tamara a. lambert

8/8/2019 Tamara a. Lambert

1/44

Unintended Consequences of Accelerated Filings: Are Mandatory Reductions in

Audit Delay Associated with Reductions in Earnings Quality?

Tamara A. LambertUniversity of Massachusetts Amherst

Isenberg School of ManagementDepartment of Accounting and Information Systems

121 Presidents DriveAmherst, MA 01003

[email protected]

Keith L. JonesGeorge Mason UniversityDepartment of AccountingEnterprise Hall, MSN 5F4Fairfax, VA 22030-4444

[email protected]

Joseph F. BrazelNorth Carolina State University

Department of AccountingCollege of Management

Campus Box 8113Nelson Hall

Raleigh, NC 27695919-513-1772

[email protected]

July 2010

We would like to thank Philip Berger, Scott Bronson, Randy Elder, Pieter Elgers, Barbara Grein,Linda Kolbasovsky, Kevin Melendrez, Roger Meuwissen, Gary Peters, Ray Pfeiffer, andworkshop participants at the University of Illinois, University of Massachusetts, Virginia TechUniversity, North Carolina State University, Northern Illinois University, University ofWisconsin-Milwaukee, Texas Tech University, Drexel University, the 2007 InternationalSymposium on Audit Research, the 2007 AAA Annual Meeting, and the 2008 Auditing Mid-Year Conference for their helpful comments.


2/44

Unintended Consequences of Accelerated Filings: Are Mandatory Reductions in

Audit Delay Associated with Reductions in Earnings Quality?

ABSTRACT: SEC rules 33-8128 and 33-8644 substantially reduce the 10-K filing period forlarge accelerated filers and accelerated filers from 90 days to 60 and 75 days, respectively. Large

accelerated filers are firms with a market value of equity greater than $700M and accelerated filershave a market value of equity between $75M and $700M. The SEC has twice reduced filingperiods by 15 days. For many firms and their auditors, these rules have led to mandatoryreductions in audit delay (i.e., the length of time from a companys fiscal year-end to the date ofthe auditors report). We investigate the potential effects of this regulation by examining underwhat contexts these mandatory reductions have been associated with changes in earnings quality.We use discretionary accruals and other measures to proxy for earnings quality. We find that largermandatory reductions in audit delay ( 15 days) more negatively impact earnings quality thansmaller mandatory reductions. This result suggests an unintended consequence of the SECs twoseparate 15-day reductions in filing deadlines (i.e., lower earnings quality). The relation betweenaudit delay reductions and earnings quality also appears to be more acute for both busy-season

audits (vs. non-busy season audits) and accelerated filers (vs. large accelerated filers). Overall, ourfindings support claims by auditors and preparers that accelerated filings would lead to reductionsin the quality of financial information supplied to external users.

Keywords: accelerated SEC filings, audit delay, discretionary accruals, earnings quality

Data availability: The data used in this study are publicly available from the sources indicated inthe text.


3/44

1

1. Introduction

SEC rules 33-8128 and 33-8644 substantially reduce the 10-K filing period for large accelerated filers

and accelerated filers from 90 days to 60 and 75 days, respectively (SEC, 2002; 2005).1For many firms and

their auditors, meeting these deadlines has led to a mandatory reduction in audit delay. Audit delay is defined as

the length of time from a companys fiscal year-end to the date of the auditors report (Ashton et al., 1987).

Concurrently, the audit failures of the early 2000s led to the passage of the Sarbanes-Oxley Act of 2002 (SOX)

and to the issuance of new auditing standards which are intended to improve audit quality by substantially

increasing the scope of the external audit (e.g., PCAOB, 2004; PCAOB, 2007). This paper explores under what

contexts requiring auditors to do more work in less time might hinder audit quality. If SEC-mandated reductions

in audit delay have led to lower audit quality, this effect should manifest itself in the dissemination of lower

quality earnings to financial statement users.

The SEC has implemented accelerated reporting deadlines largely over the objections of auditors and

preparers. Many suggest that the SEC did not give sufficient consideration to the negative consequences.2

Auditors facing shorter audit windows at year-end are forced to work less hours after year-end and/or perform a

larger percentage of their testing prior to year-end. Caramanis and Lennox (2008) find that lower audit hours

increase the extent to which managers are able to report aggressively higher earnings. Similarly, audit standards

suggest that reducing the amount of testing performed after year-end increases the risk that auditors will fail to

detect a material misstatement. Audit procedures performed after year-end (vs. at interim or before year-end)

typically provide more reliable evidence and are generally more effective (AICPA, 1983). For example,

1

According to rule 33-8128 (SEC, 2002), an accelerated filer (AF) is a firm that meets the following conditions at the end of its fiscalyear: 1) Its common equity public float was $75M or more as of the last business day of its most recently completed second fiscalquarter; 2) The company has been subject to the reporting requirements of Section 13(a) or 15(d) of the Exchange Act for a period ofat least 12 calendar months; 3) The company has previously filed at least one annual report pursuant to Section 13(a) or 15(d) of theExchange Act; and the company is not eligible to use Forms 10-KSB and 10-QSB. A large accelerated filer (LAF) is defined as anAF with a worldwide market value of outstanding voting and non-voting common equity by non-affiliates of $700M or more (SEC,2005). A non-accelerated filer (NAF) is a firm that does not meet the definition of an AF.2 See, for example, http://www.bloomberg.com/apps/news?pid=10000103&sid=atKZoxTtzuUY&refer=us,http://www.sec.gov/rules/proposed/s70802/deloittetouche.htm, http://www.sec.gov/rules/proposed/s70805/deloitte103105.pdf, andhttp://www.sec.gov/rules/proposed/s70802/bdoseidman1.htm.


4/44

2

confirmations, inspections, and recalculations of balance sheet accounts are more reliable if performed after

year-end (Arens et al., 2010). Using evidence at or around year-end to test year-end fair value estimates is likely

to be most effective as well. Additionally, fraud experts propose extending post-fiscal year-end tests (e.g.,

examining subsequent cash receipts on year-end accounts receivable balances) substantially beyond the typical

45-day period to enhance fraud detection (Hoffman and Zimbelman, 2009). Accelerated filings that lead to

substantial reductions in audit delay could impair or curb such testing and, in turn, reduce audit/earnings

quality.

The main objectives of this study are to (1) determine if larger (vs. smaller) mandatory reductions in

audit delay are more negatively associated with changes in earnings quality, and (2) examine if this effect is

more acute for busy season audits where resources are already stretched to capacity. In additional analyses we

explore whether the relation differs between large accelerated and accelerated filers.

To meet our objectives, we examine the relation between SEC-mandated reductions in audit delay and

changes in discretionary accruals in various contexts. The mandatory nature of the SECs regulation provides a

natural experimental setting for examining the effects of a hotly debated and exogenously imposed regulatory

change on earnings quality. Previous research has established a link between audit quality and discretionary

accrual levels and has used discretionary accruals to proxy for earnings quality (e.g., Myers et al., 2003; Larcker

and Richardson, 2004). Researchers have also investigated the variables that affect audit delay (e.g., Ashton et

al., 1987). Our study is the first to examine changes in audit delay as an independent variable affecting the

quality of earnings and, in turn, we contribute important empirical evidence to the debate over the acceleration

of financial reporting.

Our results provide evidence that larger mandatory reductions in audit delay (i.e., 15 days) are more

negatively associated with earnings quality than smaller mandated reductions.3 This result is robust to using two

alternative measures of earnings/financial reporting quality and to various other cut-off points in addition to 15

days. Also, the negative relation between audit delay reductions and earnings quality appears to be more acute

3 It should be noted that the SEC has twice reduced filing deadlines by 15 days (SEC, 2002; SEC, 2005) and, in our sample, theaverage mandatory reduction in audit delay was approximately 15 days.


5/44

3

for both busy-season audits and accelerated filers (vs. large accelerated filers). These findings illustrate the

conditions under which the effects of accelerated filings on earnings quality may be more severe. Our results

related to accelerated filers provide initial evidence that caution should be taken before further reducing

accelerated filer deadlines from 75 to 60 days or expanding accelerations to smaller, non-accelerated filers. In

addition, the SEC is considering reducing the filing deadline for foreign private issuers from 6 months to 90

days after fiscal year end.4 Finally, foreign regulators may want to exercise caution before reducing filing

deadlines. For example, Canada currently has a 90-day filing deadline (which was reduced from 120 days in

2004) for firms registered on the Toronto Stock Exchange.5

The remainder of the paper is organized as follows. Section 2 provides background and discusses

previous literature. Section 3 develops our hypotheses. Section 4 describes the sample selection and research

method. Section 5 displays the results. Section 6 provides a conclusion.

2. Background

2.1. SEC Regulation

Shortly after the passage of the Sarbanes-Oxley Act (SOX), the SEC issued rule 33-8128,Acceleration

of Periodic Report Filing Dates and Disclosure Concerning Website Access to Reports (SEC, 2002). This rule

was supposed to substantially shorten the Form 10-K filing deadline for accelerated filers from 90 days to 60

days for all firms with outstanding common equity by non-affiliates of $75M or more. The reduction was

stipulated to take place over a two-year period. The deadline went from 90 days to 75 days on December 15,

2003 with a further reduction to 60 days scheduled for December 15, 2004 (which was later postponed). The

objective of the deadline reduction was to provide investors with more timely, relevant information. While

quarterly and annual reports at present generally reflect historical information, a lengthy delay before that

information becomes available makes the information less valuable to investors (SEC, 2002).

4 See, for example, http://www.sec.gov/rules/proposed/2008/33-8900.pdf.5 See, for example, http://www.tntfilings.com/sedar/sedar_deadlines.asp.


6/44

4

The acceleration of filings has been a controversial and heated topic of discussion. The SEC received

302 comments on the proposal to accelerate the deadlines, 20 supported the acceleration and 282 opposed it. In

a May of 2002 response to the SEC, BDO Seidman questioned the effects of accelerations on the reliability of

financial reporting. BDO specifically cited the concurrent explosion of accounting pronouncements, increased

globalization of clients, and expanded audit standards related to fraud detection as impediments to

simultaneously reducing audit delay and maintaining high quality reporting.6 Based on negative public reaction

to the accelerated deadlines, and concerns expressed by filers and auditors over whether they would be able to

file reports on a timely basis, the SEC adopted rule 33-8507 in November of 2004. This postponed the final

phase-in date of the 60-day filing deadline to fiscal year-ends on or after December 15, 2005 (SEC, 2004). The

one year delay of the final phase-in did little to quiet public dissatisfaction. As the end of 2005 drew near,

accounting firms and issuers voiced their concerns that further compression of available audit time could lead to

lower audit/financial reporting quality.7

The SEC passed the most recent rule on the issue of accelerated filing deadlines, rule 33-8644, in

December of 2005 (SEC, 2005). This rule created a new category of accelerated filer the large accelerated

filer. A large accelerated filer (LAF) was defined as an accelerated filer (AF) with a worldwide market value of

outstanding voting and non-voting common equity by non-affiliates of $700M or more. LAFs became subject to

a reduced 60-day deadline after December 15, 2006. AFs, firms with outstanding common equity by non-

affiliates of between $75M and $700M, remained subject to a 75-day deadline. Non-accelerated filers (NAFs),

firms with outstanding common equity by non-affiliates of less than $75M, continued to be subject to the

original 90-day deadline. Public comments were still mixed as to whether or not all companies should face the

same deadline (SEC, 2005). Figure 1 illustrates the proposed changes to the 10-K filing deadline and the

changes that were ultimately enacted.

[Insert Figure 1 about here]

6 http://www.sec.gov/rules/proposed/s70802/bdoseidman1.htm. See alsohttp://www.sec.gov/rules/proposed/s70802/pricewaterhouse.htm7 See, for example, http://www.bloomberg.com/apps/news?pid=10000103&sid=atKZoxTtzuUY&refer=us,http://www.sec.gov/rules/proposed/s70805/deloitte103105.pdf, and http://www.sec.gov/rules/proposed/s70802/bdoseidman1.htm.


7/44

5

2.2. Audit Delay

The audit report date (i.e., the audit sign-off date) is the date by which the auditors have gathered

appropriate and sufficient evidence to conclude fieldwork and issue an audit opinion on a companys financial

statements (Arens et al., 2010). Audit delay is defined as the length of time from a companys fiscal year-end to

the date of the auditors report (Ashton et al., 1987). At some point afterthe audit report date, the client files the

audited 10-K with the SEC. It is important to note that on the audit report date (not the filing date) the financial

statements are finalized and their level of quality has been determined. Accelerations which cause a reduction in

filing date, but not audit delay, would likely not impact earnings quality. For this reason, reduction in audit

delay is our independent variable of interest instead of reduction in filing date.

An audit needs to be completed several days (and sometimes weeks) in advance of the filing date. Once

an auditor has completed testing and signed off on the numbers, the client requires time to prepare the 10-K

filing. The auditor then needs time to verify that the numbers reported in the 10-K are the same numbers that

were audited. This process includes an examination of the wording and numbers contained in the footnotes and

an inspection of financial statement information supplied in supplementary schedules and the Management

Discussion and Analysis section (Arens et al., 2010). It is common for the auditor to review several drafts of the

10-K before an SEC filing is complete. In the year preceding the first accelerated filing change in December of

2003, the average time period between the audit report date and the 10-K filing date for accelerated filers (who

filed on time) was 24 days. For this same sample, the time period between the audit report date and the filing

date in the year of the first deadline was reduced to eight days. Given these periods of slack between audit

report and filing dates, the use of reductions in filing dates (vs. audit delay) would portray a less accurate

picture of how the financial reporting and audit process may have been curtailed by the acceleration of filings.

We also choose to analyze the time period between year-end and the audit report date (i.e., audit delay)

rather than the filing date because auditors presumably know how much time they will need to analyze and

correct errors in the 10-K prior to the filing date. If the accelerated filing deadline only had the effect of

reducing the slack in the time between the audit report date and the filing date, then the regulation would not


8/44

6

have been nearly as controversial. It would have been an inconvenience for the auditor to truncate their review

of the 10-K, but it would have only affected the extent and timing of audit tests in relation to presentation and

disclosure assertions. Rather, much of the controversy surrounding the accelerated deadlines centered on

mandated reductions in audit delay. A mandatory reduction in audit delay would be required if a companys

audit delay, prior to one of the accelerations, was in excess of the new accelerated filing deadline. This study

empirically examines the contexts under which such SEC-mandated reductions in audit delay may have affected

the quality of audited financial statements.

2.3. Previous Literature

Little research to date has examined the SECs decision to accelerate financial reporting. Bryant-Kutcher

et al. (2010) examine whether a sample of 82 timely filers in 2002 and 2003 (pre-SEC rule 33-8644) possess

different characteristics than a sample of 103 late filers. Results show that late filers are more highly leveraged,

less liquid, and less profitable than timely filers. In addition, they find that late filers have internal control

system weaknesses and, as one would expect, longer audit delays. The primary difference between Bryant-

Kutcher et al. (2010) and this study is that they perform a descriptive comparison (excluding earnings/financial

reporting quality) of timely filers to late filers, whereas we examine the effects of accelerated filings on the

quality of earnings across firms that were forced to have reductions in audit delay. All of the firms in our

sample filed on time but were mandated to file earlier than in previous years. Given that the chief concern over

acceleration was its impact on financial statement quality; our analyses allow us to make an important

contribution to the debate over the acceleration of 10-K filings.

Research related to the determinants of audit delay report a variety of client, auditor, and financial

factors that affect audit delay. Client size and concentration of ownership, the amount of work completed at

interim, the percentage of manager and partner hours charged to an engagement, and the provision of non-audit

services have all been found to have significant negative relations with audit delay (e.g., Ashton et al., 1987;

Knechel and Payne, 2001). Audit delay has been found to bepositively related to a structured audit approach

and incremental audit effort, a change in auditors, extraordinary items, net losses, December or January fiscal


9/44

7

year-ends, financial vulnerability, client business complexity, issuance of modified audit opinions, correction of

previously reported interim earnings, and identified material weaknesses in internal controls (e.g., Ashton et al.,

1989; Knechel and Payne, 2001; Ettredge et al., 2006). While previous research has identified important

determinants of audit delay, it has not focused on the consequences of audit delay. We are not aware of any

research that has considered the impact of audit delay on earnings quality.

Studies have argued that earnings quality (using discretionary accruals as a proxy) can be used to draw

inferences about audit quality (Myers et al., 2003). Audit quality, as measured by auditor size, has been found to

be associated with earnings management and the pricing of discretionary accruals (Becker et al., 1998;

Krishnan, 2003). Positive relations have been found between discretionary accrual levels and other audit-related

variables such as auditor changes, the issuance of qualified audit opinions, auditor litigation, audit failures, and

lower audit hours (e.g., Becker et al., 1998; Caramanis and Lennox 2008). A higher level of auditor

conservatism and audit tenure has been found to be negatively associated with discretionary accrual levels

(Francis and Krishnan, 1999; Myers et al., 2003).

In the same vein as Myers et al. (2003), this paper does not purport to speak directly to audit quality;

however, it does make the argument that lower audit quality will be reflected in more extreme income

increasing financial reporting choices by management (i.e., higher discretionary accrual levels). We do not use

the absolute value of discretionary accruals, as is occasionally used in the earnings management literature (e.g.,

Frankel et al., 2002), because the focus of this paper is capturing the effect of reductions in audit delay on

audit/earnings quality. While earnings management could be income decreasing or income increasing, auditors

face a much greater risk when the client manages income upward (i.e., higher discretionary accrual levels). For

example, Kinney and Martin (1994) analyze nine data sets of audit-related adjustments from more than 1,500

audits and conclude that audit adjustments are typically income decreasing. Bonner et al. (1998) and Palmrose

and Scholz (2004) find a prevalence of non-GAAP income increasing activity leading to SEC Accounting and

Auditing Enforcement Releases and restatement announcements, respectively. In addition, prior research

indicates that client management prefer to record income increasing, over income decreasing, audit adjustments


10/44

8

(e.g., Antle and Nalebuff, 1991; Sanchez et al., 2007). Thus, as suggested by Myers et al., (2003), auditors are

primarily concerned with clients over-reporting income (i.e., higher discretionary accrual levels). Accelerated

filings that require a large mandatory reduction in audit delay may impair auditors abilities to address this

concern and tests on signed accruals allow us to examine this possibility.8

3. Hypotheses Development

3.1. Mandatory Reductions in Audit Delay and Earnings Quality

All of the factors that have been previously determined to affect audit delay are observable by the

auditor and would normally trigger a change in audit approach. Effectively, these findings are mechanical or

consistent with the audit risk model (AICPA, 2006a; AICPA, 2006b). For example, if the auditor knows he or

she needs to perform more work due to an observed increase in an audit-related risk (e.g., a material weakness

in internal control), then a longer audit delay, on average, is expected. We consider the unobservable audit

characteristic of whether management intends to use its discretion with the reporting of accruals to manage

earnings (i.e., reduce earnings quality). Graham et al. (2006) in a survey of 401 senior financial executives and

additional in-depth interviews, find that the presence of earnings management is pervasive and income

increasing accruals are seen as a method of meeting earnings benchmarks. One CFO in the study stated, You

have to start with the premise that every company manages earnings (Graham et al., 2006, 30). The authors

report that several other interviewees provided similar comments.

If management intends to manage earnings, and audits provide the biggest constraint of such practices,

then requiring auditors to perform an audit in less time after fiscal year-end could allow for a greater

opportunity for management to do so.9 On the other hand, auditors may be able to mitigate reductions in audit

delay by performing more interim procedures, relying more on the clients internal controls/systems, and using

advanced audit technology (cf., PCAOB, 2004; PCAOB, 2007; Brazel and Agoglia, 2007). Such tactics may

8 In addition to the argument contained herein for examining signed accruals to evaluate earnings quality, Hribar and Nichols (2007)find that using unsigneddiscretionary accrual models leads to an over-rejection of the null hypothesis of no earnings management andexposure to correlated omitted variables.9 This argument is similar to that provided by Caramanis & Lennox (2008, 117), clients that wish to manage earnings can anticipatethat hard-working auditors are more likely to thwart their earnings management attempts.


11/44

9

reduce the extent to which post-fiscal year-end audit procedures/evidence are needed to provide an acceptable

level of audit quality. However, as we will describe below, when mandated reductions are large (e.g., greater

than 15 days), it becomes less likely that these factors will overcome a substantial reduction in post-fiscal year-

end audit time.

As stated previously, the SEC twice reduced filing deadlines by 15 days. In the text of Rule 33-8128, the

SEC disclosed that those who objected to the proposal raised several common concerns over the extentof

acceleration (emphasis added, SEC, 2002). For example, while a 5-day reduction in audit delay may not

impact the process of confirming accounts receivable, a 20-day reduction might require receivables to be

confirmed at an interim date. In extreme cases, this change in timing could allow management to fraudulently

manipulate earnings by recording fictitious sales afterthe auditor has confirmed accounts receivable. This is a

particular concern given the concentration of frauds and restatements related to improper revenue recognition

(Beasley et al., 1999; Gullapalli, 2005b; Beasley et al., 2010).

Given the double-entry nature of accounting, many audit firms have historically taken a balance sheet

approach to the audit. That is, they devote the majority of their substantive testing (i.e., evidence evaluation) to

balance sheet (vs. income statement) accounts (e.g., Gopez 1954; Basilo, 2007). The most effective time to

inspect, confirm, and recalculate balance sheetaccounts is at year-end (Arens et al., 2010). As noted previously

with respect to accounts receivable confirmation, a substantial mandated reduction in audit delay might require

auditors to perform some (if not all) of this testing at an interim date. Concurrently, audit standards explain that

expanding (reducing) interim (year-end) testing may actually increase audit risk and exacerbate the problems

auditors face with accelerated 10-K filings (AICPA, 1983).10

The listing of audit-related adjustments to the financial statements (i.e., misstatements identified) is

typically not finalized until after year-end. With less time after year-end to audit the financial statements, the

audit team would have less time to fully evaluate both the qualitative and quantitative aspects of these

misstatements. Also, auditors obtain reliable, post-fiscal year-end evidence to evaluate year-end account

10 SAS No. 45 (AICPA, 1983) provides several examples of when moving substantive testing from post year-end to interim may beespecially challenging and should be discouraged (e.g., when circumstances predispose management to misstate financial statements).


12/44

10

balances. A substantial reduction in audit delay as a result of accelerated filings may curb these practices. For

example, suppose that a company manipulates sales through channel stuffing (i.e. selling excessive amounts

of product at discounted prices near year-end to distributors with a verbal right of return). The most effective

means of detecting this type of earnings management would be to (1) extend the amount of time after year-end

examining subsequent cash receipts (evidence that year-end accounts receivable were valid and collectible) and

(2) extend the amount of time after year-end searching for an unusually high number of returned products

(evidence that revenue recognition of shipments immediately prior to year-end was suspect). Accelerating the

filing deadline impairs such tests. In addition, auditors often require considerable time after year-end (and

substantial post-fiscal year-end evidence) to examine client estimations (e.g., the reserve for obsolete inventory,

income tax provision, fair value ofcomplex financial instruments). The recent trend from historical costs to fair

value accounting makes the year-end analysis of client estimates even more complex and important.

In addition to altering the timing of testing and the quality of evidence gathered, large mandated

reductions in audit delay may also increase time pressure on the engagement team. Experimental research has

shown that increasing time pressure decreases audit effectiveness and can result in behavior that reduces audit

quality (e.g., accepting doubtful evidence, truncating sample selections) (McDaniel, 1990; Coram et al., 2004).

As time pressure increases beyond moderate levels, research shows that auditors become overwhelmed and

suffer anxiety which has a detrimental effect on performance (DeZoort and Lord, 1997). For large mandated

audit delay reductions, audit firms generally cannot simply assign more staff to an audit to perform the same

amount of testing in a shorter year-end window. Obtaining additional staff, specifically for less

prestigious/smaller public audit clients, may be difficult as audit firms are currently operating at or near 100%

capacity (McGee, 2005; Gullapalli, 2005a; Rose, 2007). Also, if additional audit staff is available, their current

lack of assignment to a concurrent audit engagement may suggest a possible lack of competence.11

11 Recent discussions with practitioners support the supposition that human resources are currently stressed at audit firms and that,when additional staff is obtained to meet reduced audit deadlines, the staff typically lack competence and/or required industryexpertise.


13/44

11

Finally, when increased risks require an expansion of audit testing, much of the additional work cannot

be compressed into a short window. For example, discrepancies found during the confirmation of accounts

receivable must be followed up by the client and the auditor (Caster, 1990).12 Such factors delaying the audit are

out of auditors control and simply increasing the number of auditors assigned to the engagement will not speed

the process. Much post-year-end audit work requires effort and time from the audit client as well as others that

the auditor cannot control (e.g., customers, creditors, attorneys). Thus, assigning more auditors cannot take the

place of allowing more post-fiscal year-end time for evidence accumulation, testing, research, and

correspondence between the auditor, the client, and other independent third parties.

In response to accelerated filings, we expect larger mandated reductions in audit delay to more

negatively impact earnings quality than smaller reductions. Small reductions in audit delay are commonly due

to calendaring issues. For example, some firms have year-ends that fall on the last Friday of the month as

opposed to the last day of the month. Given a change in a clients year-end calendar date, auditors routinely

adjust the audit timeline by up to a week. Also, under the previously motivated premise that more post year-end

audit time leads to more effective testing and less time pressure, the size of the mandatory reduction should be

inversely related to the level of audit quality. Therefore, we do not expect small mandated reductions in audit

delay to have a serious effect on earnings quality. However, the SEC has twice reduced filing deadlines by 15

days. We suspect larger reductions in audit delay (15 days) will more adversely affect earnings quality than

smaller reductions. Formally, we hypothesize:

H1: Larger mandatory reductions in audit delay (15 days) are more negatively associated with changesin earnings quality than smaller mandatory reductions in audit delay.

3.2. Busy Season

Archival data on the amount of audit hours worked on an engagement, when fieldwork began, and other

data related to the extent of time constraints auditors experience on an engagement are not publicly available.

However, one variable that is available and relates to the amount of time pressure the auditor feels is workload

12 For example, if a customer reports that an invoice was no longer due at year-end because the product was returned, then the clientmust research when the returned product was received, compare that date to the year-end date, and communicate that information andrelated evidence to the auditor.


14/44

12

compression. Workload compression occurs during auditors busy season and exists because most

companies fiscal year-ends coincide with the calendar year-end (Lopez-Acevedo, 2007). Comments in

response to the initial SEC proposal specifically expressed concern regarding the effects of accelerations on

companies with fiscal year-ends that coincide with calendar year-ends (SEC, 2005). Our second hypothesis

examines whether the relation described in Hypothesis 1 is more acute for busy season audits.

The adverse impact of busy season audits on audit/earnings quality has been shown across a variety of

research methods. Lopez-Acevedo (2007) finds that busy season audits, defined as those with a December or

January fiscal year-end date, display greater magnitudes of discretionary accruals. In an experimental setting,

Agoglia et al., (2010) find that increased workload compression leads managers and partners to employ less

effective audit techniques. A longitudinal study of auditors at staff through partner levels finds that, by the end

of busy season, these professionals experience greater emotional exhaustion from their work and are more

depersonalized in their approach to their job (Sweeney and Summers, 2002).

Busy season engagements are therefore more likely to be served by overworked and exhausted audit

teams who may employ less effective audit procedures in order to maintain audit efficiency. Also, during busy

season, human resources at audit firms are often stretched to capacity. Busy season audits, in particular, will

have a more difficult time obtaining additional, competent staff to assist in audits where accelerated filings

require a mandatory reduction in audit delay. Last, busy season audits have historically had longer audit delays

than non-busy season audits (Knechel and Payne, 2001). Thus, it is likely that larger/more detrimental

reductions in audit delay will occur for busy season audits. In sum, the audit quality of busy season audits

should be more sensitive to mandated reductions in audit delay and this heightened sensitivity may ultimately

affect the quality of audited earnings. Therefore, our second hypothesis is:

H2: The negative association between larger mandated reductions in audit delay and changes inearnings quality is stronger for busy season audits than non-busy season audits.

4. Sample Selection and Research Method

4.1. Sample


15/44

13

We derive our sample from the years that the two mandatory filing reductions took effect. The first

deadline change (i.e., 90 days to 75 days) took effect for all accelerated filers (LAFs and AFs) filing after

December 15, 2003. The second deadline change (i.e., 75 days to 60 days) took effect for all LAFs filing after

December 15, 2006. Thus, mandatory reductions in audit delay could occur during fiscal year-ends from

December 15, 2003 through December 14, 2004 and from December 15, 2006 through December 14, 2007. As

shown in Table 1, our sample consists of 933 firm-year observations (361 from first filing deadline change and

572 from the second deadline change) where the audit delay was reduced due to a change in the filing deadline

(i.e., in the year prior to the deadline change, the audit delay was greater than the new filing deadline and was

subsequently reduced during the year of the change to meet the new deadline). The 933 firm-year observations

represent 868 unique firms (i.e., 65 LAFs were required to make audit delay reductions during both filing

deadline changes and are included in our sample twice).

[Insert Table 1 about here]

Audit Analytics is our source for audit delay data and Compustat is our source for financial data. From

our Audit Analytics sample, we exclude firms with negative audit delays or missing audit report dates.13From

our Compustat sample, we exclude firms with missing control variables and data to calculate discretionary

accruals, as well as firms with a qualified or no audit opinion. Combining the Audit Analytics and Compustat

samples provides a sample of 10,577 firm-year observations. We then exclude financial institutions and

regulated industries,14 firms that were not subject to the first filing deadline changes (NAFs), and firms that

were not subject to the second deadline reduction (NAFs and AFs). We delete observations with no change in

audit delay and firms with increases in audit delay because we are only concerned about firms that were forced

to decrease their audit delay due to the acceleration of 10-K filings. We drop observations with prior year audit

delays greater than the old filing deadline because the reduction in audit delay in the year of the deadline change

would be, in large part, due to the firm having filed late in the previous year. Finally, we exclude observations

13 We traced a sample of the firm-years with negative audit delay to the report filed on edgar.gov. It appears that the negative auditdelay is a result of a typo on the reports as they were originally filed (i.e., audit report date precedes the fiscal year-end date), andtherefore does not indicate a problem with the Audit Analytics database.14 Financial and regulated industry firms are excluded because of their unique nature of calculating discretionary accruals (Frankel etal., 2002; Tucker and Zarowin, 2006).


16/44

14

whose prior year audit delay was less than the new filing deadline. Firms in this category had their prior year

audit completed on time to meet the new deadline and any further reduction in the year of the change was less

likely due to the deadline change (vs. firms in our sample where the prior year audit delay exceeded the filing

deadline and an audit delay reduction was mandatory). For example, suppose an accelerated filer had an audit

delay of 50 days in the year prior to the first deadline change (90 days to 75 days) and an audit delay of 40 days

in the year of the change. The 10-day reduction in audit delay was unlikely mandatory because they could

have completed the audit in 50 days again in the year of the change and still had 25 days to file.15 Thus, our

final sample includes firms with audit delays greater than 75 days prior to the first deadline change and 60 days

prior to the second deadline change. As a result, their auditors were mandated to complete their audits earlier to

meet the new filing deadlines.

Prior research suggests that discretionary accruals are a noisy measure of actual earnings management

which makes inferences of earnings management difficult without conditioning on the managerial incentive to

manage earnings (Dechow and Skinner, 2000). As such, we consider a setting where management has incentive

to boost reported earnings on the basis of poor performance relative to a benchmark. Consistent with Kim et al.

(2003), we assume management has incentive to increase earnings upward via income increasing accruals if

current cash flow from operations scaled by lagged total assets is less than the industry median. Industry is

defined using the two-digit SIC code.16 Kim et al. (2003) rely on a theory of income smoothing that predicts

that managers have an incentive to boost current earnings in poor times by borrowing against future earnings to

mitigate the likelihood of dismissal (Fudenberg and Tirole, 1995). We condition using this method because, in

our context, auditors focus on firms over-reporting net income/constraining income increasing accruals (Myers

et al., 2003; Graham et al., 2006).

15 It is possible that, for some of these firms, the acceleration of filing deadlines also required a mandatory reduction in audit delay(e.g., firms with insufficient slack between their audit report date and filing deadline, firms that require a substantial amount of timepost-audit to file their 10-K). By excluding these firms we cannot conclude that our results apply to all firms that did not meet the SECaccelerated filing deadlines in the previous year (e.g., firms where accelerations required a mandatory reduction in filing date, but notaudit delay). Rather, our results pertain only to firms that had an audit delay in excess of the deadlines (i.e., the deadlines explicitlyrequired a mandatory reduction in audit delay). We feel that the sample of mandatory audit delay reductions used in our analysesprovides a natural experimental setting for examining the effects of a hotly debated and exogenously imposed regulatory change onearnings quality.16 We observe qualitatively similar results when we re-perform our analyses by defining industry using a four-digit SIC code.


17/44

15

4.2. Tests of Hypotheses

To test our hypotheses, we estimate the following regression model17:

ChDACt= 0 + 1ChDELAYt+ 2CFOMedt+ 3ChDelayt*CFOMedt+ 415-DAYt +5ChDELAYt*15-DAYt+ 615-DAYt*CFOMedt+ 7ChDELAYt*15-DAYt*CFOMedt+8SEASONt+ 9LOSS t+ 10ChCFOt+ 11ChAbsCFOt+ 12ChB/Mt+ 13ChlnMVEt+14ChLEVERAGEt+ 15EXTRAORDINARYt+ 16PARAGRAPHt+ 17ACQUISTIONt+

18ChAUDITORt+ 19NEWBIG4t+ 20OLDBIG4t-1 + 21PYDACt-1 + 22-24Year + 25-36Industry + (1)

ChDAC refers to current year discretionary accruals less prior year discretionary accruals, measured using the

Jones (1991) model with two additional modifications. First, we include ROA as described by Kothari et al.,

(2005). Second, we include a dummy variable equal to one if the company had negative cash flow from

operations and a variable that measures the interaction between the negative cash flow dummy and total cash

flow from operations (scaled by total assets) as described by Ball and Shivakumar (2006).18

Ball and

Shivakumar argue that the relation between accruals and cash flows is not linear because unrealized losses are

recognized immediately via accruals while unrealized gains are delayed.

We define audit delay as the audit report date less the fiscal year-end date (Ashton et al., 1987). The

independent variable ChDELAY in our model is the absolute value of change in audit delay. Because all of our

observations experienced a mandated reduction in audit delay, a greater ChDELAY indicates a larger mandated

reduction. Using the absolute value provides a positive value for ChDELAY which we feel is easier to interpret

in a multivariate regression. In general, we expect a positive relation between ChDELAY and ChDAC (larger

mandated reductions in audit delay are associated with increased discretionary accrual levels).

17 To test Hypothesis 2, we analyze the same regression model (removing the SEASON control variable), but partition the sample on

busy season vs. non-busy season audits. Consistent with Peterson (2009), we control for standard error bias due to repeated measuresof firm and year by including year dummies in the regression and performing a cluster regression with a firm identifier as a repeatedmeasure (SAS procedure proc surveyreg with a gvkey cluster).18 The specific model is TA = 0 + 1(1/Assetsit-1) + 2(Salesit ) + 3PPEit + 4ROAit + 5NegCFOit + 6Neg CFOit x CFOit + it.Consistent with Dechow et al. (1995), the estimate of0, 1, 2, 3, 4, 5 and 6 were obtained from the original Jones Model (withmodifications for ROA and cash flows). The change in sales is then adjusted for the change in receivables for the calculation ofdiscretionary accruals. Due to sample size restrictions, our measure of discretionary accruals is estimated from cross-sectional models.Further, Bartov et al. (2000) find that the cross-sectional Jones model and the cross-sectional modified Jones model outperform theirtime-series counterparts in detecting earnings management. We estimate model coefficients from cross-sectional industry regressionsby two-digit SIC codes. We require a minimum of 10 observations for each two-digit SIC code and year combination. We estimate themodel on all firms in Compustat. We winsorize all variables at the 1st and 99th percentile.


18/44

16

We include two separate dummy variables (i.e., 15-DAY and CFOMed) in the regression to test

Hypothesis 1. First, we include a dummy variable (15-DAY) equal to one if the firm had a large reduction in

audit delay (i.e., greater than or equal to 15 days and zero otherwise). Fifteen days was chosen as a cutoff

because (a) the SEC twice accelerated filings by 15 days (SEC, 2002; SEC, 2005) and (b) the mean mandatory

reduction in our sample (15.73 days) was approximately equal to these 15 day accelerations. To condition on

the incentive to manage earnings, we include a dummy variable CFOMed that is coded 1 if the firms cash flow

from operations scaled by lagged total assets is less than the industry (2-digit SIC) median; and 0 otherwise

(Kim et al., 2003). Our test variable is the three-way interaction between the 15-DAY, ChDELAY, and

CFOMed. A positive and significant coefficient on the three-way interaction provides support for our first

hypothesis that, for firms that have incentive to manage earnings upward (CFOMed=1), larger mandatory

reductions in audit delay (15-DAY=1) are more positively associated with changes in discretionary accruals

than smaller mandatory reductions in audit delay. It should be noted that, in our study, apositive relation with

changes in discretionary accruals indicates a negative relation with changes in earnings quality.

Consistent with prior research (Knechel and Payne, 2001; Lopez-Acevedo, 2007), we classify audits of

companies with fiscal year-ends during the months of December and January as busy season audits. To test our

second hypothesis, we partition the sample between busy season and non-busy season firms and remove the

SEASON (1= busy season, 0 otherwise) control variable from Model (1). We expect the three-way interaction

(15-DAY*ChDELAY*CFOMed) to be more significant for busy season firms than non-busy season firms.

Our regression model includes multiple control variables. Given that we examine changes in audit delay

and earnings quality, we use changes (rather than levels) measures for all continuous and auditor change control

variables. We include an indicator variable as to whether or not the firm reported a loss for the year (LOSS),

because firms are expected to manipulate accruals in a systematically different way during loss years (Lopez-

Acevedo, 2007; Frankel et al., 2002). Prior research suggests that discretionary accrual models do not

completely extract nondiscretionary accruals that are correlated with firm performance (Frankel et al., 2002).

Thus, we control for firm performance by including the change in cash flows from operations (ChCFO) (e.g.,


19/44

17

DeFond and Subramanyam, 1998; Choi et al., 2006). Because the relation between cash flows and discretionary

accruals is not linear, we also control for the absolute value of cash flows from operations (ChAbsCFO)

(DeFond and Subramanyam, 1998; Frankel et al., 2002). Following Frankel et al. (2002) and Choi et al., (2006),

we control for firm growth by including the change in the ratio of book to market value (ChB/M). We control

for firm size using the change in the natural log of the market value of equity (ChlnMVE) (Frankel et al., 2002).

We control for leverage (ChLEVERAGE), measured as the change in the ratio of total liabilities to total assets

(Frankel et al., 2002).

We also control for variables that the prior literature has found to impact audit delay (e.g., Bamber et al.,

1993): extraordinary items (EXTRAORDINARY), an explanatory paragraph added to an unqualified opinion

(PARAGRAPH), and whether the company made an acquisition during the year (ACQUISITION). In addition,

the audit delay literature finds that auditor changes are associated with changes in audit delay (Schwartz and

Soo, 1996). Our model controls for auditor change in three different ways. We include a dummy variable

(ChAUDITOR) equal to one for any change in auditor during the year. In addition, many firms were switching

between Big 4 auditors and non-Big 4 auditors during our sample period and we specifically control for these

effects. We include a dummy variable (NEWBIG4) equal to one if the firm changed from a non-Big 4 auditor to

a Big 4 auditor and a dummy variable (OLDBIG4) if the firm changed from a Big 4 auditor to a non-Big 4

auditor.

Because accruals reverse over time, consistent with Ashbaugh et al. (2003) we control for prior year

discretionary accruals (PYDAC). Finally, we include three dummy variables to control for year and, following

Ashbaugh et al. (2003), we include 12 dummy variables to control for industry. In order to maintain

parsimonious tables, the year and industry dummy variables are not tabulated.19

19 Prior research has found that material weaknesses in internal control positively affect audit delay (Ettredge et al., 2006). For oursample, the substantial effect of SOX Section 404 on audit delay would likely only impact one year (2004) and only 2004 observationswith fiscal year-ends from November 15, 2004 to December 14, 2004. We rely on the year dummy for 2004 to control for the impactof auditing the effectiveness of internal controls over financial reporting. Industries were divided into the following groups: SIC 0100-1499, SIC 1500-1999, SIC 2000-2199, SIC 2200-2399, SIC 2400-2799, SIC 2800-3299, SIC 3300-3499, SIC 3500-3999, SIC 4000-4899, SIC 4900,SIC 5000-5299, SIC 5300-5999, and SIC 7000-7999.


20/44

18

5. Results

5.1. Descriptive Statistics

Table 2 provides descriptive statistics for our variables of interest and control variables. Change in

discretionary accruals (ChDAC) ispositive (.0015), or income increasing, as our theory would predict for a

sample where mandatory audit delay reductions occurred. In addition, ChDAC, in the year following the

deadline change (ChDACt+1), is negative at the mean and median. This illustrates the reversing nature of

accruals. The mean reduction in audit delay is 15.7 days. As previously noted, the dummy variable 15-DAY

represents a partition at approximately the mean of our sample and reflects the two 15-day accelerations

mandated by the SEC. Table 2 also illustrates that busy season audits do appear to be workload compressed.

Indeed, for our sample, 73% of all audits were for fiscal years-ending in two months: December and January.

Another item of interest is that 70% of the firms in the sample had an explanatory paragraph in their audit

report. We randomly selected a few audit reports in the sample with explanatory paragraphs and found that most

were simply an emphasis of a matter (e.g., an accounting change or adoption of an accounting standard) with a

corresponding reference to a footnote.


Table 3 presents the correlation matrix. Overall, the variables ChDELAY and ChDAC are positively, but

not significantly, correlated. This suggests that the positive relation between these two variables may be

moderated by the context of the audit delay reduction.


5.2. Results of Hypotheses Testing

Table 4 provides results of two regressions with different proxies for earnings quality. The first regression

includes ChDAC as the dependent variable. The positive and significant (p = .033) coefficient on the three-way

interaction of ChDELAY*15-DAY*CFOMed provides support for our first hypothesis. For firms with

incentives to manage earnings (i.e., below the industry median for CFO), larger mandatory reductions in audit


21/44

19

delay appear to be more positively associated with changes in discretionary accruals (i.e., negatively associated

with earnings quality) than smaller reductions.

The average firm in our sample has total assets of $4.7 billion and net income of $228 million. To

illustrate the economic significance of the coefficient on the three-way interaction, the magnitude of the

coefficient (0.0038) represents an additional increase in discretionary accruals of 0.38% of total assets for a one-

day mandatory reduction when the reduction is large ( 15-days) and cash flows fall below the industry median.

For these firms, a one-day reduction in audit delay leads to an increase of approximately $18 million in

discretionary accruals or approximately 8% of average net income.

Our results related to Hypothesis 1 illustrate that, when incentive to manage earnings is higher (CFOMed

= 1), the positive relation between ChDELAY and ChDAC is stronger for larger mandatory reductions in audit

delay (reductions 15 days). Consistent with this finding and the nature of the interaction posited by

Hypothesis 1, when we re-perform the regression (untabulated) with a subsample of firms that have below

industry median CFO (i.e., CFOMed = 1 or firms with greater incentives to manage earnings) and larger

mandatory reductions in audit delay (15-DAY=1), the coefficient on ChDELAY (main effect) is positive and

significant (p < .05). On the other hand, when we re-perform the regression with a separate subsample of firms

with below industry median CFO (regardless of the size of audit delay reduction), we find that (similar to Table

4), the coefficient on ChDELAY is negative, very close to zero (-0.0001), and not significant (p = .81).

However, when we add the 15-DAY dummy and the 15-DAY*ChDELAY interaction to the regression for this

subsample, the interaction is positive and significant (p < .01). In short, the overall relation between ChDELAY

and ChDAC is not significant for our entire sample (see the correlation in Table 3) and our results are largely

driven by larger mandatory reductions in audit delay.

The second regression provides a sensitivity test by including change in working capital accruals as the

dependent variable. We consider the Dechow and Dichev (2002) empirical measure of accrual quality as a

second measure of earnings quality. Dechow and Dichev derive their measure of accrual quality as the residuals

from firm-specific regressions of changes in working capital on past, present, and future operating cash flows


22/44

20

(CFO). Specifically, Dechow and Dichev estimate the following firm-level time-series regression to derive a

measure of working capital accrual quality:

WCit = 0 + 1CFOit-1 + 2CFOit+ 3CFOit+1 + it (2)

where WC is the change in working capital from year t-1 to year t.WC is computed as Accounts

Receivable + Inventory - Accounts Payable - Taxes Payable + Other Assets. Specifically, WC is

computed from Compustat items as WC = - (#302 + #303 + #304 + #305 + #307). All variables in Model (2)

are scaled by average total assets. Dechow and Dichev use the standard deviation of the residuals from Model

(2) as a firm-specific measure of accrual quality. Dechow and Dichev require at least eight years of data to

estimate Model (2). We do not have eight years of data for all the mandatory reduction firms. Therefore,

consistent with Jones et al., (2008), we estimate Model (2) cross-sectionally and use the residual from Model (2)

as a measure of accrual quality. Jones et al. find that the Dechow and Dichev measure provides greater

predictive power for extreme earnings management (i.e., fraud) than any iteration of the discretionary accrual

models based on Jones (1991). We find the coefficient on the three-way interaction of ChDELAY*15-

DAY*CFOMed to be positive and significant (p = .009), which provides additional support for our first

hypothesis.


Table 5 provides results related to Model (1) when we partition the sample into busy season (first

regression) and non-busy season audits (second regression). The coefficient on the three-way interaction of

ChDELAY*15-DAY*CFOMed is positive and significant (p = .019) for busy season audits and the same

coefficient is not significant for non-busy season audits. Consistent with Hypothesis 2, we find that the negative

association between larger reductions in audit delay and changes in earnings quality is stronger for busy season

audits. We find qualitatively similar results (untabulated) related to Hypothesis 2 when we use ChWCAC as the

dependent variable.


5.3. Filer Type


23/44

21

The SECs decision to have three distinct groups of filers (LAFs, AFs, and NAFs), and three separate

reporting deadlines (60 days, 75 days, and 90 days, respectively), was quite controversial. As stated earlier, the

SEC received a substantial amount of comments that opposed the accelerated filing deadline. Commentators

disagreed on the proposed definition of an accelerated filer. Some companies and associations (e.g., Comcast

Corporation and Troutman Sanders LLP) expressed that all public companies should be required to adhere to

the same deadline. The American Bar Association took issue with the assumption that larger companies are

better equipped than other filers to meet the accelerated deadline, large businesses tend to be more complex,

often with international operations, multiple divisions and subsidiaries and investments from other entities from

which they often must await reports.20 Others, including the AFL-CIO and KPMG LLP, agreed with the notion

of excluding smaller companies because they may not have the necessary resources or infrastructure to meet the

accelerated deadline (SEC, 2002). We aim to shed light on this issue by examining whether the relation between

large mandated reductions in audit delay and changes in earnings quality differs by filer type (AFs vs. LAFs).

Table 6 presents the results of when we partition Model (1) on filer type. The first regression includes

accelerated filers (AFs) and the second regression includes large accelerated filers (LAFs). In accordance with

SEC rule 33-8644 (SEC, 2005), we classify an AF as a firm with a market value of equity (MVE) between

$75M and $700M and a LAF as a firm with MVE greater than $700M.21

AFs are subject to the first accelerated

filing deadline change and currently have 75 days after year-end to file their 10-K. LAFs are subject to the first

deadline change and the second deadline change and now have a 60-day deadline to file their 10-K. The

coefficient for the three-way interaction of ChDELAY*15-DAY*CFOMed is positive and marginally

significant (p = .082) for the audits of AFs. The same coefficient is positive but not significant for audits of

LAFs. As a result, we provide some evidence that the negative association between large reductions in audit

delay and earnings quality is more pronounced for AFs.


20 http://www.sec.gov/rules/proposed/s70802/skeller1.htm21 The SEC classifies filers based on public equity float and other factors (see footnote 1). We classify each firms filer status based onMVE because it is a measure that is publicly available and should be highly correlated with public equity float.


24/44

22

Based on the prior discussion that a move from year-end to interim testing (caused by mandated

reductions in audit delay) could result in greater discretionary accruals/lower earnings quality, a less

pronounced effect for LAFs is not totally unexpected. Before SEC 10-K accelerations, the sheer size of LAFs

and their more complex year-end reporting issues required that the bulk of all audit testing be performed prior to

year-end. Indeed, empirical research prior to the acceleration of filings finds a negative relation between client

size and audit delay (e.g., Bamber et al., 1993). If audits of LAFs tended to already employ more extensive

interim testing/exhibit shorter audit delays, then they were less likely to be negatively impacted by accelerated

filings. In addition, for large accelerated filers that did employ substantial year-end testing prior to accelerations,

Brazel et al. (2010) find a positive relation between client size and the degree to which auditors change the

staffing and timing of their testing. Because auditors of LAFs tend to be present at the client throughout the

year, it may have been easier to shift testing from year-end to interim to meet filing deadlines and still maintain

an acceptable level of audit quality. LAF clients also tend to be more prestigious and, if needed to meet filing

deadlines, these audits typically have a greater ability to procure higher quantities of competent audit staff (if

available).

Our result indicates that smaller publicly traded firms are most likely to exhibit a reduction in earnings

quality as a consequence of an acceleration which reduces audit delay. Currently, AFs and NAFs face 75 and 90

day 10-K filing deadlines, respectively. We therefore provide initial empirical evidence that further

accelerations for AFs may not be advisable and that expanding accelerations to smaller NAFs could lead to

lower earnings quality as well.

5.4. Sensitivity Analyses

5.4.1. Different Cut-off Points

We employ a 15-day cut-off to test Hypothesis 1. We consider whether our tabulated regression result in

Table 4 is sensitive to other cut-off points. Table 7 provides the results of regressing Model 1 multiple times

with different cut-off dummy variable interactions. The cutoff dummies range from 10 to 20 days. We report


25/44

23

only the coefficient on the variable of interest the three-way interaction of ChDELAY*15-DAY*CFOMed.

We find consistent results for cut-off points between 10 and 20 days.


5.4.2. Meet or Beat Analyst Forecast

In addition to using discretionary accruals and working capital accruals, we examine whether large

mandatory reductions in audit delay increase the likelihood that firms meet or beat the consensus analyst

forecast. As stated previously, Graham et al. (2006) find that income increasing accruals are used as a method

of meeting earnings benchmarks and auditors focus on constraining income increasing accruals (Myers et al.,

2003). Prior research finds a significant reward (penalty) for meeting or beating (missing) analysts forecasts

(Bartov et al., 2002; Kasznik and McNichols 2002; Skinner and Sloan 2002). We create a dummy variable

(Meet or Beat Expected EPS) set equal to 1 if the actual earnings per share (per First Call) is equal to or greater

than the mean analyst forecast of the last First Call update prior to the earnings announcement date; and zero

otherwise (Cheng and Warfield 2005).We include Meet or Beat Expected EPS as the dependent variable and

15-DAY, ChDELAY, and the interaction between 15-DAY and ChDELAY as independent variables (along

with the other control variables from Model (1)). Table 8 presents our results. We find that the interaction,

ChDELAY*15-DAY, is positive and significant (p = .029). In other words, the relation between mandated

reductions in audit delay and the probability of meeting or beating the consensus analyst forecast of EPS is

significantly greater if the audit reduction is greater than or equal to 15-Days (i.e., large audit delay reductions

significantly increase the probability of firms meeting or beating the consensus analyst forecast).


5.4.3. Conservatism

As a broader alternative analysis we consider the effect of mandatory reductions in audit delay on the

level of conservatism in the audited financial statements (i.e., more timely loss recognition than gain

recognition). Searching for unrecorded liabilities or understatements is one of the more difficult audit

procedures (e.g., testing the audit objectives of completeness and realizable value). For example, while an


26/44

24

auditor samples from the detail of the account balance population to perform tests of existence, the appropriate

population for sample selection related to completeness tests is not always clear. Also, auditors must develop

independent and precise estimates in order to effectively test the realizable value assertion via analytical

procedures (e.g., an estimate of the allowance for doubtful accounts). We suspect that auditors who must

compress the timing of their audits are less likely to identify, and have their clients accrue for, unrealized losses

such as asset impairments and contingent losses. Asset impairment tests and other valuation tests, as well as

tests for other accrued losses, are most effectively performed after year-end (i.e., the balance sheet date).

Therefore, we would expect mandated reductions in audit delay to be associated with less timely loss

recognition. We measure conservatism using the asymmetric operating accrual-cash flow test introduced by

Ball and Shivakumar (2005).

We employ Ball and Shivakumars cash-flow based measure of conservatism to test whether firms with

mandated reductions in audit delay demonstrate less financial statement conservatism than firms in our sample

period that did not experience a mandated reduction. Ball and Shivakumars measure of conservatism has the

potential to be more reliable in an earnings management setting than Basus (1997) returns-based measure if

one assumes that the market does not detect earnings management (i.e., stock returns do not measure true

economic return) and firms only manage accruals (i.e., cash flows accurately reflect real firm performance).

We estimate the following model:

ACCt= 0 +1DCFOt+ 2CFOt+ 3DCFOt* CFOt+ 4DMANDt+5DMANDt* DCFOt

+6DMANDt*CFOt+ 7DMANDt*DCFOt*CFOt+8SIZEt+9SIZEt*DCFOt+ 10SIZEt*CFOt +11SIZE*DCFOt*CFOt + t. (3)

WhereACCt is Accruals at time t, scaled by average total assets ((Compustat Item #123 #308) / average total

assets (item #6));DCFOt is an indicator variable equal to 1 if cash flows are less than zero in year t, zero

otherwise; CFOt is cash flows from operations in year t((item #308)/ average total assets (item #6)); DMAND

is coded 1 if the firm was included in our sample of firms with a mandatory reduction in audit delay and 0 if the


27/44

25

firm was excluded from our sample presented in Table 1 but has the necessary data required for this analysis;

SIZEis the rank of total assets in year t.22

Table 9 presents conservatism results. Consistent with Ball and Shivakumar (2005), the coefficient for

CFO is expected to be negative because earnings and cash flows are negatively correlated. The coefficient for

DCFO*CFO is expected to be positive because accrued losses are more likely to be recorded in periods of

negative cash flows. The negative and significant (p = .022) coefficient onDMAND*DCFO*CFO is consistent

with our theory. The financial statements of firms with mandated reductions in audit delay are less conservative

than the financial statements of firms without mandated reductions. To illustrate the economic significance of

the coefficient on the three-way interaction, the magnitude of the coefficient (-0.2848) implies that 28% of

additional cash flow is offset by accruals in years of negative cash flow for mandatory firms (relative to years of

negative cash flow for non-mandatory firms).23 In other words, compared to non-mandatory firms, firms with

mandatory reductions in audit delay accrue substantially less unrealized losses in cash-loss years.


5.4.4. Alternative Incentive

We consider an alternative measure for the incentive to manage earnings (i.e., an alternative measure for

CFOMed). We investigate firms that meet both of the following criteria: (1) their year-to-date earnings at the

end of the third quarter are less than prior year,and (2) their consensus analyst forecast for annual earnings is

greaterthan the prior year. Firms meeting these criteria will need to significantly improve their earnings in the

fourth quarter if they intend to achieve the consensus analyst forecast. They must do so despite the fact they

have been unable to match their prior years performance through the third quarter and the consensus analyst

forecast for annual earnings has risen from the prior year. In short, for these firms, performance has dropped but

expectations have risen. Such a scenario makes the necessity to manage earning near year-end more likely.

22Consistent with Ball and Shivakumar (2005), the model also includes a control variable for size, SIZEt, which equals the rank oftotal assets at the end of year t, standardized to the interval (0,1). The interactions SIZEt*DCFOt, SIZEt*CFOt, andSIZEt*DCFOt*CFOtare also included.23 Alternatively, we could have performed this test on the same sample as the previous tests (i.e., only mandatory filers) and used 15-DAY as the test variable that we interact with DCFO and CFO. However, the sample size is much smaller (933 vs. 9,563 firms) andthe coefficients are much higher. We find the current interpretation of our reported results is more consistent with the original Ball andShivakumar (2005) study.


28/44

26

This measure of the motive to manage earnings is particularly appropriate for our study. It focuses on

managements incentive to manage earnings in the fourth quarter. Prior research suggests that earnings

management is more likely to occur closer to year-end as earnings shortfalls become known to management

(Givoly and Ronen, 1981; Das et al., 2009). Accelerated filing deadlines will have the greatest impact on

auditors ability to detect earnings management that occurs near year-end. In short, for firms meeting the

aforementioned criteria, earnings management is more likely to occur in the fourth quarter and less likely to be

detected by auditors who must reduce their post year-end audit testing due to the accelerated filing deadlines.

The results (untabulated) are qualitatively the same as those in our main analysis. The three-way interaction

between ChDelay, 15-Day, and this alternate incentive measure is positive and significant (p < .05) when both

change in discretionary accruals (ChDAC) and working capital accruals (ChWACC) serve as dependent

variables.

6. CONCLUSION

SEC rules 33-8128 and 33-8644 substantially reduce the 10-K filing period for large accelerated and

accelerated filers from 90 to 60 and 75 days, respectively, for fiscal years ending on or after December 15, 2006

(SEC, 2002; 2005). It is also possible that similar accelerations may be imposed on NAFs and foreign issuersin

the future. For many firms and their auditors, this regulation led to a mandatory reduction in audit delay or the

length of time from a companys fiscal year-end to the date of the auditors report. The SEC implemented

accelerated reporting deadlines largely over the objections of auditors and preparers and many suggest that the

SEC did not give sufficient consideration to the potential for negative consequences. It is possible that auditors

mitigated the negative effects of audit delay reductions by performing more interim procedures, relying more on

the internal controls/systems of their clients, and using advanced audit technology (cf., PCAOB, 2004; PCAOB,

2007; Brazel and Agoglia, 2007). Such tactics may have reduced the extent to which post-fiscal year-end audit

procedures/evidence were needed to provide an acceptable level of audit quality. Still, when mandated


29/44

27

reductions were large (e.g., 15 days), it became less likely that these tactics could overcome a substantial

reduction in post-fiscal year-end audit time.

We empirically assess the disputed effects of the rule by examining the contexts under which mandated

reductions in audit delay have been associated with changes in earnings quality. Since the rules affect filer types

differently, we also examine whether the relation between reductions in audit delay and changes in earnings

quality differs by filer type. Consistent with our hypothesis, we find that larger mandatory reductions in audit

delay are more negatively associated with changes in earnings quality than smaller reductions. The evidence

suggests an unintended consequence of the SECs two separate 15-day reductions in filing deadlines (i.e., lower

earnings quality). The relation between larger mandated reductions and earnings quality appears to also be more

acute for busy-season audits and AFs.

Overall, our findings support claims by auditors and preparers that the acceleration of 10-K filings has

the capacity to reduce the quality of financial information supplied to external users. However, these adverse

effects appear to depend on the context of the acceleration (size of audit delay reduction, fiscal year-end of the

firm, and filer type). We do provide some initial evidence regarding the potential negative effects of

accelerating deadlines for AFs and NAFs in the future. Our empirical evidence should also be of considerable

use when foreign regulators consider accelerating their filing processes in the future.

Given the substantial effect of this rule on the financial reporting and audit processes, and the current

dearth of research investigating this topic, we believe the results of this study will spur future research related to

the regulations impact on preparers, auditors, investors, and the capital markets. For example, future studies

could measure the increased relevance of accelerated filings and examine whether it possibly outweighs the

reductions in information quality described herein. Researchers could investigate how investors

perceive/balance the benefits of more timely financial information with the cost of potentially less reliable

financial statements. Finally, a fruitful area of research might be to examine how auditors can cope with

accelerated filings by making their audits more efficient, while not sacrificing their effectiveness.


30/44

28

References

Agoglia, C.P., Brazel, J.F., Hatfield, R.C., Jackson, S.B., 2010. How do Audit Workpaper Reviewers Cope withthe Conflicting Pressures of Detecting Misstatements and Balancing Client Workloads? Auditing: AJournal of Practice & Theory, (forthcoming).

American Institute of Certified Public Accountants (AICPA), 1983. Statement on Auditing Standards No. 45:Omnibus Statement on Auditing Standards-1983. New York, NY.

American Institute of Certified Public Accountants (AICPA), 2006a. Statement on Auditing Standards No. 107:Audit Risk and Materiality in Conducting an Audit. New York, NY.

American Institute of Certified Public Accountants (AICPA), 2006b. Statement on Auditing Standards No. 110:Performing Audit Procedures in Response to Assessed Risks and Evaluating the Audit Evidence ObtainedNew York, NY.

Antle, R., Nalebuff, B., 1991. Conservatism and Auditor-Client Negotiations. Journal of Accounting Research29, 31-54.

Arens, A.A., Elder, R.J., Beasley, M.S., 2010. Auditing and Assurance Services, an Integrated Approach. UpperSaddle River, NJ: Prentice Hall.

Ashbaugh, H., LaFond, R., Mayhew, B.W., 2003. Do Nonaudit Services Compromise Auditor Independence?Further Evidence. The Accounting Review 78, 611-639.

Ashton, R.H., Willingham, J.J., Elliott, R.K., 1987. An Empirical Analysis of Audit Delay. Journal ofAccounting Research 25, 275-292.

Ashton, R.H., Graul, P.R., Newton, J.D., 1989. Audit Delay and the Timeliness of Corporate Reporting.Contemporary Accounting Research 5, 657-673.

Ball, R., Shivakumar, L., 2005. Earnings Quality in UK Private Firms: Comparative Loss RecognitionTimeliness. Journal of Accounting and Economics 39, 83-128.

Ball, R., Shivakumar, L., 2006. The Role of Accruals in Asymmetrically Timely Gain and Loss Recognition.Journal of Accounting Research 44, 207-242.

Bamber, E.M., Bamber, L.S., Schoderbek, M.P., 1993. Audit Structure and Other Determinants of Audit ReportLag: An Empirical Analysis. Auditing: A Journal of Practice & Theory 12, 1-23.

Bartov, E., Givoly, D., Hayn, C., 2002. The Rewards to Meeting or Beating Earnings Expectations.

Journal of Accounting and Economics 33, 173204.

Bartov, E., Gul, F.A., Tsui, J.S.L., 2000. Discretionary-Acruals Models and Audit Qualification. Journal ofAccounting and Economics 30, 421-452.

Basilo, T. A., 2007. Reducing Sarbanes-Oxley Compliance Costs: Is the Top Down, Risk-Based AuditApproach a Solution, or a Mistake? The CPA Journal 77, 6-9.


31/44

29

Basu, S., 1997. The Conservatism Principle and Asymmetric Timeliness of Earnings. Journal of Accountingand Economics 24, 3-37.

Beasley, M., Carcello, J., Hermanson, D., 1999. Fraudulent Financial Reporting: 1987-1997:An Analysis of U.S. Public Companies. Committee of Sponsoring Organizations of the TreadwayCommission.

Beasley, M., Carcello, J., Hermanson, D., Neal, T., 2010. Fraudulent Financial Reporting: 1998-2007

An Analysis of U.S. Public Companies. Committee of Sponsoring Organizations of the TreadwayCommission.

Becker, C.L., DeFond, M.L., Jiambalvo, J., Subramanyam, K.R, 1998. The Effect of Audit Quality on EarningsManagement. Contemporary Accounting Research 15, 1-24.

Bonner, S.E., Palmrose, Z-V., Young, S.M., 1998. Fraud Type and Auditor Litigation: An Analysis of SECAccounting and Auditing Enforcement Releases. The Accounting Review 73, 503-532.

Brazel, J.F., Agoglia, C.P., 2007. An Examination of Auditor Planning Judgments in a Complex AccountingInformation System Environment. Contemporary Accounting Research 24, 1059-1083.

Brazel, J. F., Carpenter, T.D., Jenkins, J.G., 2010. Auditors Use of Brainstorming in the Consideration ofFraud: Reports from the Field. The Accounting Review 85, 1273-1301.

Bryant-Kutcher, L., Peng, E.Y., Zvinakis, K., 2010. The Impact of the Accelerated Filing Deadline onTimeliness of 10-K filings. Working Paper, Colorado State University.

Caramanis, C., Lennox, C., 2008. Audit Effort and Earnings Management. Journal of Accounting andEconomics 45, 116-138.

Caster, P., 1990. An Empirical Study of Accounts Receivable Confirmations as Audit Evidence. Auditing: A

Journal of Practice & Theory 9, 75-91.

Choi, J.-H., Kim, J.-B., Zang, Y., 2006. The Association between Audit Quality and Abnormal Audit Fees.Working Paper, Seoul National University.

Coram, P., Ng, J., Woodliff, D.R., 2004. The Effect of Misstatement on the Propensity to Commit ReducedAudit Quality Acts under Time Budget Pressure. Auditing: A Journal of Practice & Theory 23, 159-167.

Das, S., Shroff, P.K., Zhang, H., 2009. Quarterly Earnings Patterns and Earnings Management. ContemporaryAccounting Research 26, 797-831.

Dechow, P.M., Dichev, I.D., 2002. The Quality of Accruals and Earnings: The Role of Accrual EstimationErrors. The Accounting Review 77, 35-59.

Dechow, P.M., Skinner, D.J., 2000. Earnings Management: Reconciling the Views of AccountingAcademics, Practitioners, and Regulators. Accounting Horizons 14, 235-250.

Dechow, P.M., Sloan, R.G., Sweeney, A.P., 1995. Detecting Earnings Management. TheAccounting Review 70, 193-225.


32/44

30

DeFond, M.L., Subramanyam, K.R., 1998. Auditor Changes and Discretionary Accruals. Journal of Accountingand Economics 25, 35-67.

DeZoort, F.T., Lord, A.T., 1997. A Review and Synthesis of Pressure Effects Research in Accounting. Journalof Accounting Literature 16, 28-85.

Ettredge, M.L., Li, C., Sun, L., 2006. The Impact of SOX Section 404 Internal Control Quality Assessment onAudit Delay in the SOX Era. Auditing: A Journal of Practice & Theory 25, 1-23.

Frankel, R.M., Johnson, M.F., Nelson, K.K., 2002. The Relationship between Auditors Fees for NonauditServices and Earnings Management. The Accounting Review 77 (Supplement), 71-105.

Francis, J.R., Krishnan, J., 1999. Accounting Accruals and Auditor Reporting Conservatism. ContemporaryAccounting Research 16, 135-165.

Fudenberg, K., Tirole, J., 1995. A Theory of Income and Dividend Smoothing Based on Incumbency Rents.Journal of Political Economy 103, 75-93.

Givoly, D., Ronen, J., 1981. Smoothing Manifestations in Fourth Quarter Results of Operations: Some

empirical evidence. ABACUS 17, 17493.

Gopez, E. C., 1954. Auditing with Accent on the Income Statement. The Accounting Review 29, 571-574.

Graham, J.R., Harvey, C.R., Rajgopal, S., 2006. Value Destruction and Financial Reporting Decisions.Financial Analysts Journal 62, 27-39.

Gullapalli, D., 2005a. Take this Job and File it; Burdened by Extra Work Created by the Sarbanes-Oxley Act,CPAs Leave the Big Four for Better Life. The Wall Street Journal, New York, May 4, pg C1.

Gullapalli, D., 2005b. To Err is Human, to Restate Financials, Divine: Companies Redo Reports in Record

Numbers, Partly due to Sarbanes-Oxley. The Wall Street Journal, New York, January 20, pg C3.

Hoffman, V.B., Zimbelman, M.F., 2009. Do Strategic Reasoning and Brainstorming Help Auditors ChangeTheir Standard Audit Procedures in Response to Fraud Risk? The Accounting Review 84, 811-837.

Hribar, P., Nichols, D.C., 2007. The Use of Unsigned Earnings Quality Measures in Tests of EarningsManagement. Journal of Accounting Research 45, 1017-1053.

Jones, J. J., 1991. Earnings Management During Import Relief Investigations. Journal of Accounting Research29, 193-228.

Jones, K., Krishnan, G., Melendrez, K., 2008. Do Models of Discretionary Accruals Detect Actual Cases ofFraudulent and Restated Earnings? Contemporary Accounting Research 25, 499-531.

Kasznik, R., McNichols, M.F., 2002. Does Meeting Earnings Expectations Matter? Evidence fromAnalyst Forecast Revisions and Share Prices. Journal of Accounting Research 40, 727759.

Kinney, W.R., Martin, R.D., 1994. Does Auditing Reduce Bias in Financial Reporting? A Review of


33/44

31

Audit Related Adjustment Studies. Auditing: A Journal of Practice and Theory 13, 149-156.

Kim, J., Chung, R., Firth, M., 2003. Auditor Conservatism, Asymmetric Monitoring, and EarningsManagement. Contemporary Accounting Research 20, 323-359.

Knechel, W.R., Payne, J.L., 2001. Additional Evidence on Audit Report Lag. Auditing: A Journal of Practice &Theory 20, 137-146.

Kothari, S.P., Leone, A., Wasley, C., 2005. Performance Matched Discretionary Accrual Measures. Journal ofAccounting and Economics 39, 163-167.

Krishnan, G.V., 2003. Audit Quality and the Pricing of Discretionary Accruals. Auditing: A Journal of Practice& Theory 22, 109-126.

Larcker, D.F., Richardson, S.A., 2004. Fees Paid to Audit Firms, Accrual Choices, and Corporate Governance.Journal of Accounting Research 42, 625-658.

Lopez-Acevedo, D.M., 2007. The Effect of Workload Compression on Audit Quality. Working Paper,University of Texas at San Antonio.

McDaniel, L.S., 1990. The Effects of Time Pressure and Audit Program Structure on Audit Performance.Journal of Accounting Research 28, 267-285.

McGee, S., 2005. CPA Recruitment Intensifies as Accounting Rules Evolve. The Wall Street Journal. NewYork, March 22, pg. B-6.

Myers, J.N., Myers, L.A., Omer, T.C., 2003. Exploring the Term of the Auditor-Client Relationship and theQuality of Earnings: A Case for Mandatory Auditor Rotation? The Accounting Review 78, 779-799.

Palmrose, Z-V., Scholz, S.W., 2004. The Circumstances and Legal Consequences of Non-GAAP Reporting:

Evidence from Restatements. Contemporary Accounting Research 21, 139-180.

Peterson, M., 2009. Estimating Standard Errors in Finance Panel Data Sets: Comparing Approaches.The Review of Financial Studies 22, 435-480.

Public Company Accounting Oversight Board (PCAOB), 2004. Auditing Standard No. 2, An Audit of InternalControl over Financial Reporting in Conjunction with an Audit of Financial Statements. Washington, DC.

Public Company Accounting Oversight Board (PCAOB), 2007. Auditing Standard No. 5, An Audit of InternalControl over Financial Reporting that is Integrated with an Audit of Financial Statements. Washington,DC.

Rose, J., 2007. Recruiters Take Hip Path to Fill Accounting Jobs. The Wall Street Journal, New York,September 18, pg B8.

Sanchez, M.H., Agoglia, C.P., Hatfield, R.C., 2007. The Effect of Auditors Use of a Reciprocity-BasedStrategy on Auditor Client Negotiations. The Accounting Review 82, 241-263.

Schwartz, K.B., Soo, B.S., 1996. The Association between Auditor Changes and Reporting Lags. ContemporaryAccounting Research 13, 353-370.


34/44

32

Securities and Exchange Commission (SEC), 2002. Release No. 33-8128 Acceleration of Periodic Report FilingDates and Disclosure Concerning Website Access to Reports.

Secur

tamara a. lambert

Documents