SOX Section 404 Material Weakness Disclosures and Audit Fees

AUDITING:A JOURNAL OF PRACTICE&THEORY

V ol.25,No.1

May2006

pp.99–114

SOX Section404Material Weakness

Disclosures and Audit Fees

K.Raghunandan and Dasaratha V.Rama

SUMMARY:Section404of the Sarbanes-Oxley Act and Auditing Standard No.2

(PCAOB2004)require management and the auditor to report on internal controls over

?nancial reporting.Section404is arguably the most controversial element of SOX,and

much of the debate around the costs of implementing section404has focused on

auditors’fees(Ernst&Young2005).In this paper,we examine the association between

audit fees and internal control disclosures made pursuant to section404.Our sample

includes660manufacturing?rms that have a December31,2004?scal year-end and

?led the section404report by May15,2005.We?nd that the mean(median)audit

fees for the?rms in our sample for?scal2004is86(128)percent higher than the

corresponding fees for?scal2003.Audit fees for?scal2004are43percent higher for

clients with a material weakness disclosure compared to clients without such disclo-

sure;however,audit fees for?scal2003are not associated with an internal control

material weakness disclosure(in the10-K?led following?scal2004).We also?nd that

the association between audit fees and the presence of a material weakness disclosure

does not vary depending on the type of material weakness(systemic or non-systemic).

INTRODUCTION

I n this paper we examine the association between material weakness disclosures made

pursuant to section404of the Sarbanes-Oxley Act(U.S.House of Representatives 2002)(SOX)and audit fees.Section404and the subsequent Public Company Account-ing Oversight Board(PCAOB)standards relating to this section(PCAOB2004)require(1) a management report on internal control over?nancial reporting,(2)auditor attestation of such a management report,and(3)an auditor’s report on internal control.The internal control reporting requirements have been the most controversial elements of SOX.Many SEC registrants and others have complained about the high costs associated with the new requirements(SEC2005b)and audit fees account for a signi?cant part of section404 implementation costs(American Bankers’Association[ABA]2005;Financial Executives Institute[FEI]2005;Ernst&Young2005).

Our analysis focuses on audit fees paid by660manufacturing?rms that have a De-cember31,2004?scal year-end and?led their404reports by May15,2005.While some submissions to the SEC and Congress have referred to surveys highlighting the magnitude of costs associated with section404,our study provides systematic empirical evidence about

K.Raghunandan and Dasaratha V.Rama are Professors at Florida International University.

We thank Joe Carcello,Bill Messier,Dan Simunic(editor),and two anonymous reviewers for many useful com-ments.We thank Ping Wang and Yun-chia Yan for research assistance.

Submitted:July2005

Accepted:December2005

99

100Raghunandan and Rama the higher audit fees following the implementation of section404.We?nd that the audit fees for2004are in general much higher than the corresponding fees for2003—the mean and median fees for2004are higher by86and128percent,respectively.Not surprisingly, we?nd that the audit fees are higher for the58?rms that disclosed material weaknesses in internal controls over?nancial reporting pursuant to section404of SOX.Of greater interest are our results related to the magnitude of the higher fees,the effects on prior-year audit fees,and the effects of types of internal control weaknesses on audit fees.

We?nd that audit fees for?scal2004are43percent higher for?rms that report a section404material weakness disclosure than for?rms without such disclosure.However, the material weakness indicator variable is not signi?cant in explaining audit fees for the same?rms for the?scal year ending December31,2003.We?nd that voluntary disclosures about internal control material weakness disclosures before section404became effective are not associated with higher audit fees for either2004or2003.Audit fees are higher irrespective of whether the internal control weaknesses can be categorized as systemic,and the impact on audit fees does not differ signi?cantly for systemic or non-systemic internal control weakness disclosures.

SOX became law in July2002and the associated SEC implementation rule was issued in early2003,so both companies and their auditors were aware of the internal-control-related reporting requirements well before2004.In addition,auditing standards have long required the auditor to consider the internal controls of the client in audit planning.Hence, the fact that the section404material weakness indicator is signi?cant in explaining audit fees for?scal2004,but not for?scal2003,is surprising.The results suggest that many auditors did not test internal controls extensively in2003but began to test extensively for internal controls(and accordingly modify their audit programs)only in2004and charged higher fees in the presence of internal control weaknesses in2004.The change in audit approach could perhaps be attributed to the fact that the PCAOB issued AS No.2only in March2004(with subsequent approval by the SEC in June2004)and additional guidance was issued by the PCAOB in June,July,October,and November2004.

Bond-rating agencies have noted that their reactions will differ depending on the type of material weakness disclosures:systemic or non-systemic(Moody’s Investor Service 2004;Fitch Ratings2005).However,we?nd that the association between audit fees and material weakness disclosures do not vary depending on the type of internal control weak-ness.These results suggest that the responses of auditors to different types of internal control weaknesses may differ from those of?nancial statement users.

In summary,our paper provides empirical evidence about audit fees for a group of manufacturing?rms that are subject to the requirements of section404of SOX.Our paper adds to both the audit fees and the internal control/section404literatures by documenting the association between audit fees and section404material weakness disclosures.

BACKGROUND AND RESEARCH QUESTIONS

SOX was enacted in July2002in the aftermath of the Enron,Andersen,and WorldCom failures.Section404(1)of SOX requires the SEC to prescribe rules requiring that each annual report?led with the SEC under the1934Securities Exchange Act shall‘‘contain an internal control report,which shall:

(1)state the responsibility of management for establishing and maintaining an adequate

internal control structure and procedures for?nancial reporting;and

Auditing:A Journal of Practice&Theory,May2006

SOX Section 404Material Weakness Disclosures and Audit Fees 101

Auditing:A Journal of Practice &Theory,May 2006

(2)contain an assessment,as of the end of the most recent ?scal year of the issuer,of

the effectiveness of the internal control structure and procedures of the issuer for

?nancial reporting.’’

In addition,section 404(2)speci?es that:

With respect to the internal control assessment required by subsection (a),each registered

public accounting ?rm that prepares or issues the audit report for the issuer shall attest to,

and report on,the assessment made by the management of the issuer ...Any such attestation

shall not be the subject of a separate engagement.

Section 404has arguably been the most controversial element of SOX.Many SEC

registrants and business associations have complained about the costs associated with the

internal control reporting requirement and some have called for the revision (if not repeal)

of such requirements (e.g.,American Electronics Association [AEA]2005;FEI 2005;Mi-

crosoft,Inc.2005;Powell 2005).Both the U.S.Senate and the U.S.House of Represen-

tatives held hearings that focused on the costs associated with the internal control reporting

requirements and other elements of SOX (U.S.Senate 2004;U.S.House of Representatives

2005).In such hearings,SEC Chairman William Donaldson acknowledged that ‘‘the inter-

nal control reporting and auditing requirements ...have required signi?cant outlays of time

and expense’’(Donaldson 2005).In light of such vociferous criticism,the SEC (1)thrice

postponed the adoption date for section 404,1(2)commissioned a round-table to discuss

implementation issues related to section 404,and (3)formed an advisory committee on

smaller companies to ‘‘examine the impact of the Sarbanes-Oxley Act and other aspects of

the federal securities laws on smaller companies’’(SEC 2005a).

Audit Fees and Section 404

Audit fees account for a substantial part of the costs associated with the implementation

of section 404(cf.,ABA 2005;FEI 2005).The CEO of Ernst &Young (EY)noted that

‘‘much of the debate regarding the high costs of 404implementation seems to be centered

on audit fees’’(EY 2005).Thus,patterns of increases in audit fees (as a result of the

implementation of section 404)are suf?ciently interesting to be examined in their own

light.

Professional organizations and others interested in section 404related issues have cited

evidence from surveys about the additional costs associated with the new rules (FEI 2005;

Charles River Associates [CRA]2005).For example,a widely cited FEI study indicates

that member ?rms spent ‘‘an average of $4.3million for added internal costs and additional

fees spent on auditors and other consultants and software associated with complying with

Section 404.’’While results from such surveys have been widely cited in submissions before

the SEC and Congress as part of efforts to change existing regulations,the nature of such

surveys preclude drawing de?nitive conclusions about the effect of section 404on audit

fees.It is not apparent how much of the increased costs are attributable to additional

audit fees,per se,as opposed to other costs.Also,Deloitte &Touche (2005)noted that

while ‘‘average per company costs have been estimated to be in the millions of dollars ...

1For registrants that are classi?ed as accelerated ?lers,section 404became effective for ?scal year-ends after November 15,2004.However,even within this category,for registrants with market ?oat of under $700million,the SEC granted an additional 45days to ?le the section 404reports if the ?scal year-end was on or before February 28,2005.For non-accelerated ?lers and foreign registrants,section 404is effective for ?scal year-ends on or after July 15,2007.

102Raghunandan and Rama

Auditing:A Journal of Practice &Theory,May 2006

the reported average cost estimates are skewed dramatically upwards by the higher costs

of the largest organizations.’’In addition,there is no empirical evidence about how audit

fees differ for ?rms with and without material weakness disclosures.2Thus,our study ?lls

a void in the literature about the effects of section 404on audit fees of ?rms with and

without internal control material weakness disclosures.

Auditing standards have long recognized the signi?cance of internal control in auditing,

since weak internal controls over ?nancial reporting result in less reliable numbers from

the accounting system.However,auditors had the option of not relying on internal controls

and to rely solely on the results from substantive tests.Thus,prior to SOX the only required

audit procedure related to internal controls was that auditors had to obtain an understanding

of clients’internal controls;tests of internal controls were necessary only if the auditor

decided to rely on the internal controls.

SOX changes the requirements related to internal controls,and requires the auditor to

evaluate management’s assertion about the effectiveness of internal controls.In addition,

AS No.2(PCAOB 2004)requires auditors to perform their own independent assessment

related to internal controls over ?nancial reporting and issue a report based on such as-

sessment.Hence,it is likely that the audit approaches would be substantially different in

the post-SOX period from the pre-SOX period.The CEO of EY acknowledged this and

noted that ‘‘the effort required by the audits of internal control under 404and AS2is

substantially greater than the effort traditionally required for consideration of internal con-

trol solely in connection with audits of ?nancial statements’’(EY 2005).

The presence of a material weakness creates signi?cant additional work for auditors

including additional (1)testing and changes in the audit program,(2)partner time related

to discussions with client management,and (3)documentation related to the decision to

classify a weakness as a material weakness as opposed to a signi?cant de?ciency (which

does not require disclosure in SEC ?lings).Such additional work can be expected to lead

to higher audit fees.

In addition,auditors and others also have noted that there would be adverse conse-

quences to issuers that disclose material weaknesses in internal controls.A senior manager

with PricewaterhouseCoopers (PWC)noted that when internal control over ?nancial re-

porting is determined to be ineffective it increases ‘‘the probability that accounting problems

exist’’(Wilfert 2005).Credit-rating agencies have suggested that certain types of internal

control weaknesses would be a factor in their rating decisions (Moody’s Investor Services

2004;Fitch Ratings 2005).Hence,the presence of a material weakness can also lead to an

increase in the auditor’s risk premium related to audit fees.

Together,the above discussion suggests the ?rst research question:

RQ1:What is the audit fee premium for ?rms disclosing a material weakness in

internal controls over ?nancial reporting compared to ?rms without such

disclosures?

SOX became law in July 2002,and both issuers and auditors were aware of the internal

control requirements of section 404.Audit ?rms have noted that ‘‘much of the preparation

for 404implementation started in 2003’’(e.g.,EY 2005).In addition,it is unlikely that the

2Charles River Associates (CRA)(2005)estimates the effect of section 404on audit fees,but is based on a small sample of 90?rms.Neither the CRA survey nor the FEI survey analyzes separately ?rms with and without material weakness disclosures.

SOX Section 404Material Weakness Disclosures and Audit Fees 103

Auditing:A Journal of Practice &Theory,May 2006

material weaknesses that were disclosed in section 404reports (in early 2005)developed

only in 2004but were not present in 2003.Hence,it is reasonable to expect that the effects

of material weaknesses in internal controls would be re?ected in the audit fees for ?scal

2003,well before the required implementation date of 2004.3However,as noted below,the

timeline of actions by regulators suggests that there may not be an association between

material weakness disclosures pursuant to section 404and audit fees for ?scal 2003.

Section 103of SOX requires the PCAOB to issue rules relating to the ‘‘auditor’s testing

of the internal control structure and procedures of the issuer.’’Pursuant to this mandate,the

PCAOB issued Auditing Standard No.2but only in March 2004,and the SEC approved

this standard in June 2004.EY (2005)notes that ‘‘only then,subsequent to AS2being

issued,did auditors have the opportunity to actively start identifying implementation issues

and seeking answers to their questions from the PCAOB staff.Additional interpretive guid-

ance was issued by the SEC and PCAOB during the latter half of 2004(in June,July,

October,and November)and in January 2005.’’

Thus,there are differing arguments about whether we are likely to ?nd an association

between the disclosures of a material weakness pursuant to section 404in ?lings made

with the SEC after ?scal 2004and audit fees for ?scal 2003.This leads to the second

research question:

RQ2:Were audit fees higher even in ?scal 2003for ?rms that subsequently

disclosed a material weakness in internal controls over ?nancial reporting?

METHOD

Model

We use the following model to examine the association between audit fees and the type

of internal control report:

LAFEE ?b 0?b 1*LNTA ?b 2*RECINV ?b 3*FORGN ?b 4*SQSEG ?b 5*LIQ

?b 6*DA ?b 7*ROA ?b 8*GC ?b 9*BIG4?b 10*MW .

The variables are de?ned as follows:

LAFEE

?natural log of audit fees;LNTA

?natural log of total assets at year-end;RECINV

?receivables and inventory as a proportion of total assets at year-end;FORGN

?1if the ?rm has foreign operations,else 0;SQSEG

?square root of the number of operating segments;LIQ

?current ratio at year-end;DA

?total debt divided by total assets at year-end;ROA

?return-on-assets (operating income/total assets);3Neither AS No.2(PCAOB 2004)nor the SEC rules related to section 404(SEC 2003)require the disclosure of when the internal control problems developed;these rules,as well as SOX itself,only require the disclosure based on internal control evaluations as of the ?scal year-end.Disclosure practices vary widely across ?rms.While some clients indicate that the internal control problems had existed for some time,others do not make any statement about when the internal control problems https://www.360docs.net/doc/f218376609.html,pounding the dif?culty is the fact that some of the ?rms in the latter group disclose restatements relating to prior years.Hence,we cannot infer when the problems developed if the disclosures are silent about the date of origin of the internal control problem.

104Raghunandan and Rama

Auditing:A Journal of Practice &Theory,May 2006

GC ?1if audit report modi?ed for going-concern,else 0;

BIG4?1if Big 4auditor,else 0;and

MW ?1if there is a material weakness disclosure in audit report,else 0.

Our control variables are derived based on prior research.Almost all models that seek

to explain audit fees follow the approach in Simunic (1980)and use various measures of

client size,complexity,and risk such as LNTA ,RECINV ,FORGN ,SQSEG ,and GC .We

include DA ,LIQ ,and ROA in our model given the signi?cance of these or similar variables

in recent research (e.g.,Francis and Wang 2005;Whisenant et al.2003).

Sample

We begin with a list of 2,956?rms that ?led their section 404reports by May 15,

2005,per the AuditAnalytics database.We restrict our analysis to manufacturing ?rms with

a December 31?scal year-end for the following reasons.First,prior research suggests that

audit fees vary across industries (e.g.,Simunic 1980;Casterella et al.2004).Second,it is

likely that both the presence of material weaknesses in internal control and the effects of

such weaknesses on audit effort and fees would vary across industries.4Hence,we delete

2,019?rms that are not in the manufacturing (i.e.,SIC codes 2000–3999)sector.

Section 404and AS No.2(PCAOB 2004)both have led to major changes in auditing.

Detailed guidance from the PCAOB was available only in the second half of 2004,and

given the time pressures on auditors and clients,we delete 147?rms that have a ?scal year-

end other than December 31,2004.Since section 404is not mandatory for foreign ?rms

until July 15,2007,we delete 24foreign ?rms from our sample.We rely on the Audit-

Analytics database for audit fee data,and delete 35?rms with missing fee data.Given the

unique nature of section 404work,and the steep learning curve associated with a new

client in the initial year of a 404audit,we delete 67?rms where either 2003or 2004is

an initial year audit for the auditor.Finally,we delete four ?rms that have very small values

for audit fees or client size.5Thus,our ?nal sample includes 660?rms.We obtain all of

our data from the Compustat and AuditAnalytics databases.

RESULTS

Table 1provides descriptive data about the sample for ?scal years ending December

31,2004and 2003.The mean total assets for 2004and 2003are $6,537and $5,832million,

respectively;the median values are much lower in both years,and provide evidence about

the skewness in the sample.The values for all of the control variables suggest that the 2004

and 2003values are quite similar;only LIQ has a change of more than 5percent in the

mean value across the years (changing from 4.18in 2003to 3.91in 2004).Receivables

and inventory constitute just less than a quarter of total assets and just under two-thirds of

the sample has foreign operations.Ninety-?ve percent of the sample is audited by the Big

4,and 2percent of the sample ?rms receive going-concern modi?ed audit opinions.Fifty-

eight of the ?rms (8.8percent)report one or more section 404material weaknesses in

internal control.The table also shows that the mean audit fees for 2004is $3.08million,

and represents an 86percent increase from the mean fees of $1.66million for 2003;the

4

Prior research (e.g.,Asare and Davidson 1995;Kreutzfeld and Wallace 1986)?nds that internal control problems are much more likely in inventory and receivables than in other asset accounts;further,auditing standards have long required auditors to use speci?c audit procedures in these two areas.The proportion of receivables and inventory (as a percent of total assets)varies across industries,so the impact of additional testing related to section 404can be expected to have differential impact across industries.5Two ?rms had total assets less than $10million and two other ?rms had audit fees less than $100,000.

SOX Section 404Material Weakness Disclosures and Audit Fees 105

Auditing:A Journal of Practice &Theory,May 2006TABLE 1

Sample Statistics

(n ?660)

Variable Fiscal 2004

Mean S.D.

Median Fiscal 2003Mean S.D.Median Total Assets ($MM)6,35736,689

5425,83234,953509Audit Fees ($000)3,0825,881

1,2121,6603,812532LNTA 20.29 1.91

20.1420.15 1.9220.05RECINV 0.240.16

0.230.230.150.23SQSEG 1.510.55

1.41 1.440.53 1.00FORGN 0.650.48

1.000.640.48 1.00LIQ 3.91 3.93

2.69 4.18 4.46 2.73DA 0.450.29

0.420.470.300.45ROA ?0.040.26

0.04?0.050.230.03GC 0.020.15

0.000.020.150.00BIG40.950.22

1.000.950.22 1.00MW 0.090.28

0.00———MWSYS 0.060.24

0.00———MWOTH 0.030.180.00———

The sample includes all manufacturing ?rms that ?led their section 404reports with the SEC by May 15,2005.

Variable De?nitions:Total Assets ?total assets;LNTA ?natural log of client’s total assets;RECINV ?proportion of total assets in accounts receivable and inventory;SQSEG ?square root of the number of segments;FORGN ?1if the ?rm has foreign operations,else 0;INITIAL ?1if ?rst year with auditor,else 0;LIQ ?ratio of current assets divided by current liabilities;DA ?total debt divided by total assets;ROA ?return on assets (operating income divided by total assets);GC ?1if audit opinion modi?ed for going concern,else 0;BIG4?1if Big 4auditor,else 0;MW ?1if material weakness in internal control,else 0;MWSYS ?1if ?rm disclosed systemic material weakness in internal control,else 0;and MWOTH ?1if ?rm disclosed non-systemic material weakness in internal control,else 0.

median audit fee increases by 128percent,from $0.53million in 2003to $1.21million in

2004.

Table 2presents the correlation matrix for the independent variables.The highest cor-

relation is only .48,suggesting that multicollinearity is not likely to be a problem.This is

con?rmed later by VIF scores from our regression,none of which exceeds 1.6.6

Table 3presents details about the mean and median audit fees for 2003and 2004,when

we partition the sample by size quartiles (based on 2004total assets).A comparison of the

?rst two rows of the table show that in every quartile there is a signi?cant increase in

average audit fees from 2003to 2004,ranging from an increase of 136percent in the

second quartile to an increase of 79percent in the fourth quartile.Parametric and nonpar-

ametric tests indicate that the fee increases in the overall sample,as well as in each of the

6VIF scores less than 10indicate that inferences are not confounded by multicollinearity problems (Kennedy 2003).

106Raghunandan

and

Rama

Auditing:

A

Journal of Practice &Theory,May 2006TABLE 2

Pearson Correlations for 2004(2003)

Variable RECINV FORGN SQSEG LIQ DA ROA GC BIG4MW MWSYS MWOTH

LNTA 0.11**(0.14**)0.31**(0.35**)0.47**(0.46**)?0.40**(?0.40**)0.40**(0.48**)0.42**(0.40**)?0.05(?0.02)0.20**(0.20**)?0.01(?0.00)0.01(0.02)?0.02(?0.03)RECINV 0.23**(0.26**)0.25**(0.20**)?0.34**(?0.42**)0.06(0.15**)0.37**(0.45**)?0.08*(0.00)?0.05(?0.07)0.12**(0.13**)0.07(0.06)0.10*(0.13**)

FORGN 0.24**(0.28**)?0.24**(?0.28**)0.07(0.13**)0.25**(0.22**)?0.14**(?0.09*)0.11**(0.11**)0.07(0.10*)0.03(0.07)0.06(0.06)

SQSEG ?0.38**(?0.32**)0.31**(0.30**)0.29**(0.22**)?0.01(?0.09*)0.05(0.07)0.06(0.06)0.05(0.05)0.02(0.04)

LIQ ?0.44**(?0.45**)?0.19**(?0.26**)?0.04(?0.06)?0.07(?0.06)?0.05(?0.05)?0.04(?0.04)?0.03(?0.03)DA ?0.05(0.09*)0.24**(0.21**)0.12**(0.08*)0.05(0.05)0.05(0.07)0.01(?0.01)ROA ?0.18**(?0.04)0.04(0.04)?0.01(0.01)0.03(0.03)?0.05(0.02)GC 0.03(?0.01)?0.05(0.02)?0.04(0.04)?0.03(?0.03)BIG4?0.03(?0.03)0.03(0.02)?0.08*(?0.08*)**,*Indicate p ?.01and p ?.05,respectively.The variables are de?ned as in Table 1.

SOX Section 404Material Weakness Disclosures and Audit Fees 107

Auditing:A Journal of Practice &Theory,May 2006TABLE 3

2004Audit Fees Compared with 2003Audit Fees and Predicted Fees

Full Sample

Client Size Quartile 1234Mean (Median)FY 2003Fees ($)

1,659,692(532,000)243,346(195,100)384,272(317,830)1,030,373(877,500)4,651,869(2,959,000)Mean (Median)FY 2004Fees ($)

3,082,660(1,212,280)515,414(416,950)906,360(758,000)2,117,903(1,891,330)8,329,858(5,877,500)Mean (Median)of Ratio of 2004Fees to 2003Fees

2.35(2.08) 2.41(2.11) 2.58(2.31) 2.32(2.13) 2.09(1.88)Mean (Median)of Ratio of Actual to Predicted Fees (using 2003fee model) 2.25(2.02) 2.54(2.42) 2.01(1.82) 2.16(1.97) 2.21(2.02)four size-based quartiles,are signi?cant (p ?.01in each instance).The third row of the table shows that in each size quartile the means of the ratios of 2004fees to 2003fees are more than 2.0.The fourth row of the table indicates the means and medians of the ratio of actual 2004fees to the predicted fees for 2004using the 2003audit fee model (discussed in the next paragraph);the results show that for each quartile the ratio is more than 1.8.Finally,analyses of the fees for the two years indicate that audit fees for ?scal 2004are higher than the audit fees for ?scal 2003for 641of the 660?rms in our sample.Overall,the data suggest that the fee increases are not being driven by ?rms of any particular size.

Table 4presents the results from two regressions—one each for ?scal 2004and 2003fees.Both regressions are highly signi?cant,and the explanatory powers of both regressions are fairly high and quite similar (the adjusted R 2s are .82and .81for the 2004and 2003regressions,respectively).The variable of interest,MW ,is positive and signi?cant in the 2004regression but not signi?cant in the 2003regression,indicating that the presence of a section 404material weakness disclosure is associated with higher audit fees in 2004but not in 2003.The coef?cient for MW in the regression is 0.360,so the effect of having a material weakness in internal control corresponds to a 43percent increase in audit fees.7

The results show the signi?cance and magnitude of ?ve of the coef?cients (LNTA ,FORGN ,LIQ ,ROA ,BIG4)are quite similar across the two periods;t-tests indicate that we cannot reject the null hypothesis of no difference in the magnitude of the coef?cients of these ?ve variables in the two regressions.However,the pattern of results with respect to the other coef?cients is quite interesting.

While RECINV is positive and signi?cant in both regressions,a t-test indicates that the magnitude of the coef?cient is higher (p ?.10)in 2004than in 2003;as noted earlier,receivables and inventory are two areas that are most prone to errors and misstatements.It is reasonable to expect that receivables and inventory will have more complex internal control procedures,thus leading to greater section 404speci?c audit fees.SQSEG is positive and signi?cant in the 2004regression,but not in the 2003regression.Firms with more business segments require more/more complex internal controls.Intuitively,the greater the number of different business segments,the more work the auditor must perform in assessing internal controls resulting in higher fees.

7The dependent variable is natural log of audit fees,so the impact of a change in MW from 0to 1is given by e 0.36(?1.43).

108

Raghunandan and Rama

Auditing:A Journal of Practice &Theory,May 2006TABLE 4

Regression Results Model:LAFEE ?b 0?b 1*LNTA ?b 2*RECINV ?b 3*FORGN ?b 4*SQSEG ?b 5*LIQ

?b 6*DA ?b 7*ROA ?b 8*GC ?b 9*BIG4?b 10*MW

Variable

20042003Intercept

2.989(.00) 1.836(.00)LNTA

0.502(.00)0.539(.00)RECINV

0.598(.00)0.411(.02)FORGN

0.206(.00)0.183(.00)SQSEG

0.168(.00)0.020(.67)LIQ

?0.017(.01)?0.019(.00)DA

0.054(.53)0.191(.04)ROA

?0.440(.00)?0.440(.00)GC

0.301(.03)0.071(.63)BIG4

0.444(.00)0.416(.00)MW 0.360(.00)0.088(.27)F ?293.6;p ?.001Adj.R 2?.82F ?277.0;p ?.001Adj.R 2?.81

p-values are in https://www.360docs.net/doc/f218376609.html,FEE ?natural log of audit fees.Other variables are as de?ned in Table 1.

The GC variable is signi?cant in 2004but not in 2003.Firms in ?nancial distress may be more likely to cut back on control-related activities that may be viewed as discretionary spending items (Carcello et al.2005);hence,auditors may need to perform additional work in ?rms with a going-concern uncertainty.8However,DA is signi?cant in 2003but not in 2004;this suggests that cost drivers of audit fees could have shifted post-SOX-404.That is,leverage may affect auditors’work less than in previous years,consistent with auditors performing a more exhaustive audit in the post-SOX-404period as opposed to an audit tailored to business speci?c risks.

Overall,the results on the control variables reveal changes in the audit market and in the way auditors perform their work.Apart from the presence of a material weakness itself,factors that may be associated with the presence of a material weakness also lead to higher 8An alternative explanation is that perhaps auditors require larger risk premiums now for distressed ?rms.How-ever,it is unlikely that such an increase is caused essentially by section 404.

SOX Section 404Material Weakness Disclosures and Audit Fees 109Auditing:A Journal of Practice &Theory,May 2006

audit fees.9However,additional research work is required before de?nitive conclusions can be drawn about the impact of section 404on the audit process.The results do suggest possible opportunities for future research (e.g.,evaluate the impact of section 404in other industries as well as differences across industries,longer time periods,etc.).

Additional Analyses:Types of Material Weaknesses

Bond-rating agencies have noted that their responses to section 404disclosures will vary depending on the type of material weakness.For example,Fitch Ratings (2005)clas-si?es some internal control weakness disclosures as pervasive/systemic control weaknesses and notes that:

Certain material weaknesses might constitute pervasive risks,such as problems with ‘‘tone at the top’’or the quality of personnel in charge of the ?nancial reporting functions...Ratings for companies not previously identi?ed as having such pervasive weaknesses by Fitch will need to be looked at carefully,and negative action is likely ...Material weaknesses in internal controls can also occur at the transaction level,potentially affecting information such as speci?c account balances.However,auditors usually can perform tests fairly easily to ensure that account balances are fairly stated and disclosures are adequate.

Similarly,Moody’s Investors Service (2004)states that material weakness disclosures will be classi?ed as Type A or Type B—analogous to the account/transaction or systemic weakness categories of Fitch—and that their impact on bond ratings will be different,with adverse consequences more likely for Type B internal control weaknesses.

As seen in Table 5,37of the 58?rms with section 404material weakness disclosures disclose problems that are systemic in nature.For example,27of the ?rms note problems with the quality and/or training of accounting personnel,while 21?rms disclose prob-lems with the reconciliation of accounts and ?nancial statement preparation.

We examine the association between audit fees and the type of material weakness by replacing the MW variable with two indicator variables:MWSYS and MWOTH .MWSYS takes a value of 1if the issuer reported a systemic problem,and 0otherwise,while MWOTH takes the value of 1if there is a material weakness that is not systemic,and 0otherwise.10We include MWSYS and MWOTH in the regression (but drop the MW variable).Panel B of Table 5presents the results from such regressions.11Both MWSYS and MWOTH are signi?cant for ?scal 2004fees but not for ?scal 2003fees.Tests indicate that the differences in magnitude between the two coef?cients are not statistically signi?cant in either regres-sion.Thus,the results indicate that audit fees are higher for issuers that disclose a section 404internal control weakness disclosure,irrespective of the type of material weakness disclosure.

Additional Analyses:Voluntary Disclosures in 2004

Section 302of SOX deals with another type of disclosures related to controls within an entity:disclosure controls.There is signi?cant overlap between disclosure controls and internal controls over ?nancial reporting (SEC 2003).However,there are many differences 9

t-tests indicate that the coef?cients of SQSEG ,DA ,and GC are signi?cantly different (p ?.05)in the 2004and 2003regressions;the change in the coef?cient of RECINV is signi?cant at p ?.10.10

Thus,by construction,every instance of a systemic problem implies the presence of problems related to trans-actions—but not vice-versa.11For the sake of brevity,we do not report the coef?cients of the other variables in the table since all of the coef?cients are quite similar to those reported earlier in Table 4.

110

Raghunandan and Rama Auditing:A Journal of Practice &Theory,May 2006TABLE 5

Types of Material Weaknesses and Audit Fees

Panel A:Types of Section 404Internal Control Weaknesses Disclosed

Problems in speci?c types of transactions or accounts only:

21Systemic or Broad Weaknesses:Quality and training of accounting personnel 27Segregation of duties 9Reconciliation of accounts,?nancial statement preparation 21Information Systems related problems 7Quality of internal audit and/or audit committee 2Total ?rms with systemic or broad weaknesses

37a

Panel B:Association between Types of Weaknesses and Audit Fees Model:LAFEE ?b 0?b 1*LNTA ?b 2*RECINV ?b 3*FORGN ?b 4*SQSEG ?b 5*LIQ

?b 6*DA ?b 7*ROA ?b 8*GC ?b 9*BIG4?b 10*MWSYS

?b 11*MWOTH

Variable

20042003Other control variables

Omitted from table Omitted from table MWSYS

.389(.00).113(.22)MWOTH .344(.00)

.117(.35)F ?267.2;p ?.001Adj.R 2?.82F ?251.9;p ?.001Adj.R 2?.81

p-values are in parentheses.

a Many ?rms had more than one of the above problems,so the numbers do not add to 37.

Variables De?nitions:MWSYS ?1if ?rm disclosed systemic material weakness in internal control,else 0;and MWOTH ?1if ?rm disclosed non-systemic material weakness in internal control,else 0.

Other variables are as de?ned as in Tables 1and 3.For the sake of brevity,we do not report the coef?cients for the control variables.The coef?cients for all of the control variables in the two regressions are substantively similar to (magnitude within 5percent of,and not signi?cantly different from)the coef?cients reported in Table 2.

between disclosures under sections 302and 404.One of the differences is that auditor association with section 302disclosures is only indirect,in contrast to the direct assessment and reporting requirements under section 404.Section 302of SOX became effective on August 29,2002,while section 404of SOX became effective for ?scal year-ends after November 15,2004.An industry newsletter (ComplianceWeek )notes that some ?rms vol-untarily disclosed problems in internal control over ?nancial reporting in 2004while making section 302disclosures,before the implementation date for section 404.It is possible that audit fees are also higher for ?rms that voluntarily disclosed internal control problems prior to the required implementation date for section 404.

We obtained the list of voluntary internal control weakness disclosures made by ?rms in 2004prior to the required 404disclosures from ComplianceWeek and checked if any of the 660?rms in our sample had made prior section 302disclosures about internal control

SOX Section 404Material Weakness Disclosures and Audit Fees 111Auditing:A Journal of Practice &Theory,May 2006

weaknesses in SEC ?lings.12Nine ?rms in our sample voluntarily disclosed a material weakness in internal control during 2004,but only seven of these ?rms later also disclosed a weakness under section 404;an additional 21?rms in our sample voluntarily disclosed internal control weaknesses earlier in 2004without using the term ‘‘material weakness’’and six of these ?rms subsequently reported a material weakness under section 404.13We included a dummy variable in our regression model to capture the effects of voluntary reporting of internal control weaknesses prior to the implementation date of section 404.VOLMW equals 1if the ?rm voluntarily disclosed a weakness in internal controls in 2004,and 0otherwise.Since we want to distinguish ?rms that voluntarily disclosed a material weakness from those that only disclosed under section 404,we include both MW and VOLMW in the regression.VOLMW is not signi?cant in the fee regressions for either 2004or 2003,and the signi?cance of the other variables remains substantively similar to those reported in Table 4.

Additional Analysis:Nonaudit Fees

Given the ready availability of nonaudit fees,and the recent surge in interest related to various nonaudit-fee-related issues,we also examined the nonaudit fees and the nonaudit fee ratio (nonaudit fees as a proportion of audit fees)for the sample ?rms in 2003and in 2004.We ?nd that the nonaudit fee ratio does not differ,either in 2003or in 2004,for ?rms with and without a material weakness disclosure.14

Sensitivity Tests

We perform the following size-related tests.First,we partition the sample into two groups based on total assets (below or above median)and perform separate regressions on the two samples.Our results related to the variable of interest,MW ,hold in both samples.Second,we delete outlier observations that were more than two standard deviations away from the mean.Third,for the continuous variables we use the values at the 1st and 99th percentiles to winsorize the sample.In each instance,our results remain qualitatively similar to those presented in the paper.

Our sample includes 32?rms audited by non-Big 4?rms.We delete these 32?rms,and thus restrict the sample to clients of the Big 4?rms.The results with the subsample of Big 4clients are similar to those reported in the paper.In addition,to check if clients of any particular Big 4?rm were driving the results we add three dummy variables in the regression;the coef?cient for PricewaterhouseCoopers is positive and signi?cant indicating the presence of a PWC fee premium,but the inferences related to the other variables remain unchanged.

SUMMARY AND CONCLUSIONS

Section 404of the Sarbanes-Oxley Act has arguably been the most controversial ele-ment of SOX.There have been numerous complaints about the high costs associated with that part of SOX,and ‘‘much of the debate regarding the high costs of 404implementation 12

Thus,we assume that none of the other ?rms in our sample had disclosed an internal control weakness under section 302in quarterly ?lings with the SEC.As a test,we examined a random sample of 20?rms that did not appear in the ComplianceWeek list for each quarter in 2004and found that none of them had made any internal control weakness disclosures.13

Some of these ?rms used terms such as ‘‘signi?cant de?ciency’’(used in AS No.2)or ‘‘reportable condition’’(used in prior AICPA standards.)14However,the average nonaudit fee ratio (nonaudit fees/audit fees)declines from 0.65in 2003to 0.31in 2004.While the average audit fee increases from $1.66million in 2003to $3.08million in 2004,the average nonaudit fee declines from $1.26million in 2003to $1.12million in 2004.

112Raghunandan and Rama seems to be centered on audit fees’’(Ernst&Young2005).In this paper,we examine the association between section404disclosures and audit fees.Our sample includes660man-ufacturing?rms with a December31,2004?scal year-end that had?led their section404 reports by May15,2005.

We?nd that the mean(median)audit fees for?scal2004is86(128)percent higher than the corresponding audit fees for?scal2003;while the fact that the fees are higher is itself not surprising,the magnitude of the increase in fees gives some empirical evidence about the changed audit environment in2004compared to2003.

Results from our regression model indicate that the audit fee for?scal2004for?rms disclosing a material weakness in internal control over?nancial reporting is43percent higher than for?rms without such a disclosure.However,we?nd that the audit fees for ?scal2003are not associated with the presence of a material weakness disclosure(in the 10-K?led following?scal2004).Auditing standards have long required auditors to evaluate the client’s internal controls(unless the auditor decided not to rely on internal controls), so the result related to2003fees is initially surprising.Unless the material weaknesses developed only in2004,the evidence suggests that auditors(1)were not testing internal controls extensively in2003and therefore did not charge higher fees in the presence of internal control weaknesses in2003,but(2)began to test extensively for internal controls (and accordingly modify their audit programs)only in2004and charged higher fees in the presence of internal control weaknesses in2004.An alternative,but not mutually exclusive, explanation is that some companies and their auditors may not have started their section 404work until2004,but once the auditor and/or client identi?ed a material weakness the audit fees went up drastically.Of course,detailed guidance from the PCAOB relating to section404audits was available only in March2004,with additional guidance provided throughout the year.Hence,it is reasonable that audit programs were adjusted during2004, resulting in higher fees for?scal2004,particularly in the presence of a material weakness in internal control over?nancial reporting.

Some?rms have voluntarily disclosed weaknesses in internal control in other?lings with the SEC during2004,before the required implementation date of section404.We ?nd that such voluntary disclosure is not associated with higher audit fees in either?scal 2004or?scal2003;however,our inferences must be viewed with caution given the small sample of?rms with such voluntary disclosures.

Bond-rating agencies have stated that user perceptions about internal control weakness disclosures will vary depending on the type of internal control weakness,with disclosures about systemic or pervasive weaknesses leading to more adverse consequences than dis-closures of problems that are account-or transaction-speci?c.However,we?nd that the association between audit fees and internal control disclosures does not vary depending on the type of material weakness.While this may appear to contradict the views expressed by bond-rating agencies about the differences in risks associated with the two types of disclo-sures,it could simply re?ect the fact that auditors may increase the level of substantive testing in the presence of either type of internal control problem—at least in the initial year of identifying a material weakness.

The results also suggest many interesting avenues for future research.How do users react to the disclosure of material weaknesses?For example,is the cost of capital higher Auditing:A Journal of Practice&Theory,May2006

SOX Section 404Material Weakness Disclosures and Audit Fees 113Auditing:A Journal of Practice &Theory,May 2006

for ?rms disclosing a material weakness in internal control?15Another issue for future research is to examine if auditors’reactions vary depending on the type of internal control weakness.For example,are auditors more likely to resign from engagements in the presence of systemic or pervasive internal control problems?Auditors may have been reluctant to resign from engagements in 2004after a material weakness determination,and thus add to the clients’problems in a year of signi?cant change,but may be more likely to resign once the initial-year work associated with section 404is over.Another interesting issue is to examine the pattern of audit fees for clients disclosing different types of internal control weaknesses.The Big 4have noted in submissions to the SEC that they expect audit fees to be lower in future years,given the learning-curve effect related to section 404work (e.g.,PricewaterhouseCoopers 2005;EY 2005).However,the pattern of reductions in audit fees may vary depending on (1)the existence of a material weakness,and (2)the type of material weakness.These are empirical questions for future research.

REFERENCES

American Bankers’Association (ABA)https://www.360docs.net/doc/f218376609.html,ments submitted to the SEC Roundtable on Im-

plementation of Internal Control Reporting Provisions.Washington,D.C.:SEC.Available at:https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp/soxcomp-brunner.pdf.

American Electronics Association (AEA).2005.Sarbanes-Oxley Section 404:The ‘‘Section’’of Un-

intended Consequences and Its Impact on Small Business .Washington,D.C.:AEA.

Asare,S.,and R.Davidson.1995.Expectation of errors in unaudited book values:The effect of

control procedures and ?nancial condition.Auditing:A Journal of Practice &Theory 14(Spring):1–18.

Carcello,J.,D.Hermanson,and K.Raghunandan.2005.Factors associated with U.S.public com-

panies’investment in internal auditing.Accounting Horizons (June):65–80.

Casterella,J.R.,J.R.Francis,B.L.Lewis,and P.L.Walker.2004.Auditor industry specialization,

client bargaining power,and audit pricing.Auditing:A Journal of Practice &Theory (Spring):123–140.

Charles River Associates (CRA).2005.Sarbanes-Oxley Section 404Costs and Remediation of De?-

ciencies.Washington, D.C.:Charles River Associates.Available at:https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp/soxcomp-all-attach.pdf.

Deloitte &Touche.2005.Feedback on experiences with the implementation of the auditing and

reporting requirements of Section 404of the Sarbanes-Oxley Act of 2002.Available at https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp/soxcomp-koch.pdf.

Donadson,W.2005.Testimony:Concerning the Impact of the Sarbanes-Oxley Act .U.S.House of

Representatives,Committee on Financial Services.Washington,D.C.:Government Printing Of?ce.

Ernst &https://www.360docs.net/doc/f218376609.html,ments submitted to the SEC’s Roundtable on Implementation of Internal

Control Reporting Provisions.Implementation of Sarbanes-Oxley Internal Controls Provisions .Available at:https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp/soxcomp-turley.pdf.

Financial Executives Institute (FEI)https://www.360docs.net/doc/f218376609.html,ments of the Committee on Corporate Reporting.

Submitted to the SEC Roundtable on Implementation of Internal Control Reporting Provisions.15If auditors provide an unquali?ed opinion in their report on the ?nancial statements,regardless of the presence of a material weakness,there may be no reason for users to assess the quality of the ?nancial information any differently and thus react differently.On the other hand,if users discount the value of a ?rm more severely following the disclosure of systematic material weaknesses,then this could be driven by factors other than ?nancial information quality.One implication of the above argument is that research examining differences in user reactions to the presence of material weaknesses (or different types of weakness disclosures)may need to control for other factors,such as those related to business risk.

114Raghunandan and Rama Washington,D.C.:SEC.Available at:https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp/soxcomp-brod.

pdf.

Fitch Ratings.2005.Fitch’s Approach to Evaluating Management and Auditor Assessments of Internal Control.New York,NY:Fitch.

Francis,J.R.,and D.Wang.2005.Impact of the SEC’s public fee disclosure requirement on subse-quent period fees and implications for market ef?ciency.Auditing:A Journal of Practice& Theory(Supplement):145–160.

Kennedy,P.2003.A Guide to Econometrics.5th Edition.Malden,MA:Blackwell Publishing. Kreutzfeld,R.,and W.Wallace.1986.Error characteristics of audit populations:Their pro?le and relationships to environmental factors.Auditing:A Journal of Practice&Theory6(Fall):20–

43.

Microsoft,https://www.360docs.net/doc/f218376609.html,ments submitted to the SEC Roundtable on Implementation of Internal Control Reporting Provisions.Washington,D.C.:SEC.Available at:https://www.360docs.net/doc/f218376609.html,/news/ press/4-497/microsoft033105.pdf.

Moody’s Investors Service.2004.Special Comment:Section404Reports on Internal Control.New York,NY:Moody’s Investors Service.

Powell,S.S.2005.Costs of Sarbanes-Oxley are out of control.Letters to the Editor.Wall Street Journal(March21):A-17.

https://www.360docs.net/doc/f218376609.html,ments submitted to the SEC’s Roundtable on Implementation of Internal Control Reporting Provisions.Available at:https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp/ soxcomp-stauffer.pdf.

Public Company Accounting Oversight Board(PCAOB).2004.An Audit of Internal Control Over Financial Reporting Performed in Conjunction with An Audit of Financial Statements.Auditing Standard No.2.Washington,D.C.:PCAOB.

Securities and Exchange Commission(SEC).2003.Final Rule:Management’s Reports on Internal Control Over Financial Reporting and Certi?cation of Disclosure in Exchange Act Periodic Reports.Release Nos.33-8238,34-47986.Washington,D.C.:SEC.

———.2005a.Press Release:SEC Chairman Donaldson Announces Members of Advisory Committee on Smaller Public Companies.Washington,D.C.:SEC.

———.2005b.Spotlight On:Implementation of Internal Control Reporting Provisions.Washington,

D.C.:SEC.Available at:https://www.360docs.net/doc/f218376609.html,/spotlight/soxcomp.htm.

Simunic,D.1980.The pricing of audit services:Theory and evidence.Journal of Accounting Research 18(1):161–190.

U.S.House of Representatives.2002.The Sarbanes-Oxley Act of2002.Public Law107-204[H.R.

3763].Washington,D.C.:Government Printing Of?ce.

———.2005.The Impact of the Sarbanes-Oxley Act.Hearing before the Committee on Financial Services.April21.Washington,D.C.:Government Printing Of?ce.

U.S.Senate.2004.Examining the Impact of the Sarbanes-Oxley Act.Hearings before the Senate Banking Committee.September9.Washington,D.C.:Government Printing Of?ce. Whisenant,S.,S.Sankaraguruswamy,and K.Raghunandan.2003.Joint determination of audit and non-audit fees.Journal of Accounting Research(September):721–744.

Wilfert,G.E.2005.The?rst wave of section404reporting.Orange County Business Journal28

(6):4.

Auditing:A Journal of Practice&Theory,May2006