,p = 0,038, one-tailed).

Similarly, we hypothesised that the 24 respond-ents who cited 'dissatisfaction with audit quality'as the second most important reason for consider-ing auditor change would rate 'audit quality' as amore important characteristic than the others. Inthis case, although the difference between the twogroups was in the expected direction, it was notstatistically significant ( t=1.08, p = 0,144), Thethird most important reason for considering auditorchange, 'changes in company's top management',has no related auditor characteristic.

Of the 139 companies that had considered chang-ing their auditors, 37 (27%) did actually effect achange while 102 (73%) did not. The reasons givenby the non-changers are summarised in Table 7,Fee reduction by the incumbent auditor was mostfrequently cited, either singly or in combinationwith other factors. The most common single reason

Table 6The Three Most Important Stated Reasons for Considering Auditor Change

Stated reason

1, Level of audit fee2, Dissatisfaction with audit

quality (i,e, auditor's abilityto detect problems)

3, Changes in company's topmanagement

4, High turnover of auditengagement staff

5, Need for group auditorrationalisation

6, Merger/takeover with/byanother company

7, Use of inexperienced auditengagement staff

Notes:•Superscripts indicate the citation frequency of stated reason within each top three slot (thus, for example,

six respondents cited 'changes in company's top management' as the most important reason for consideringauditor change and this was the fifth most frequently cited such reason),

•Stated reasons are shown in decreasing frequency of total citations across the three top slots.

Mostimportant

45'24=

,16 =

47 =

9'

213 =

Individual ranking frequency *2nd most 2rd mostimportant important

23'6"=

7'=

9=

73=

410

, 1 3 =

10' =10' =

6"

55

46=

—

8'

Total'

7840

19

15

15

13

11

Group citationfrequency

(fromTable 5)

12

3

6

5

7=

9=

Page 11

AUTUMN 1995 237

Table 7Reasons for Not Changing Auditors Following Consideration of Change

Reason

1. Incumbent auditor offered reduced audit fee2. Avoidance of disruption and loss of management time3. Incumbent auditor offered improved quality of service4. Change of audit partner5. 'Other'

Note: Reasons are shown in descending frequency of combined frequency of citation.

Solereasoncited

1926143

13

Number of companies

Contributor)reason

231314—

7

Combined> frequency

of citation

4239283

20

%{n = 102)

413827

320

given was, however, avoidance of disruption andloss of management time. Three respondents men-tioned change of audit partner as the single reasonfor not proceeding with the change. 'Other' reasonsincluded inertia (four cases), decision pending(three cases), satisfaction with incumbent auditors(three cases), lack of suitable alternatives in thearea (one case), advice from merchant bank (onecase) and new shareholder did not insist (one case).

We examined the association between the topthree most important stated reasons for consideringauditor change and the main sole reasons for notsubsequently changing auditors. Where level ofaudit fee prompted auditor change consideration,50% did not change because the incumbent offereda reduced audit fee, 35% did not change to avoiddisruption and loss of management time and15% did not change because the incumbent offeredan improved quality of service. The correspondingfigures for dissatisfaction with audit quality andchanges in company's top management were 12%,44%, 44% and 25%, 25%, 50% respectively.

DiscussionThe auditor choice process is clearly a complex

one. Not only does it involve multi-attributedecision analysis, on the part of one or more indi-viduals, but these attributes are constantly chang-ing within a dynamic decision setting. Despite this,our analysis of 29 potentially desirable audit firmcharacteristics does reveal eight uncorrelated under-lying dimensions: reputation/quality; acceptabilityto third parties; value for money; ability to providenon-audit services; small audit firm; specialistindustry knowledge; non-Big Six large audit firm;and geographical proximity. These findings aregenerally consistent with the findings of Lynn (1987)in the US and of a recent UK survey (Fitzgerald,1992). The UK survey was restricted to the con-sideration of six auditor characteristics and coveredbusinesses of all sizes, although no precise samplingdetails are reported. Regular partner contact wasranked most important while the lowest competitiveprice was ranked least important.

Given the current debate concerning audit feecompetition, an interesting finding to emerge fromthe present study is that, although value for moneyis ranked as the sixth most important audit firmcharacteristic in general, the low absolute levelof audit fee is ranked only 22nd (see Table 2).In contrast, fee level is the most frequently citedreason for consideration of a change in auditor.Fee reduction by incumbent auditors was also themost frequently cited reason for not changingauditors. There are two possible explanations ofthis result. First, it should be noted that the valuefor money concept has two dimensions—price andquality. During piloting, value for money was aconcept suggested to us several times as an im-portant auditor characteristic, however it was notsuggested to us as a change driver. It would appearthat when considering change, auditees think interms of price or quality, rather than their jointeffect. We therefore maintained this perceptualdistinction in the questionnaire. Second, it may bethat expected audit fees exhibit greater variationbetween audit firms than other auditor character-istics, and thus will induce change more frequently.

In comparing Table 5 with Tables 2 and 4 it canbe seen that in many cases auditor change isconsidered due to an emergent problem whichaffects a highly ranked auditor characteristic. Forexample, a high turnover of audit engagement staffhas an adverse impact on the quality of workingrelationships with audit engagement staff. In othercases, a change in the structural characteristicsof the auditee (e.g. company growth) changes therelative importance of auditor characteristics, withthe incumbent no longer able to provide the re-quired service. In a final class of cases, the structuralcharacteristics of the auditor change (e.g. auditfirm merger), with the result that the incumbentagain fails to provide the service required.

ConclusionThe purpose of this paper is to contribute towardsan understanding of auditor/client relationships.

Page 12

238 ACCOUNTING AND BUSINESS RESEARCH

since the changing nature of these relationships, inaggregate, profoundly affects the market for auditservices in the UK. Questionnaire responses from210 listed companies are analysed to reveal funda-mental aspects of the auditor choice process, suchas the importance of audit firm characteristics, thereasons for changing auditor and the reasons forretaining the incumbent.

Significant within-group variation existed incompanies' views on the desirability of audit firmcharacteristics. This variation is partially explainedin terms of company size and industry group.Large companies rated several characteristics(including technical competence, quality issues andspecialist industry knowledge) of significantly moreimportance than small companies; large companieswere, however, less concerned about the absoluteaudit fee, perhaps because it represents a smallerproportion of operating costs for such companies.Capital goods companies rated the quality of work-ing relationship with audit partner and commongroup auditor of significantly more importance thanfinancial companies, and willingness to providedetailed cost information of more importance thanconsumer goods companies. Financial companies,not surprisingly, rated specialist industry knowl-edge of more importance than the other threeindustrial groups.

Factor analysis revealed eight uncorrelated auditfirm dimensions of importance to companies—the top four being reputation/quality, acceptabilityto third parties, value for money and ability toprovide non-audit services. The nature of the muchdiscussed 'Big Six factor' is partially revealed byour findings. This factor is itself multidimensional,encompassing, in particular, reputation/quality,acceptability to third parties and specialist industryknowledge. This provides empirical support for thequality, credibility and specialist knowledge argu-ments of DeAngelo (1981b), Dopuch and Simunic(1980 and 1982) and Simunic and Stein (1986)respectively.

Although the absolute level of audit fee was notranked as an important audit firm characteristicover the sample as a whole, audit fees were theprincipal cause of consideration of auditor change.The 45 respondents who cited 'level of audit fee' asthe most important reason for auditor change alsotended to rate 'low absolute level of audit fee' asimportant. Changes in the structure of the auditfirm, changes in the personnel of the audit teamand top management changes within the auditeewere also common destabilising influences. Thesechanges principally include audit firm merger,change in audit partner (which is recommendedin the Cadbury Code of Best Practice (1992)) andturnover of audit staff. Moreover, two-thirds of allcompanies have, to a greater or lesser degree,considered changing their auditors recently. Thismay be taken as an indication of underlying

instability in auditor/client relationships. Of thosewho considered change, but did not actually change,fee reduction by the incumbent was the mostfrequently cited reason for retaining the status quo.

This paper has two main limitations which aresuggestive of further research. First, it must beemphasised that we have investigated the reasonsfor the consideration of auditor change and notthe reasons for actual changes. Both issues areof interest in their own right, with the formerpermitting inferences about the latter. The numberof respondents within our sample who actuallychanged auditors is quite low (n = 36), and doesnot permit meaningful analysis.'• Second, this paperis based on the analysis of declared responses,collected by means of a postal questionnaire. Cor-roboration of these results using publicly availablesecondary data is desirable.

ReferencesAlbrecht, W. D. and Lamy, R. E. (1992), 'The Determinants

of the Market Reaction to an Announcement of a Change inAuditor', AAA conference paper.

Bar-Yosef, S. and Livnat, J. (1984), 'Auditor Selection:An Incentive-Signalling Approach", Accounting and BusinessResearch, Autumn, pp. 301-309.

Beattie, V. A. and Fearnley, S. (1994), 'The Changing Structureof the Market for Audit Services in the UK—A DescriptiveStudy', British Accounting Review, December, pp. 301-322.

Bedingfield, J. P. and Loeb, S. E. (1974), 'Auditor Changes—An Examination', Journal of Accountancy, March, pp. 66-69.

Burton, J. C. and Roberts, W. (1967), 'A Study of AuditorChanges', Journal of Accountancy, April, pp. 31-36.

Cadbury Report (1992), The Financial Aspects of CorporateGovernance (Gee & Co.).

Carpenter, C. G. and Strawser, R. H. (1971), 'Displacement ofAuditors When Clients Go Public', Journal of Accountancy,June, pp. 55-58.

Chow, C. W. and Rice, S. J. (1982), 'Qualified Audit Opinionsand Auditor Switching', Accounting Review, April, pp. 326-335.

Citron, D. B. and Taffler, R. J. (1992), 'The Audit Report underGoing Concern Uncertainties: An Empirical Analysis',Accounting and Business Research, Vol. 22(88), pp. 337-345.

Craswell, A. T. (1988), 'The Association Between QualifiedOpinions and Auditor Switches', Accounting and BusinessResearch, Winter, pp. 23-31.

Davison, A. G., Stening, B. W. and Wai, W. T. (1984). 'AuditorConcentration and the Impact of Interlocking Directorates'.Journal of Accounting Research, Spring, pp. 313-317.

DeAngelo. L. E. (1981a). 'Auditor Independence. "LowBalling", and Disclosure Regulation'. Journal of Accountingand Economics, Vol. 3, pp. 113-127.

DeAngelo. L. E. (1981b), 'Auditor Size and Audit Quality'.Journal of Accounting and Economics, December.pp. 183-199.

DeAngelo, L. E. (1982), 'Mandated Successful Efforts andAuditor Choice', Journal of Accounting and Economics,Vol. 4, pp. 171-203.

DeBerg, C. L., Kaplan, S. E. and Pany, K. (1991), 'AnExamination of Some Relationships Between Non-AuditServices and Auditor Change', Accounting Horizons, March,pp. 17-28.

'•This sample size is extended in a follow-up study whichfocuses on actual changers.

Page 13

A U T U M N 1995 239

Dopuch, N. and Simunic, D. (1980), 'The Nature ofCompetition in the Auditing Profession: A Descriptive andNormative View', in Buckley, J. W. and Weston, J. F. (eds.).Regulation and the Accounting Profession (Lifetime LearningPublications).

Dopuch, N. and Simunic, D. (1982), 'Competition in Auditing:An Assessment', Fourth Symposium on Auditing Research(University of Illinois).

Dye, R. A. (1991), 'Informationally Motivated AuditorReplacement", Journal of Accounting and Economics, Vol. 14,pp. 347-374.

Eichenseher, J. W. (1985), 'The Effects of Foreign Operationson Domestic Auditor Selection', Journal of Accounting,Auditing and Finance, Vol. 8(3), pp. 195-209.

Eichenseher, J. W., Hagigi, M. and Shields, D. (1989), 'MarketReaction to Auditor Changes by OTC Companies', Auditing:A Journal of Practice and Theory, Fall, pp. 29-40.

Eichenseher, J. W. and Shields, D. (1983), 'The Correlatesof CPA—Firm Change of Publicly-Held Corporations',Auditing: A Journal of Practice and Theory, Spring, pp. 23-37.

Ettredge, M. and Greenberg, R. (1990), 'Determinants of FeeCutting on Initial Audit Engagements', Journal of AccountingResearch, Vol. 28(1), Spring, pp. 198-210.

Firth, M. and Smith, A. (1992), 'Selection of Auditor Firms byCompanies in the New Issue Market', Applied Economics,Vol. 24, pp. 247-255.

Fitzgerald, N. (1992), 'Survey Findings: Putting a Priority onClarity', CA Magazine, November, pp. 24-33.

Francis, J. (1984), 'The Effect of Audit Firm Size on AuditPrices: A Study of the Australian Market', Journal ofAccounting and Economics, Vol. 6, pp. 133-151.

Francis, J. R. and Wilson, E. R. (1988), 'Auditor Changes: AJoint Test of Theories Relating to Agency Costs and AuditorDifferentiation', Accounting Review, October, pp. 663-682.

Fried, D. and Schiff, A. (1981), 'CPA Switches and AssociatedMarket Reactions', Accounting Review, April, pp. 324-341.

Haskins, M. E. and Williams, D. D. (1990), 'A ContingentModel of Intra-Big Eight Auditor Changes', Auditing: AJournal of Practice and Theory, Fall, pp. 55-74.

Healy, P. and Lys, T. (1986), 'Auditor Changes Following BigEight Mergers with Non-Big Eight Audit Firms', Journalof Accounting and Public Policy, Vol. 5, pp. 251-265.

Jensen, M. and Meckling, W. (1976), 'Theory of the Firm:Managerial Behaviour, Agency Costs and OwnershipStructure', Journal of Financial Economics, October, pp. 305-360.

Johnson, W. B. and Lys, T. (1990), 'The Market for AuditServices: Evidence from Voluntary Auditor Changes',Journal of Accounting and Economics, Vol. 12, pp. 281-308.

Kaplan, S. E., Menon, K. and Williams, D. D. (1990), 'TheEffect of Audit Structure on the Audit Market', Journal ofAccounting and Public Policy, Vol. 9, pp. 197-215.

Kim, J.-O. and Mueller, C. W. (1978), Introduction to FactorAnalysis: What It Is and How To Do It (Sage Publications).

Krishnan, J. (1992), 'An Empirical Model of Opinion ShoppingBehaviour' (AAA conference paper).

Levinthal, D. A. and Fichman, M. (1988), 'Dynamics of Inter-organizational Attachments: Auditor-Client Relationships',Administrative Science Quarterly, Vol. 33, pp. 345-369.

Lindahl, F. W. (1992), 'How Much Do We Know AboutAuditor Changes? The Effect of Observed and UnobservedHeterogeneity' (AAA conference paper).

Lynn, S. A. (1987), 'Segmenting the CPA Services Market',Advances in Accounting, Vol. 4, pp. 239-265.

Menon, K. and Williams, D. D. (1991), 'Auditor Credibilityand Initial Public Offerings', Accounting Review, April,pp. 313-332.

Nichols, D. R. and Smith, D. B. (1983), 'Auditor Credibilityand Auditor Changes', Journal of Accounting Research,Autumn, pp. 534-544.

Oppenheim, A. N. (1966), Questionnaire Design and AttitudeMeasurement (Heinemann).

SAS (1990), SAS/STAT User's Guide Volume 1, 4th edition(SAS Institute, Cary, NC).

Schwartz, K. B. and Menon, K. (1985), 'Auditor Switches byFailing Firms', Accounting Review, April, pp. 248-261.

Siegel, S. and Castellan, N. J. (1988), Non-parametric Statisticsfor the Behavioural Sciences (McGraw-Hill).

Simon, D. T. and Francis, J. R. (1988), 'The Effects of AuditorChange on Audit Fees: Tests of Price Cutting and PriceRecovery', Accounting Review, April, pp. 255-269.

Simunic, D. A. and Stein, M. (1986), 'On the Economics ofProduct Differentiation in Auditing', in Proceedings of the1986 Touche Ross I University of Kansas Symposium on AuditingProblems, pp. 69-99.

Smith, D. B. (1986), 'Auditor "Subject To" Opinions,Disclaimers, and Auditor Changes', Auditing: A Journal ofPractice and Theory, Fall, pp. 95-108.

Smith, D. B. (1988), 'An Investigation of Securities andExchange Commission Regulation of Auditor Change Dis-closures: The Case of Accounting Series Release No. 165',Journal of Accounting Research, Spring, pp. 134-145.

Stokes, D. (1992), 'The Market for Audit Services: Evidencefrom Audit Firm Mergers' (AAA conference paper).

Titman, S. and Trueman, B. (1986), 'Information Qualityand the Valuation of New Issues', Journal of Accounting andEconomics, Vol. S, pp. 159-172.

Wallace, W. A. (1980), 'The Economic Role of the Auditin Free and Regulated Markets', The Touche Ross and Co.Aid to Education Program, reprinted in Auditing Monographs(Macmillan Publishing Co., 1985).

Wallace, W. A. (1987), 'The Economic Role of the Auditin Free and Regulated Markets: A Review', Research inAccounting Regulation, Vol. 1, pp. 7-34.

Wallace, R. S. O. and Mellor, C. J. (1988), 'Non-response Biasin Mail Accounting Surveys: A Pedagogical Note', BritishAccounting Review, August, pp. 131-139.

Williams, D. D. (1988), 'The Potential Determinants of AuditorChange', Journal of Business Finance and Accounting,Summer, pp. 243-261.

Page 14