Consequences of changing the auditor’s report: Evidence from the U.K.

Elizabeth Gutierrez
Universidad de Chile

Miguel Minutti-Meza*
University of Miami

Kay W. Tatum
University of Miami

Maria Vulcheva
Florida International University

March 1, 2016

Abstract: The audit regulator in the United Kingdom made significant changes to the auditor’s report
for large public companies with fiscal years beginning on or after October 1, 2012. The new report has to
describe significant risks of material misstatement, the application of materiality, and the audit’s scope.
We study changes in audit fees, audit quality, and investors’ reaction to the report’s public release in the
two years before and after the new rules. We find mixed evidence of a change in audit fees, ranging from
an increase of nearly four percent to no change, depending on model specification. We do not find
evidence of changes in audit quality or investors’ reaction to the report’s public release. Further
examining the content of the new reports, we document that the total length of the report, the length of the
risks’ discussion, and the number of risks are positively associated with audit fees. Moreover, materiality
is inversely related to audit quality. Thus, although the new report format had a small impact on audit fees
and did not significantly change audit quality or investors’ reaction to the report, the mandated risk and
materiality disclosures are associated with cross-sectional variation in audit fees and quality.

The authors thank Keith Czerney, Pietro Bianchi, Like Jiang (discussant), Jake Krupa, Stephania Mason
(discussant), Robert Pawlewicz, Matthew Phillips, Grace Pownall, Sundaresh Ramnath, Joe Schroeder, Michael
Willenborg, and seminar participants at AAA Auditing Section Midyear Meeting 2016, AAA International
Accounting Section Midyear Meeting 2016, Florida International University, the International Symposium on Audit
Research (ISAR) 2015, Ohio State University, Public Company Accounting Oversight Board (PCAOB), and
Singapore Management University for helpful comments and suggestions. All errors are our own. *Corresponding
author: [email protected].

Electronic copy available at: http://ssrn.com/abstract=2741174

 Consequences of changing the auditor’s report: Evidence from the U.K.

1. INTRODUCTION

Audit regulators worldwide are taking steps to enhance the content of the auditor’s report,

including the Financial Reporting Council (FRC) in the United Kingdom (U.K.) and Ireland, the

International Auditing and Assurance Standards Board (IAASB) across countries, and the Public

Company Accounting Oversight Board (PCAOB) in the U.S. In most countries, including the

United States (U.S.), the auditor’s report uses standardized language for the vast majority of

companies. Most importantly, the auditor’s opinion has a “pass or fail” nature regarding

compliance with accounting principles. Arguably, the auditor’s report provides financial

statement users with little company-specific information about the auditor’s work.

The FRC has promulgated changes issuing International Standard on Auditing (ISA) 700

(U.K. and Ireland, revised June 2013), “The Independent Auditor’s Report on Financial

Statements” (FRC 2013). This standard mandates significant modifications to the auditor’s

report for companies with a premium listing of equity shares on the London Stock Exchange

(LSE) main market with fiscal years beginning on or after October 1, 2012 (i.e., fiscal year-end

on or after September 2013).1 The new report is significantly longer than the previous

standardized report. The auditor must describe the most significant risks of material

misstatement, the application of materiality, and the audit’s scope.2

Concurrently with the FRC’s new standard, the IAASB and PCAOB issued proposals for

new and revised auditor reporting standards (IAASB 2013; PCAOB 2013). A significant new

1
The main market is the LSE’s principal market for listed companies from the U.K. and overseas. There are three
listing categories in the main market: premium, standard, and high growth. Premium listing is only open to equity
shares issued by trading companies and closed and open-end investment entities. Premium listed companies must
comply with the U.K.’s highest standards of regulation and corporate governance.
http://www.londonstockexchange.com/companies-and-advisors/main-market/main/market.htm. In our analyses we
focus on premium listed non-financial companies incorporated in Great Britain.
2
In Appendix A we provide an example of a U.K. auditor’s report prepared in 2013 in accordance with the new
rules.

Electronic copy available at: http://ssrn.com/abstract=2741174

requirement in both the IAASB and PCAOB standards is to communicate key or critical audit

matters (KAMs or CAMs, respectively), similar to the U.K. requirements regarding significant

risks of material misstatements.3 The IAASB issued a suite of auditor reporting standards in

January 2015, including ISA 701 “Communicating Key Audit Matters in the Independent

Auditor’s Report” (IAASB 2015c). The proposal to revise the auditor report model is on the

PCAOB’s standard-setting agenda for the second quarter of 2016.4

In this study we examine the consequences of the new auditor’s report in the U.K., in

terms of audit costs, audit quality, and investors’ reaction to the report’s filing. We take

advantage of the cut-off date for the adoption of the new report’s requirements, applicable to

companies with fiscal years starting on or after October 1, 2012. In our first set of analyses, we

examine data from two years before and after the cut-off date, and test for changes in audit fees,

as a proxy for audit costs; in absolute discretionary accruals, as a proxy for audit quality; and, in

three-day cumulative abnormal returns and trading volume around the public dissemination of

the annual report, which includes the auditor’s report, as proxies for investors’ reaction to the

report’s content.5

The regulatory cut-off date effectively results in having a treatment group of new report

adopters and a control group of non-adopters immediately before and after the September 2013

implementation date. In our second set of analyses, we examine the one-year pre-post adoption

changes for companies with fiscal year-end between September 2013 and February 2014 (i.e.,

adopters of the new report format starting in 2013), using as a benchmark the group of

3
The PCAOB proposal also contains other changes, including enhancing the report’s writing and adding statements
about auditor independence, auditor tenure, and the auditor’s responsibility for, and evaluation of, other information.
4
http://pcaobus.org/Standards/Documents/201512-standard-setting-agenda.pdf
5
In additional analyses we examine the incidence of small earnings and total accruals as alternative proxies for audit
quality; and, absolute three-day abnormal returns and three-day trading spread around the report’s filing date as
alternative proxies for investors’ reaction. We do not find evidence that the new report resulted in changes in the
incidence of small earnings, total accruals, absolute abnormal returns, or trading spread.

Electronic copy available at: http://ssrn.com/abstract=2741174

companies with fiscal year-end between March and August 2013 (i.e., non-adopters of the new

report format in 2013). The control group, consisting of companies subject to the same listing

requirements, but fiscal year-end immediately before the September 2013 cut-off date, helps to

isolate the effects of the new report from time trends and other potential confounding factors

(i.e., short-term difference-in-differences analysis).6

We document mixed evidence regarding a change in audit fees. Examining two years

before and after the new rules, the new report is associated with an increment in audit fees of

approximately four percent. This estimate is reduced to approximately three percent after

including company fixed effects in our pre-post audit fee model. However, comparing

companies reporting six months before and after the regulatory cut-off, as well as a year prior,

we do not find evidence of a difference-in-differences effect of the new report on audit fees.

Furthermore, we do not find evidence that the new report had an overall impact on audit quality

or investors’ reaction. This lack of findings is consistent with several FRC’s statements,

including ISA 700 (U.K. and Ireland, revised June 2013), which explicitly mention that the intent

of the new requirements is to make public information previously discussed with the audit

committee and that meeting these requirements is not likely to be very costly (FRC 2013a,

2013b, 2013c).

We also conduct cross-sectional analyses pertaining to the two-year period following the

adoption of the new rules. We examine the association between audit fees, audit quality, and

investors’ reaction and several quantifiable elements of the new reports, including the total

length of the report, the length of the risks’ discussion, the number of risks, and the materiality

6
In our analyses we use all non-financial companies in the LSE Main Market Premium that meet the cut-off criteria,
with available financial statement data in Thomson Reuters Worldscope and hand-collected data from annual
reports. Our sample selection criteria are discussed in more detail in Section 4.

3
threshold.7 Under the new rules, the length of the report has tripled, increasing by approximately

two pages. The section on risks of material misstatement discloses on average four risks. The

average (median) materiality is 0.6 (0.5) percent of total assets.8 We document that the length of

the report, the length of the risks’ discussion, and the number of risks are all positively related to

audit fees, suggesting an association between auditor effort or risk premium and the auditor’s

assessment and disclosure of important audit issues. We also find that materiality is positively

related to absolute discretionary accruals, suggesting that materiality is inversely related to audit

quality. However, we do not find evidence of a consistent association between the report’s

characteristics and either three-day abnormal returns or abnormal trading volume.

Our combined findings suggest that the risk and materiality disclosures in the report are

predictably related to cross-sectional variation in audit fees and quality. Arguably, if the risks

and materiality disclosures were strictly boilerplate, they would not be associated with audit fees

and quality. Nevertheless, comparing the two years before and after the adoption of the new

rules, the new report had a relatively small impact on audit fees and did not significantly change

audit quality or investors’ reaction around the filing date.

Our study adds to the mixed evidence provided by other contemporaneous papers

examining the effect of the new U.K. audit regulation (Lennox et al., 2015; Reid et al., 2015a,

2015b). On one hand, Lennox et al. (2015) focus on changes in firm value after the adoption of

the new rules, as well as on the incremental impact of the risk disclosures on firm value. They

7
We do not explicitly assess the effect of scope disclosures, beyond examining the effects of the total length of the
auditor’s report, because the scope disclosures are difficult to quantify. In many cases the scope disclosures discuss
the group structure of the audit without providing additional details. See for example, the scope section of the report
included in Appendix A.
8
We scale the materiality amount in British Pounds by total assets in order to use the same denominator for all
companies. However, our results do not depend on scaling by total assets. We also use the natural logarithm of the
materiality amount in our main analyses.

4
document that the risk disclosures are not incrementally informative to investors.9 Furthermore,

they do not find evidence of either a change in market reaction in the short window surrounding

the report’s filing, or a significant change in audit fees after the adoption of the new rules. On the

other hand, Reid et al. (2015, 2015b) find a marginal increase in audit fees, a decrease of 0.019

in absolute abnormal accruals, a decrease of 24 percent in the odds of meeting or beating

analysts’ forecasts, and an increase of approximately 12.7 percent in abnormal trading volume

around the filing of annual reports after the adoption of the new rules. While closely related to

these studies, our study differs in three aspects. First, we examine a longer sample period that

spans four years around the adoption of the new rules. Our sample allows us to investigate

changes beyond the first year after the adoption of the new rules. Second, we use the September

2013 regulatory cut-off date to implement difference-in-differences analyses, where the control

group is comprised of companies listed in the same market, with annual reports in the months

just preceding the regulatory cut-off.10 Third, we perform a series of cross-sectional tests

examining the association between the report’s content and audit fees, audit quality, and

investor’s reaction.

Our findings can be of interest to regulators and other parties considering changes to the

auditor’s report. The U.K. is a large and developed market and our findings may be useful in

assessing similar changes to the auditor’s report in other similar jurisdictions. With regard to the

U.K., the findings of our study may help regulators and other interested parties, aiming to

enhance the usefulness of the currently mandated new report’s disclosure in the U.K. (FRC

9
Lennox et al. (2015) examine the effect of risk disclosures as input of long-window equity valuation models.
10
Both Lennox et al. (2015) and Reid et al. (2015a; 2015b) use companies listed in the LSE’s AIM market as a
control group. AIM is an alternative international market for comparatively smaller growing companies. There are
major differences between premium and AIM listed companies. A premium listing entails comparatively more
rigorous standards and compliance with the U.K. Corporate Governance Code. Reid et al. (2015a; 2015b) also use
companies listed in the U.S. New York Stock Exchange (NYSE) as an alternative control group. There are
considerable differences between the NYSE and LSE premium listing requirements, as well as different litigation
environments in the U.S. and U.K.

5
2015).11 This study contributes to the literature on auditor’s reports and their association with the

cost of audits, audit quality, and investors’ reaction (Bedard et al. 2015; Butler et al., 2004;

Carcello and Li, 2013; Czerney et al., 2014a, 2014b; Hay et al., 2006; Menon and Williams,

2010). We also contribute to the literature on risk disclosures in financial reports (Campbell et

al., 2014; Kravet and Muslu, 2013). Finally, this study extends the descriptive evidence provided

by financial analysts and auditors on the consequences of the auditor’s report change in the U.K.

(Citi, 2014, 2015; KPMG 2014; PwC 2015).

2. BACKGROUND AND HYPOTHESIS DEVELOPMENT

Effective changes to the auditor’s report in the U.K. in 2013

Currently, in most countries, including the U.S., the auditor’s report uses standardized

language for a vast majority of companies. However, investors and regulators have expressed

growing concern about the usefulness of the auditors’ reports. Arguably, the pass-fail nature of

auditors’ reports may not give financial statement users enough relevant information. For

example, “until now, auditors’ letters have been among the least interesting part of annual

reports. If the opinions said the accounting was proper – and virtually all did – and did not voice

concern whether the company could stay in business, the letters were basically the same. There

was no reason for an investor to read them” (Norris, 2014).

Audit regulators have taken steps to change the auditor’s report. The U.K. FRC has

moved forward by issuing ISA 700 (U.K. and Ireland, revised June 2013), “The Independent

Auditor’s Report on Financial Statements” (FRC, 2013a). This new auditing standard mandates

significant changes to the auditor’s report for most large companies listed on the LSE with fiscal

11
Reviews of the first two years of reports conducted by the FRC and involving discussions with investors and
auditors suggest that there is room for improvement, for example by increasing the granularity of risk reporting and
providing more explanations about the materiality, scope, and risk disclosures (FRC, 2015, 2016).

6
years beginning on or after October 1, 2012.12 The new standard (paragraph 19A of ISA 700)

requires auditors to: (1) describe the most significant risks of material misstatement; (2) disclose

how the auditor applies the concept of materiality, including a materiality threshold for the

statements as a whole and performance materiality; and, (3) explain the scope of the audit.

The new FRC standard complements changes made to the U.K. Corporate Governance

Code in October 2012 (FRC, 2012a, 2012b). The Code requires additional specific disclosures

for audit committees, including (1) a description of the significant issues that the audit committee

considered in relation to the financial statements, and how these issues were addressed; (2) an

explanation of how the audit committee has assessed the effectiveness of the external audit

process and the approach taken to the appointment or reappointment of the external auditor; (3)

information about the length of tenure of the current audit firm and when a tender was last

conducted; and, (4) if the external auditor provides non-audit services, an explanation of how the

audit committee ensures that auditor objectivity and independence are safeguarded (paragraph

C.3.8 of the U.K. Corporate Governance Code FRC 2014).

Changes to the auditor’s report adopted by the IAASB and considered by the PCAOB

A significant proposed change by both the IAASB and the PCAOB is the requirement to

communicate key (critical) audit matters (KAMs or CAMs), which are similar to the U.K. new

disclosure requirements regarding significant risks of material misstatements.

In July 2013 the IAASB issued a proposal for revising the auditor’s report (IAASB

2013). Following the proposed changes, in January 2015 the IAASB issued revised standard ISA

700 “Forming an Opinion on Financial Statements” and new standard ISA 701 “Communicating

12
In the U.K. the new rules apply to companies with a Premium listing of equity shares regardless of whether they
are incorporated in the U.K. or elsewhere. In Ireland, these include Irish incorporated companies with a primary or
secondary listing of equity shares on the Irish Stock Exchange.

7
Key Audit Matters in the Independent Auditor’s Report” (IAASB 2015).13 The key changes are:

(1) inclusion and discussion of KAMs; (2) declaration of auditor independence and compliance

with ethical standards; (3) signature of the engagement partner with a “harm’s way exemption”;

(4) prescribed order of the presentation of the auditor’s report; (5) emphasis on the audit opinion

and other firm-specific information; (6) auditor assessment of the appropriateness of the

management’s going concern assumption and review of circumstances that might endanger this

assumption; and, (7) improved disclosure of auditor responsibilities.

Furthermore, the PCAOB issued a proposed auditing standard on the auditor’s report in

August 2013 (PCAOB 2013). The proposed standard retains the basic elements from the existing

standard, including the requirement to express an opinion about whether or not the financial

statements are fairly presented in accordance with GAAP (i.e., “pass or fail” opinion). However,

the proposed standard makes significant changes to the existing report. It requires the auditor to

communicate CAMs, enhances certain language in the current report, and adds new elements.

The new elements are a statement about auditor independence, a statement about the auditor’s

responsibility for other information, and auditor tenure. The proposal to revise the auditor report

model is on the PCAOB’s standard-setting agenda for the second quarter of 2016

Nevertheless, we note that the FRC, IAASB, and PCAOB new and proposed

requirements regarding significant risks of material misstatements, KAMs, and CAMS do not

have identical definitions (EY 2014). The significant risks of material misstatements definition

refers to risks that had the greatest effect on audit strategy, allocation of resources, and directing

efforts of the audit team (FRC 2013). The KAM’s definition refers to areas of most significance

in the audit, determined from matters discussed with those charged with governance, that

13
The standards will be effective for audits of financial statements for periods ending on or after December 15,
2016. Early adoption is permitted.

8
required auditor attention (IAASB 2015b). The CAM’s definition refers to areas of most

significant auditor difficulty, including those that require significant auditor judgment (PCAOB

2013).

Effect of changes to the auditor’s report on audit fees, audit quality, and investors’ reaction

First, we are interested in determining whether the regulatory changes had an overall

impact on the cost of audits. Prior research has found positive association between audit fees and

client size, risk, complexity, and auditor litigation risk (Carcello and Li 2013; Hay et al., 2006;

Seetharaman et al., 2002). Therefore, a potential change in auditor effort and risk premium,

resulting from increased disclosures in the auditor’s report under the new regime, can impact the

cost of audits. Thus, we formulate the following alternative hypothesis:

H1: All else equal, the adoption of the new auditor’s report is associated with an

increase in audit fees.

We are also interested in determining whether the regulatory changes had an overall

impact on audit quality. The literature on audit quality has generally demonstrated that

reputation, litigation, and regulatory concerns shape the incentives that drive auditors to supply

audit quality (DeFond and Zhang, 2014). Additional disclosures in the auditor’s report might put

companies and their auditors under more careful scrutiny by users, especially when it comes to

risk disclosures, and thus increase audit quality. Therefore, to test the effect of the new regulation

on audit quality, we formulate the following alternative hypothesis:

H2: All else equal, the adoption of the new auditor’s report is associated with an

increase in audit quality.

Finally, we examine whether the regulatory changes are associated with investors’

reaction to the filing of the auditor’s report. The new auditor’s report disclosures on the

9
significant risks of material misstatement, the application of materiality, and the scope of the

audit were not available to investors before the regulatory changes. These disclosures might

provide new information to the market and motivate a revision in the aggregate investors’ beliefs

about the value of the firm (i.e., price reaction), and a revision of differential pre-announcement

beliefs and announcement interpretations for individual investors (i.e., trading volume reaction)

(Bamber et al., 2011).14

Previous studies of the auditor’s report in the U.S. demonstrate that non-standard

modifications, such as going concern opinions, have information content when they are not

expected (Menon and Williams 2010; Taffler et al., 2004). More generally, Czerney et al.

(2014a) document investors’ reaction to various forms of explanatory language in the auditor’s

report (e.g., adoption of new accounting standards, division of auditor responsibility, previous

restatements, and emphasis of matter paragraphs) and suggest that the report’s language may

have an attention directing effect. To investigate whether the new format of the auditor’s report

has incremental information content, we propose the following alternative hypothesis:

H3: All else equal, the adoption of the new auditor’s report is associated with investors’

reaction to the filing of annual reports.

However, there are three broad reasons why the changes to the report may not have an

observable effect on audit fees, audit quality, and investors’ reaction in the years immediately

following the adoption of the new rules.

First, regulators did not expect a large revision to the auditors’ methodologies, but instead

aimed to make public several matters already discussed privately between auditors and audit

committees. Several FRC statements explicitly mention that the new report’s requirements intend

14
We do not make prediction about the direction of price reaction, because we expect this direction to be determined
based on whether the investors perceive the new disclosures as good or bad news.

10
to make public information previously discussed with the audit committee and that meeting these

requirements is not likely to be very costly. The report’s standard ISA (U.K. and Ireland) 700

(FRC, 2013) mentions in paragraph A13A “assessed risks of material misstatement are likely to

have been identified by the auditor in meeting the requirements of ISA (U.K. and Ireland) 315.”

A Consultation Paper (FRC, 2013c, paragraph 23) highlights that “investors did not think that

cost would be a significant issue as the costs would largely be related to the additional

consideration of public reporting of matters likely to have been already fully addressed in the

audit and reported to the audit committee.” A Feedback Statement (FRC, 2013b, paragraph 6.1),

dealing with implementation issues, notes that “as the auditor is already required to disclose the

information to be included in the audit report to the audit committee the work effort in preparing

it and familiarizing the audit committee with it should be minimal.” Thus, audit quality may

already have accounted for significant risks, even if the auditor or the client did not publicly

disclose such risks (i.e., on average, risks were adequately addressed and mitigated before the

new rules). Increased disclosure on what the auditors already do may not necessarily increase the

auditors’ litigation risk and auditors may not be able to significantly increase fees for additional

hours spent on documenting and writing the new reports.

Second, the additional information may not be strictly new to investors or significant enough

to change the average perceptions about a company’s financial reporting quality. Disclosing

uncertainty regarding revenue recognition, tax accounting, valuation, etc. may already be

expected by investors given other information sources (e.g., management discussions, financial

analyst reports, audit committee reports) and investors may already infer the risk profile of a

company from other observable characteristics (size, presence of intangible assets including

11
goodwill, etc.).15

Third, the new rules could have a long-term impact not captured by our sample and proxies

for the cost of audits, audit quality, and investors’ reaction. According to FRC (2016, p. 4),

although investors have welcomed extended auditor reporting, and greatly value the enhanced

information it provides, they feel that more could still be done to enhance the reports. Investors

noted that they want more granular descriptions of risks, more transparency about assumptions

made by management and benchmarks used by auditors, and greater precision. For example,

“words like ‘significant’ are not necessarily sufficiently informative” (FRC 2016, p.21).

3. RESEARCH DESIGN AND MODEL SPECIFICATION

Pre-post adoption and difference-in-differences research design

We take advantage of the cut-off date for the adoption of the new auditor’s report

requirements, applicable to LSE main market premium listed companies with fiscal years

starting on or after October 1, 2012 (i.e., year-ends on or after September 2013). In our first set

of analyses, using data from two years before and two years after the cut-off date (see Figure 1-

A), we examine pre-post adoption changes in audit fees as a proxy for the cost of audits; in

absolute discretionary accruals as a proxy for audit quality; and, in three-day cumulative

abnormal trading volume and abnormal returns around the public dissemination of the annual

report, which includes the auditor’s report, as proxies for investors’ reaction to the auditor’s

report.

The regulatory cut-off date effectively results in having a treatment group of new report

adopters and a control group of non-adopters immediately before and after the September 2013

implementation date. In a second set of analyses, we examine the one-year pre-post adoption

15
In private discussions with personnel from a Big-4 firm with experience with the report changes we confirmed
that in the short term there were no fundamental changes to their methodology and that the ability to increase fees as
a result of the new changes was limited.

12
changes for companies with fiscal year-end between September 2013 and February 2014 (i.e.,

adopters of the new report format), including as a benchmark the one-year pre-post changes for

companies with fiscal year-end between March and August 2013 (i.e., non-adopters of the new

report format, see Figure 1-B). Although the sample in the difference-in-difference analysis is

relatively small, we aim to capture short-term changes in outcomes around the adoption of the

new rules, while controlling for other confounding factors.

Next, we conduct additional cross-sectional analyses in the two years following the new

report rules (see Figure 1-C). We examine the association between the auditor’s report content

and audit fees, audit quality, and investors’ reaction. We start by examining the extent of the new

disclosures by focusing on the total length of the auditor’s report. Later on, we focus on

quantifiable elements of the new report, the risks’ discussion and the materiality threshold.

Audit fee analyses

Our measure of audit fees is the logarithm of total fees paid to the auditor for audit

services in each year. We manually obtain audit fees from the notes to the financial statements in

the annual report. To examine the effect of the new report on audit fees, we estimate the

following pre-post model, using all firm-year observations with available data two years before

and after the regulatory cut-off date in September 2013:

AFEESi,t = β0 + β1POSTi,t + β2SIZEi,t + β3ROAi,t + β4LEVi,t + β5CURRi,t

+ β6NSEGi,t + β7FORSALESi,t + β8BIG4i,t + β9CHAUDi,t + β10GCi,t
+ β11DECYEi,t + β12USLISTi,t + IndustryFE + εit, (1a)

where for company i in year t, POSTi,t is a time indicator variable, equal to one for the two years

following the new report requirements (starting in September 2013), and zero otherwise,

capturing the overall pre-post adoption effect of the new report; SIZEi,t is the natural logarithm of

13
total assets; ROAi,t is net income before extraordinary items (NIBE) divided by total assets;

LEVi,t is long-term debt divided by total assets; FORSALESi,t is U.K. sales divided by total sales;

CURRi,t is current assets divided by total assets; NSEGi,t is the natural logarithm of the number of

business segments; USLISTi,t is equal to one if the company is cross-listed on the NYSE, AMEX,

or NASDAQ stock exchanges, and zero otherwise;16 CHAUDi,t is equal to one if the company

has an audit firm change in the current fiscal year, and zero otherwise; GCi,t is equal to one if the

auditor’s report includes a going concern qualification, and zero otherwise; BIG4i,t is equal to

one if the company is audited by a Big-4 auditor, and zero otherwise; and, IndustryFE are 1-digit

SIC Codes industry fixed effects.17 The control variables aim to isolate the effects of client risk,

size, and complexity, and have been commonly used in previous audit fee research (Hay et al.,

2006).

For our difference-in-differences analyses we estimate the following model, using firm-

year observations with fiscal year-end six months before (March to August 2013) and after

(September 2013 to February 2014) the regulatory cut-off date in September 2013, as well as the

same companies in the prior year:

AFEESi,t = β0 + β1ADOPTERi,t + β2YR2013i,t + β3ADOPTER*YR2013i,t

+ β4SIZEi,t + β5ROAi,t + β6LEVi,t + β7CURRi,t + β8NSEGi,t
+ β9FORSALESi,t + β10BIG4i,t + β11CHAUDi,t + β12GCi,t
+ β13DECYEi,t + β14USLISTi,t + IndustryFE + εit, (1b)

where for company i in year t, ADOPTERi,t is an indicator variable for the September cut-off date

of the new report requirements, equal to one for companies with fiscal year-ends in the six

16
Including over-the-counter (OTC) listings in the USLISTi,t definition does not change our results.
17
All continuous variables in the regression models, except for the investors’ reaction dependent variables
(abnormal return and volume), are winsorized at the one and 99 percent levels to mitigate the presence of outliers.
Our results are similar if we winsorize or truncate the investors’ reaction dependent variables (abnormal return and
volume) at the one and 99 percent levels. Standard errors are clustered by company.

14
months from September to February, and zero in the six months from March to August;

YR2013i,t is a time indicator variable, equal to one for companies with fiscal year-ends in the six

months around the adoption date in September 2013 (March 2013 to February 2014), and zero

for the same companies in the year before (March 2012 to February 2013); ADOPTER*YEARi,t is

the interaction between ADOPTERi,t and YR2013i,t capturing the incremental one-year difference

for adopters of the new report format, compared to the one-year difference for non-adopters of

the new report format (i.e., short-term difference-in-differences estimator); and, all other

variables are as described above. We also include company fixed effects in models 1a and 1b and

report the results in Section 6.

Finally, for our cross-sectional analyses of the association between audit fees and report

characteristics we estimate the following model, using firm-year observations for the two years

after the regulatory cut-off date in September 2013:

AFEESi,t = β0 + β1REPORT_DISCLi,t + β2SIZEi,t + β3ROAi,t + β4LEVi,t

+ β5CURRi,t + β6NSEGi,t + β7FORSALESi,t + β8BIG4i,t + β9CHAUDi,t
+ β10GCi,t + β11DECYEi,t + β12USLISTi,t + IndustryFE + εit, (1c)

where for company i in year t, REPORT_DISCLi,t is one of the following independent variables:

TOTAL_NWRDSi,t, the natural logarithm of the total number of words in the auditor’s report;

RISK_NWRDSi,t, the natural logarithm of the total number of words in the risks section of the

auditor’s report; NRISKSi,t, the number of risks disclosed in the auditor’s report; PERCMATi,t, the

materiality amount in British Pounds divided by total assets; or, LOGMATi,t, the natural

logarithm of the materiality amount.18 All other variables are as defined above.

18
For example, in the report in Appendix A there are two risks disclosed by the auditor (1) Fraud in revenue
recognition, and (2) Risk of management override of internal controls. The number of words in the risks section is

15
Audit quality analyses

Our proxy for audit quality in our main analyses is the absolute value of discretionary

accruals. This measure arguably captures audit quality with respect to financial reporting quality,

where higher audit quality is defined as greater assurance that the financial statements faithfully

reflect the firm’s underlying economics conditioned on the financial reporting system and its

innate characteristics (DeFond and Zhang, 2014). Moreover, this measure has been extensively

used in prior research on audit quality (e.g., Carcello and Li, 2013; Lawrence et al., 2011;

Reichelt and Wang, 2010). To examine the effect of the new report on audit quality, we estimate

the following pre-post model:

DACCRi,t = β0 + β1POSTi,t + β2SIZEi,t + β3ROAi,t + β4LEVi,t + β5BIG4i,t

+ β5USLISTi,t + IndustryFE + εit, (2a)

where for company i in year t, DACCRi,t is absolute discretionary accruals, estimated using the

Jones model including ROA, scaling all variables by average assets.19 The control variables aim

to isolate the effects of client characteristics and have been used in previous research on audit

quality (e.g., Lawrence et al., 2011). All other variables are as defined above.

For our difference-in-differences analyses, using a similar research design as in our audit

fee analyses, we estimate the following model:

the word count of the paragraphs under each risk heading, 92 for the first risk and 90 for the second risk, total 182.
The disclosed total materiality is £1.3 million.
19
Discretionary accruals are estimated using an annual cross-sectional model for each SIC-2 digit industry (Kothari
et al., 2005) for all the companies on the London Stock Exchange. In order to obtain these companies, we use the
WSCOPEUK list from Worldscope for securities on LSE, that are major and primary (EXDSCD = ‘LN’ MAJOR =
‘Y’ and ISINID = ‘P’) and that are not Investment Trusts or Closed-End Funds (TYPE = ‘EQ’). Discretionary
accruals are the residual of the following model:
𝑇𝐴𝐶𝐶𝑅!,! = 𝛼 + 𝛽! 𝑅𝑂𝐴!,! + 𝛽! ∆𝑆𝑎𝑙𝑒𝑠!,! + 𝛽! 𝑃𝑃𝐸!,! + 𝜀
where for firm 𝑖 and year 𝑡, 𝑇𝐴𝐶𝐶𝑅!,! is (income before extraordinary items - cash flow from operations)/average
total assets; 𝑅𝑂𝐴!,! is (net income before extraordinary items)/average total assets; ∆𝑆𝑎𝑙𝑒𝑠!,! is (𝑆𝑎𝑙𝑒𝑠! −
𝑆𝑎𝑙𝑒𝑠!!! )/average total assets; and 𝑃𝑃𝐸!,! is gross property, plant, and equipment/average total assets.

16
 DACCRi,t = β0 + β1ADOPTERi,t + β2YR2013i,t + β3ADOPTER*YR2013i,t
+ β4SIZEi,t + β5ROAi,t + β6LEVi,t + β7BIG4i,t +β14USLISTi,t
+ IndustryFE + εit, (2b)

where all variables are as defined above. We also include company fixed effects in models 2a

and 2b and report the results in Section 6. Next, for our cross-sectional analyses we estimate the

following model, using firm-year observations for the two years after the regulatory cut-off:

DACCRi,t = β0 + β1REPORT_DISCLi,t + β2SIZEi,t + β3ROAi,t + β4LEVi,t

+ β5BIG4i,t + β6USLISTi,t + IndustryFE + εit, (2c)

where all variables are as defined above.

Analyses of investors’ reaction

Our two proxies for investors’ reaction are (1) cumulative abnormal returns surrounding

the date of the public distribution of the annual report, which includes the auditor’s report (i.e.,

report filing date), and (2) the sum of three-day abnormal trading volume around the report filing

date. Abnormal returns reflect the average change in investors’ belief due to an announcement

event. Abnormal trading volume, in turn, reflects the change in individual investors’ beliefs

(Garfinkel and Sokobin, 2006). The later proxy arguably captures not only information content,

but also information asymmetry and investor disagreement regarding new information. Prior

literature suggests that while somewhat positively correlated, the two measures can be

considered largely independent of each other (Bamber and Cheon, 1995). Any given information

event can result in high abnormal volume and almost no change in returns and vice versa

(Kandel and Pearson, 1995). Thus, using the two measures together increases our chances of

finding an investor reaction to the new auditor’s report. Furthermore, utilizing both measures is

important given the comparatively small size of our sample (Cready and Hurtt, 2002).

17
 The auditor’s report is included in the release of each company’s full annual report. The

U.K. Companies Act mandates that annual reports must be publicly available. Specifically, DTR

6.3 requires a public announcement when the report is submitted to the National Storage

Mechanism and available for viewing on each company’s website. We search the company

announcements of annual filings using the Regulatory News Service (RNS) of the LSE, as well

as the National Storage Mechanism (NSM).20 The release of the full annual report follows an

earlier earnings announcement disclosure. For the companies in our sample the earnings

announcement is a very complete form of disclosure, containing full financial statements and

mentioning whether the auditor’s opinion was unqualified. Moreover, the current listing rules

require preliminary announcements to be agreed with the auditor prior to publication (FRC,

2015). See Appendix C for an illustration about how we identify and calculate daily returns

around the annual report’s public release dates. To examine the effect of the new report on

investors’ reaction, we estimate the following pre-post model:

REACTIONi,t = β0 + β1POSTi,t + β2SIZEi,t + β3ROAi,t + β4LEVi,t + β5USLISTi,t

+ β6IOWNi,t + β7CHNIi,t + β8STDRETi,t + β9MTBi,t + β10BETAi,t
+ IndustryFE + εit, (3a)

where for company i in year t, REACTIONi,t is one of two proxies for investors’ reaction,

ABRETi,t, the sum of the three-day abnormal returns by company, calculated each day as the

20
For example, see the announcement by ITE Group plc on December 24, 2013 in RNS that the annual report is
available at: http://www.londonstockexchange.com/exchange/news/market-news/market-news-detail/11816292.
html. The FRC requirements are described in detail at: http://fshandbook.info/FS/html/handbook/DTR/6/3. We
could not identify reliable report filing dates in Worldscope, Bloomberg, or other online open sources. We find that
the annual report filing dates in Worldscope and Bloomberg often correspond to the earnings release date as per the
RNS announcements. The LSE RNS website can be accessed searching each company at: http://
www.londonstockexchange.com/prices-and-markets/markets/prices.htm. The NSM can be accessed searching each
company at: http://www.morningstar.co.uk/uk/NSM. An alternative source is the website
http://www.investegate.co.uk. For a small number of observations we cannot identify the specific announcement of
the release of the annual report. In these cases we use the date of the announcement of the Annual General Meeting
(if available), or the date of the Annual General Meeting, as release dates.

18
company returns = (Price Closet – Price Closet-1)/Price Closet-1 minus same-day returns for the

LSE value-weighted portfolio;21 or, ABVOLi,t, the natural logarithm of the ratio of the company’s

mean event-period volume divided by the company’s mean estimation-period volume. Event-

period volume is daily volume over the two-day event window beginning on the report filing

date, scaled by shares outstanding. Estimation-period volume is measured over the trading period

beginning 61 days before the earnings announcement date and ending 40 days later (i.e., 21 days

before the earnings announcement date);22 IOWNi,t is the natural logarithm of the number of the

company's shares held by institutional investors/total number of shares outstanding as of the

current fiscal year end;23 CHNIi,t is NIBE in year t minus NIBE in year t-1 scaled by total assets

in year t-1; STDRETi,t is the standard deviation of daily stock returns in year t; MTBi,t is equity

market value divided by book value; and, all other variables are as defined above; and, BETAi,t is

the slope coefficient from regressing the daily return on the company's stock on daily returns of

the value-weighed portfolio, over 220 day period (-250, -21) relative to the filing date of the

current year financial statements.

For our difference-in-differences analyses, using a similar research design as in our audit

fee and audit quality analyses, we estimate the following model:

REACTIONi,t = β0 + β1ADOPTERi,t + β2YR2013i,t + β3ADOPTER*YR2013i,t

+ β4SIZEi,t + β5ROAi,t + β6LEVi,t + β7USLISTi,t + β8IOWNi,t
+ β9CHNIi,t + β10STDRETi,t + β11BETAi,t + β12MTBi,t
+ IndustryFE + εit, (3b)

21
We calculate market variables by aggregating companies’ individual data. We obtain the constituents of the LSE
from WSCOPEUK list from Worldscope for securities on LSE, that are major and primary (EXDSCD = ‘LN’
MAJOR = ‘Y’ and ISINID = ‘P’) and that are not Investment Trusts or Closed-End Funds (TYPE = ‘EQ’).
22
Our measure of abnormal volume is consistent with DeFond et al., (2007) and Miller (2010)
23
We aggregate the percentage as reported by Thomson Reuters where we only exclude ‘Individuals’ from the
ownership reported.

19
where all variables are as defined above. We also include company fixed effects in models 3a

and 3b and report the results in Section 6. For our cross-sectional analyses we estimate the

following model, using firm-year observations for the two years after the regulatory cut-off:

REACTIONi,t = β0 + β1REPORT_DISCLi,t + β2SIZEi,t + β3ROAi,t + β4LEVi,t

+ β5USLISTi,t + β6IOWNi,t + β7CHNIi,t + β8STDRETi,t + β9BETAi,t
+ β10MTBi,t + IndustryFE + εit, (3c)

where all variables are as defined above.

4. SAMPLE SELECTION AND DESCRIPTIVE STATISTICS

Sample selection

Table 1 shows the samples used in our analyses. We start our sample with all 691

companies incorporated in Great Britain, with ordinary stocks listed in the LSE Premium as of

December 31, 2013 and traded in the LSE Main Market.24 We merge this list of companies with

the Thomson Reuters Worldscope database and find 340 non-financial companies with SEDOL

identifiers (i.e., eliminate SIC codes 6000 to 6999). There is a large number of banking,

insurance, trust, and investment companies traded in the LSE Main Market (KPMG, 2014). Our

study focuses on non-financial companies similar to prior literature (e.g., Carcello and Li, 2013;

Czerney et al., 2014a, 2014b; Seetharaman et al., 2002).25

We read and manually code information from annual reports (e.g., annual report and

auditor’s report length; name of audit firm and audit partner; number and description of

24
The companies listed in the LSE are available at: http://www.londonstockexchange.com/statistics/companies-and-
issuers/companies-and-issuers.htm. We used the list available on April 30, 2015.
25
We identified only three “early” adopters in our sample. Vodaphone Group included an auditor’s report based on
the FRC’s exposure draft for fiscal year-end March 31, 2013. Sky plc included an auditor’s report based on the
revised FRC standard issued in June 2013 for fiscal year-end June 30, 2013. Ashmore Group included an auditor’s
report based on the revised FRC standard issued in June 2013 for fiscal year-end June 30, 2013. Vodaphone Group
and Sky plc are included in our sample, for these companies and consistent with the September 2013 cut-off, we
considered March 31, 2014 and June 30, 2014 the first regular year of adoption.

20
significant risks of material misstatements in the auditor’s report; materiality amount; audit fees;

and, number and description of significant accounting issues in the audit committee report).

Finally, we complement our data with financial and returns variables from the Thomson Reuters

databases (Worldscope, Fundamentals and Datastream).

Our main pre-post analyses of audit fees use 1,320 firm-year observations with available

data from two years before (660 observations) and two years after (660 observations) the

regulatory cut-off. In sensitivity analyses we use 1,248 firm-year observations from companies

with data for the full four-year panel. Our differences-in-differences analyses use 664 firm-year

observations that meet the criteria of our research design. There are 126 and 206 companies

reporting six months before and after the regulatory cut-off, also with data in the year prior. Our

cross-sectional analyses of the report length and risk disclosures use 651 firm-year observations

where we could extract the auditor report’s text and count the number of words. Our cross-

sectional analyses of materiality disclosures use 657 firm-year observations.26

Our main pre-post analyses of audit quality use 1,108 firm-year observations with

available data from two years before (568 observations) and two years after (540 observations)

the regulatory cut-off. In sensitivity analyses we use 912 firm-year observations from companies

with data for the full four-year panel. Our differences-in-differences analyses use 572 firm-year

observations that meet the criteria of our research design. There are 110 and 176 companies

reporting six months before and after the regulatory cut-off, also with data in the year prior. Our

cross-sectional analyses of the report length and risk disclosures use 533 firm-year observations

where we could extract the auditor report’s text and count the number of words. Our cross-

sectional analyses of materiality disclosures use 537 firm-year observations.

26
There are a few cases when the companies do not disclose a specific materiality amount. We code these cases as
missing.

21
 Our main pre-post analyses of investors’ reaction use 1,236 firm-year observations with

available data from two years before (615 observations) and two years after (621 observations)

the regulatory cut-off. In sensitivity analyses we use 1,112 firm-year observations from

companies with data for the full four-year panel. Our differences-in-differences analyses use 610

firm-year observations that meet the criteria of our research design. There are 111 and 194

companies reporting six months before and after the regulatory cut-off, also with data in the year

prior. Our cross-sectional analyses of the report length and risk disclosures use 605 firm-year

observations where we could extract the auditor report’s text and count the number of words.

Our cross-sectional analyses of materiality disclosures use 611 firm-year observations. The main

additional data restrictions compared to our fee analyses are the availability of market data and

identifying the date on which each company’s annual report is available to investors (i.e., filing

date) in the LSE RNS and NSM databases.

5. RESULTS

Audit fee analyses

Table 2, Panel A, reports the descriptive statistics for the variables used in our audit fee

analyses. The mean of AFEESi,t is 13.095 when POSTi,t =1 (untabulated, approximately 486,500

British Pounds in the post adoption period), compared to a mean of 13.014 when POSTi,t =0

(untabulated, approximately 448,600 British Pounds in the pre adoption period). However, the

difference in means between the pre and post adoption periods is not statistically significant (at

the ten percent level).

The companies in our sample are relatively large, the mean SIZEi,t is 13.554 when POSTi,t

=1 (untabulated, approximately 767 million pounds in assets in the post adoption period). Most

of the companies have a Big 4 auditor (mean BIG4i,t = 0.936 when POSTi,t =1). The only

22
statistically significant differences in means between the pre and post adoption periods are a one-

percent decrease in ROAi,t (significant at the five-percent level), a three-percent increase in

auditor rotation CHAUDi,t (significant at the one-percent level), and a two-percent increase in the

incidence of going concern opinions GCi,t (significant at the five-percent level).

The length of the auditor’s report has increased by approximately two pages under the

new rules (untabulated, approximately 2,400 words in the new reports, compared to 757 words in

the old reports). The new section on risks of material misstatements has on average four risks

(untabulated, approximately 660 words, approximately 30 percent of the words in the new

report). The average materiality PERCMATi,t is 0.65 percent of total assets (the median is 0.5

percent, or half of one percent).

Table 2, Panel B, Columns 1-2 show coefficients and t-statistics for our pre-post adoption

analyses of audit fees, using data from two years before and two years after the regulatory cut-off

(Model 1a). The coefficient on POSTi,t = 0.038 in Column 1 (statistically significant at the one-

percent level) indicates that fees increased by approximately four percent after the adoption of

the new rules. However, the pre-post analysis doesn’t allow us to rule out a generalized increase

in fees due to other time-varying economic factors besides the adoption of the new report. As

documented in prior studies (Hay et al., 2006), we find that audit fees are positively associated

with client size (SIZEi,t), leverage (LEVi,t), current assets (CURRi,t), number of segments

(NSEGi,t), foreign sales (FORSALESi,t), receiving a going concern opinion (GCi,t), December

fiscal year-ends (DECYEi,t), and U.S. cross-listing status (USLISTi,t).

Table 2, Panel B, Columns 3-4 show coefficients and t-statistics for our differences-in-

differences analyses, using firm-year observations with fiscal year-end six months before (March

to August 2013) and after (September 2013 to February 2014) the regulatory cut-off in

23
September 2013, as well as the same companies in the prior year (Model 1b, see Section 3 for

additional details). The coefficient on ADOPTER*YR2013i,t = 0.021 in Column 3 (statistically

insignificant at the ten-percent level) does not show a short-term difference-in-differences effect

on audit fees. Hence, the 4 percent increase in the pre-post analysis cannot be entirely attributed

to the change in regulation. Also, one could argue that the “true” cost of complying with the new

audit report model is beyond audit fees. For example, if auditors and clients negotiate and split

compliance work to be done as well as audit fees, the “true” costs of the new rules may not be

directly reflected in audit fees.

Table 2, Panel C documents that audit fees are positively associated with the total length

of the report (Column 1, the coefficient on TOTAL_NWRDSit = 0.420 is statistically significant at

the one-percent level); the length of the risks’ discussion (Column 3, the coefficient on

RISK_NWRDSit = 0.082 is statistically significant at the one-percent level); and, the number of

risks (Column 5, the coefficient on NRISKSit = 0.074 is statistically significant at the one-percent

level). These findings suggest an association between auditor’s effort, or auditor’s risk premium,

and the auditor’s assessment and disclosure of difficult audit issues. In other words, the number

and length of the auditor’s risk disclosures are reflected in audit fees. Although unlikely, the

significantly positive relationship between audit fees and risk disclosure could also be

attributable to the risk premium paid by riskier clients. In contrast, in Table 2, Panel D we do not

find evidence of an association between audit fees and materiality. In Column 1, the coefficient

on PERCMATit = 0.128 and in Column 3 the coefficient on LOGMATit = 0.117 are not

statistically significant at the ten-percent level.27

27
Prior research has documented that client’s risks and materiality are associated with audit fees (Bell et al., 2000;
Houston et al., 1999; Pratt and Stice, 1994). A comparatively lower level of materiality and greater risk of
misstatement can lead to an increase in the resources required to obtain audit evidence, and therefore to an increase
in audit fees. Bell et al. (2000) find that the increase in the perceived business risk of the audit engagement results in

24
Audit quality analyses

Table 3, Panel A, reports the descriptive statistics of the data used in our audit quality

analyses. The absolute discretionary accruals for the companies in our sample have a mean of

0.048 in the pre and post adoption period (mean DACCRi,t = 0.048 when POSTi,t =0 and POSTi,t

=1). The only statistically significant difference in means between the pre and post adoption

periods in is a one-percent decrease in ROAi,t (significant at the ten-percent level).

Table 3, Panel B, Columns 1-2 show coefficients and t-statistics for our pre-post adoption

analyses of audit quality, using data from two years before and two years after the regulatory cut-

off date (Model 2a). The coefficient on POSTi,t = -0.001 in Column 1 (statistically insignificant

at the ten-percent level) indicates that audit quality did not change after the new rules were

adopted. As documented in prior studies (e.g., Kothari et al., 2005), we find that absolute

discretionary accruals are negatively associated with client size (SIZEi,t) and profitability

(ROAi,t). We document similar results for our differences-in-differences analyses (Model 2b) in

Table 3, Panel B, Columns 3-4. The coefficient on ADOPTER*YR2013i,t = -0.007 in Column 3

(statistically insignificant at the ten-percent level) does not show a short-term difference-in-

differences effect on audit quality.

In Table 3, Panel C we do not find evidence of an association between audit quality and

the total length of the report (Column 1, the coefficient on TOTAL_NWRDSit = -0.008 is

statistically insignificant at the ten-percent level); the length of the risks’ discussion (Column 3,

the coefficient on RISK_NWRDSit = -0.003 is statistically insignificant at the ten-percent level);

and, the number of risks (Column 5, the coefficient on NRISKSit = -0.001 is statistically

insignificant at the ten-percent level). However, in Table 3, Panel D we find that audit quality is

higher audit fees, through the increase in the billable hours reported by the auditor. Nevertheless, as pointed out by
Bell et al. (2000), specific risks cannot be incorporated in the audit fee determination if they cannot be quantified, if
they are quantifiable only at the industry level, or if market conditions do not allow for the fees to reflect audit risk.

25
positively associated with materiality, suggesting that comparatively smaller materiality is

associated with higher audit quality. In Column 1, the coefficient on PERCMATit = 0.034 and in

Column 3 the coefficient on LOGMATit = 0.021 are statistically significant at the one-percent

level. These findings suggest that audit quality decreases (i.e., absolute accruals increase) as

materiality increases.28

Analyses of investors’ reaction

Table 4, Panel A, reports the descriptive statistics of the data used in our investors’

reaction analyses. Both abnormal returns and abnormal volume are small on the report filing

dates (mean ABRETi,t = 0.001 and ABVOLi,t = -0.029 when POSTi,t =1). The only statistically

significant difference in means between the pre and post adoption periods is a small decrease in

the standard deviation of stock returns (STDRETi,t) and an increase in market-to-book (MTBi,t),

both significant at the one-percent level. Our descriptive statistics suggest that the public release

of the annual report is generally not a significant source of news for investors in the pre or post

period. Appendix C provides an illustration of daily returns on the dates of the earnings release

versus the subsequent public release of the annual report.

Table 4, Panel B, Columns 1-2 and 5-6 show coefficients and t-statistics for our pre-post

adoption analyses of investors’ reaction, using abnormal returns and abnormal volume

respectively (Model 3a). The coefficients on POSTi,t = -0.001 in Column 1, and POSTi,t = -0.022

in Column 5, are statistically insignificant at the ten-percent level. Table 4, Panel B, Columns 3-

28
Investors that are supportive of the new auditor’s report argue that audit judgment (e.g., the auditor’s assessment
of risks and materiality) can increase the understanding of how audit firms apply auditing standards and can also
enable investors to better evaluate the quality of the audit (FRC, 2013). As investors and regulators gain more
insight into the audit process, auditors’ incentives to supply audit quality may increase due to a change in their
litigation and regulatory risk. Lower levels of materiality and greater risk of misstatement are arguably associated
with the amount and type of audit evidence, and therefore, lead to an increase in audit quality. Contrary to this view,
opponents to the new auditor’s report indicate that the risk of misstatements and materiality disclosures may not
enhance the understanding of investors because these issues are too complex to be summarized in the auditor’s
report (FRC, 2013).

26
4 and 7-8 show coefficients and t-statistics for our differences-in-differences analyses. We do not

find a difference-in-difference effect on investors’ reaction. These combined findings indicate

that investors’ reaction to the auditor’s report content did not change, on average, after the new

rules were adopted.

In Table 4, Panel C we do not find evidence of an association between investors’ reaction

and the total length of the report, the length of the risks’ discussion, or the number of risks.

Finally, in Table 4, Panel D we only find a weak association between abnormal returns and the

natural logarithm of materiality (Column 3, the coefficients on LOGMATit is 0.007, statistically

significant at the ten-percent level).29

6. ADDITIONAL AND SENSITIVITY ANALYSES

Fixed effects, companies with data for the full panel, and bootstrapped standard errors

In order to mitigate the effect of time-invariant client characteristics, we also include

company fixed effects in our pre-post and difference-in-differences models. In our audit fee pre-

post model (1a) we find that the coefficient on POSTi,t is 0.027 (untabulated, statistically

significant at the five-percent level), indicating that audit fees increased by approximately three

percent in the two years after the adoption of the new rules. After including company fixed

effects in our pre-post and difference-in-differences models, we do not find evidence that the

new report had an overall impact on audit quality or investors’ reaction, confirming our main

analyses. Furthermore, in order to isolate auditor “style” and idiosyncratic differences among

29
Some research suggests that the disclosure of materiality in the auditor’s report might help reduce the audit
expectation gap (De Martinis and Burrowes, 1996). However, Houghton et al. (2011) also show in a series of
interviews and focus groups that the concept of materiality is not easy to grasp for financial statement users. Users
frequently associate materiality with the size of the items being audited instead of the conservatism and costs of the
audit. Users also seem to focus more on the quantitative rather than the qualitative aspects of materiality. If investors
have trouble in understanding materiality or the materiality threshold is relatively standardized, its disclosure might
not produce a market reaction. Finally, materiality should be the maximum level of misstatement that will not affect
a reasonable financial statement user. The market may not react to disclosed materiality levels if auditors are setting
materiality to an acceptable level (i.e., the market agrees with the auditor’s determination of materiality).

27
audit firms in their reports, we also estimate our pre-post models including auditor fixed effects,

instead of a Big-4 indicator variable, and find similar inferences to those documented in the main

analyses. We also estimate our pre-post models including only those companies that have data

for the full four-year pre-post period. These analyses are comparatively stricter about using each

company in the pre period as its own control for the post period. In our fee model (1a) we find

that the coefficient on POSTi,t is 0.026 (untabulated, statistically significant at the ten-percent

level), indicating that fees increased by approximately three percent in the two years after the

adoption of the new rules. Also, using observations with four years of data in Models 2a and 3a,

we do not find evidence that the new report had an overall impact on audit quality or investors’

reaction, confirming our main analyses. Finally, in order to mitigate concerns that our findings

are the result of low power, we compute standard errors for our pre-post and difference-in-

differences models using bootstrap with 2,000 replications. We find similar inferences to those

documented in the main analyses.

One-year differences pre-post adoption and between the first two years of adoption

In order to isolate one-year changes in the pre-post adoption years, we limit our samples

to the two years around the adoption of the new rules (t and t-1). In our fee model (1a) we find

that the coefficient on POSTi,t is 0.028 (untabulated, statistically significant at the one-percent

level), indicating that fees increased by approximately three percent in the first year after the

adoption of the new rules. We do not find evidence that the new report had an overall one-year

impact on audit quality or investors’ reaction, confirming our main analyses.

In order to compare the two years post adoption (t and t+1), we also estimate a modified

version of our pre-post models. We limit our samples to the post adoption years and include an

indicator variable for the second year of adoption (t+1). We do not find evidence of differences

28
between the two years post adoption in audit fees, audit quality, or investors’ reaction. In the

three analyses, the indicator variable for year t+1 is statistically insignificant (at the ten-percent

level). Finally, we examine the composition and number of risks in the two years post adoption.

We classify individual risks in 16 categories, based on our interpretation of the main issue by

reading the risk headings and disclosures. Table 5 shows the percentage of companies reporting

each risk type in years t and t+1. Revenue recognition is the predominant risk type, disclosed in

71 and 61 percent of the reports in year t and t+1, respectively. Risks associated with impairment

of goodwill and long-term assets, tax accounting, and business combinations are also commonly

disclosed in the reports. Examining the pair-wise correlation between risks, we find that some

associations are positive and some negative. However, we do not find a clear pattern. The only

two risk types that appear often together are revenue recognition and internal control issues

(untabulated, pair-wise correlation is 0.30, statistically significant at the five percent level).

However, among other risks, correlation patterns are unclear. The most important change

between year t and t+1 is a decrease in the incidence of including internal control issues. This

shift seems to follow concerns about the boilerplate nature of some risk disclosures (Citi, 2014).

The mean number of risks in year t and t+1 is 3.97 and 3.88, respectively (untabulated,

the median is 4 in both years). In order to compare the effect of changes in the number of risks in

the two years post adoption t and t+1, we also estimate a modified version of our cross-sectional

models of the association between NRISKSi,t and audit fees, audit quality, and investors’ reaction

(Models 1c, 2c, and 3c). We limit our samples to the post adoption years and include an indicator

variable for the second year of adoption (t+1), as well as an interaction between the second year

indicator and NRISKSi,t. The interaction coefficient between the second year indicator and

NRISKSi,t in the fee model is 0.049 (untabulated, statistically significant at the one-percent level),

29
in the audit quality model is 0.004 (untabulated, statistically significant at the ten-percent level),

and in the abnormal returns and abnormal volume models is not statistically significant (at the

ten-percent level). These results indicate that changes in the number of risks year-over-year can

be meaningful in terms of audit fees and audit quality.

Accounting issues disclosed by the audit committee

There were other changes to corporate reporting in the U.K. around the same time, in

particular the changes to the U.K. Corporate Governance Code in October 2012 (FRC, 2012b).

These changes require additional disclosures by the audit committee of the board of directors in

the corporate governance section of the annual report. To assess the potential importance of the

accounting risks in the audit committee report versus the risks in the auditor’s report, we include

NRISKSi,t the number of risks of material misstatements disclosed in the auditor’s report and also

ACNRISKSi,t the number of risks of material misstatements disclosed in the audit committee’s

report in Models (1c), (2c), and (3c). There is a large degree of overlap between these two

sources of significant risks. In untabulated analyses, the pairwise correlation between NRISKSi,t

and ACNRISKSi,t is 0.54 (statistically significant at the one percent level). The FRC report

examining the first two years of auditor’s reports (FRC, 2016, Table 13) documents that there is

74 and 85 percent overlap in the first and second year of adoption, respectively, on the types of

risk reported by auditors and audit committees. Including both NRISKSi,t and ACNRISKSi,t in our

cross-sectional fee analyses (Model 1c) we find that ACNRISKSi,t is positively associated with

fees (untabulated coefficient 0.036, statistically significant at the five-percent level) and

NRISKSi,t is not associated with audit fees. In our cross-sectional analyses of audit quality and

investor’s reaction (Models 2c and 3c) we do not find statistically significant coefficients (at the

ten percent level) for NRISKSi,t and ACNRISKSi,t. Overall, it is difficult to conclude whether the

30
risks in the auditor’s report are more informative than the risks in the report of the Audit

Committee.

Reportable differences to the audit committee in the materiality section

In addition to reporting a materiality amount, companies also disclose a smaller threshold

for reportable issues to the audit committee. For example, in the report in Appendix A

materiality is 1.3 million pounds, but the auditor discloses that any differences over 62,000

pounds will be discussed with the audit committee, as well as misstatements below that amount

that, in the auditor’s view, warrant reporting for qualitative reasons. The correlation between

materiality and reportable differences scaled by total assets is 0.66 (both statistically significant

at the one percent level). In Models (1c), (2c), and (3c), we included both PERCMATi,t the

materiality amount divided by total assets, and PERCREPDIFFi,t the reportable differences

amount divided by total assets. In untabulated analyses, we find a positive association between

PERCMATi,t and discretionary accruals (statistically significant at the five-percent level), but no

association between PERCREPDIFFi,t and audit fees, discretionary accruals, or investors’

reaction.

Additional proxies for audit quality

As an alternative proxy for audit quality, we use the incidence of small earnings as an

inverse proxy for audit quality (Carcello and Li, 2013), defined as an indicator variable equal to

one if the change in net income before extraordinary items from year t-1 to year t, scaled by

market value of equity at year t-1, is in the interval [0,0.02], and zero otherwise. We estimate

logistic regression models using the incidence of small earnings as a dependent variable and the

same independent variables as in models (2a), (2b), and (2c), excluding industry fixed effects,

given the low incidence of the small earnings variable (the untabulated mean of this variable in

31
our pre-post sample is 0.014). In untabulated analyses, using this proxy we do not find evidence

that the new report resulted in changes in audit quality (i.e., pre-post or difference-in-differences

analyses). We do not find an association between this proxy and the length of the report or risk

section. We find a negative association between this proxy and the number of risks and the

length of the risk section (at the five percent level). We find a positive association between this

proxy and LOGMATi,t, the natural logarithm of the materiality amount (statistically significant at

the ten percent level).

We also use performance-adjusted (i.e., ROA-matched) discretionary accruals, following

Kothari et al. (2005). In untabulated analyses, using this proxy we do not evidence that the new

report resulted in changes in audit quality (i.e., pre-post or difference-in-differences analyses).

We do not find an association between this proxy and the length of the report or risk section.

However, we find a positive association between this proxy and both our materiality variables

(statistically significant at the five percent level). The association between materiality and these

two alternative proxies (i.e., small earnings and performance-adjusted discretionary accruals)

confirm our findings that materiality is inversely related to audit quality.

Finally, in order to investigate whether there was a change in total accruals after the

adoption of the new rules, we estimate our pre-post and difference-in-differences models using

absolute total accruals as the dependent variable. The correlation between absolute total accruals

and discretionary accruals is 0.6 (untabulated, statistically significant at the one percent level). In

untabulated analyses, we also do not find evidence that the new report resulted in changes in

total accruals. The combined findings of these alternative analyses demonstrate that our

inferences are not driven by our estimation of discretionary accruals.

Additional proxies for investors’ reaction and information asymmetry

32
 We used the average spread on the dissemination date of the annual report, including the

auditor’s report, as a proxy for investors’ reaction and changes in information asymmetry.

Average bid-ask spread is widely used as a proxy of information asymmetry (Lee et al., 1993;

Leuz and Verrecchia, 2000). According to theoretical models, market specialists widen the bid-

ask spread when there is greater information asymmetry between informed and liquidity traders

(Glosten and Milgrom, 1985). In untabulated analyses we do not find evidence that the new

auditor’s report had an effect on information asymmetry. We compute the average effective bid-

ask spread during the filing date scaled by trade price (spread = (ask - bid)/m, where m=(ask +

bid)/2).

We also compute abnormal volume using the market’s volume as a benchmark, as the

sum of the three-day trading volume scaled by common shares outstanding among all UK listed

companies for the period (-1,1) surrounding the filing date of the annual financial statements.

Using this alternative abnormal volume calculation, in untabulated analyses we find inconsistent

results, indicating a decrease in abnormal volume after the implementation of the new

requirements (both in the pre-post and difference-in-differences analyses). We also find a

positive association between this alternative proxy and materiality (LOGMATi,t). However,

considering the evidence from all our analyses together, it is difficult to conclude that the new

report had an effect on investors’ reaction.

Finally, we compute absolute three-day cumulative abnormal returns, as well as

cumulative abnormal returns over the period (-5,5) surrounding the filing date of the annual

financial statements. We find similar results (untabulated) to those documented in the main

analyses.

7. CONCLUSION

33
 The auditor’s report is currently the auditor’s only direct communication with

shareholders about the audit process and its outcome. Investors and regulators have expressed

growing concerns about the format of the auditor’s report. In most countries, including the U.S.,

the auditor’s report uses standardized language for the vast majority of companies. The FRC has

moved forward by issuing a new standard that mandates significant changes to the auditor’s

report for most large companies listed on the LSE with fiscal years beginning on or after October

1, 2012.

In this study we examine two years pre and two years post the new report requirements in

the U.K. and provide evidence on the consequences of changing the report’s content in terms of

the cost of audits, audit quality, and investors’ reactions to the information in the report.

Examining the pre-post change for the group of adopter companies alone, we find that audit fees

increased by approximately four percent. However, using an alternative research design

examining a treatment group of new report adopters and a control group of non-adopters

reporting immediately before and after the regulatory cut-off in September 2013, we do not find

evidence of short-term consequences to the new report requirements. Moreover, we do not find

evidence that the new report changes audit quality or investors’ reaction. These findings are

consistent with the FRC’s statements, including the new report’s standard ISA 700, explicitly

mentioning that the new requirements intend to make public information previously discussed

with the audit committee and that meeting these requirements is not likely to be very costly

(FRC, 2013, 2013b, 2013c).

We also conduct cross-sectional analyses examining the association between the new

auditor’s report content and audit fees, audit quality, and investors’ reaction using the sample of

adopter companies in the two years after the cut-off date. We find that the length of the auditor’s

34
report, the length of the risks’ section, and the number of risks are positively associated with

audit fees, suggesting an association between auditor’s risk and effort and the report’s content.

We also find that the materiality amount is positively associated with absolute discretionary

accruals, suggesting that comparatively smaller materiality is associated with higher audit

quality. Arguably, if the risk and materiality disclosures were strictly boilerplate, they would not

be associated with audit fees and quality.

Our study can be of interest to regulators aiming to change the auditor’s report.

Furthermore, this study contributes to the literature examining the consequences of the auditor’s

report content. We also contribute to the literature on risk disclosures in financial reports.

Finally, this study extends the descriptive evidence provided by practitioners on the

consequences of the auditor’s report change in the U.K. However, we note four broad caveats to

our findings. First, there could be other compliance costs, absorbed by clients or auditors and not

reflected in audit fees. Second, the companies in our study are relatively large, are at the “high-

end” of reporting quality, and have a rich information environment, factors which may decrease

the usefulness of the new auditor’s report. For these companies, the public release of the annual

report is generally not a significant source of news for investors. For smaller companies, with

comparatively poorer information environment and fewer information sources, the auditor’s

disclosures may be incrementally informative. Third, it is difficult to fully disentangle the market

reaction to the auditor’s report from other disclosures in the annual report, including the

disclosures by the audit and other board committees (risk, compensation, etc.). Fourth, the new

rules could have a long-term impact not captured by our sample and proxies for the cost of

audits, audit quality, and investors’ reaction.

35
 References

Bamber, L. S., Barron, O. E., Stevens, D. E., 2011. Trading volume around earnings announcements and
other financial reports: Theory, research design, empirical evidence, and directions for future
research. Contemporary Accounting Research 28, 431–471.
Bamber, L.S., Cheon, Y., 1995. Differential price and volume reactions to accounting earnings
announcements. The Accounting Review 70, 417–441.
Bedard, J., Gonthier-Besacier, N., Schatt, A., 2015. Cost and benefits of reporting key audit matters in the
audit report: The French experience. Working Paper. Laval University.
Bell, T. B., Landsman, W. R., Shackelford, D. A., 2000. Auditors' perceived business risk and audit fees:
analysis and evidence. Journal of Accounting Research 39, 35–43.
Butler, M., Leone, A., Willenborg, M., 2004. An empirical analysis of auditor reporting and its
association with abnormal accruals. Journal of Accounting and Economics 37, 139–165.
Campbell, J. L., Chen, H., Dhaliwal, D. S., Lu, H., Steele, L.B., 2014. The information content of
mandatory risk factor disclosures in corporate filings. Review of Accounting Studies 19, 396–455.
Carcello, J. V., Li, C., 2013. Costs and benefits of requiring an engagement partner signature: Recent
experience in the United Kingdom. The Accounting Review 88 (5), 1511–1546.
Citi Research (Citi), 2014. New U.K. auditor’s reports. A review of the new information.
Citi Research (Citi). 2015. UK auditor’s reports. Findings from the latest FTSE 100 auditor’s reports.
Cready, W., Hurtt, D., 2002. Assessing investor response to information events using return and volume
metrics. The Accounting Review 77 (4), 891–909.
Czerney, K., Schmidt, J. J., Thompson, A. M., 2014a. Do investors respond to explanatory language
included in unqualified audit reports? Working paper. University of Illinois and University of
Texas at Austin.
Czerney, K., Schmidt, J. J., Thompson, A. M., 2014b. Does auditor explanatory language in unqualified
audit reports indicate increased financial misstatement risk? The Accounting Review 89, 2115–
2149.
DeFond, M., Zhang, J., 2014. A review of archival auditing research. Journal of Accounting and
Economics 58, 275–326.
DeFond, M., Hung, M., Trezevant, R., 2007. Investor protection and the information content of annual
earnings announcements: International evidence. Journal of Accounting and Economics, 37–67.
EY, 2014. Enhancing the auditor’s report. Overview. http://www.ey.com/Publication/vwLUAssets/EY-
overview-enhancing-the-auditors-report/$FILE/EY-overview-enhancing-the-auditors-report.pdf
(accessed 02.10.15).
Financial Reporting Council (FRC), 2012a. Guidance on audit committees. London, England: The

36
 Financial Reporting Council Limited. https://www.frc.org.uk/Our-Work/Publications/Corporate-
Governance/Guidance-on-Audit-Committees-September-2012.aspx (accessed 02.10.15).
Financial Reporting Council (FRC), 2012b. Discussion paper. Revisions to the UK Corporate
Governances Code and guidance on audit committees. London, England: The Financial Reporting
Council Limited. https://www.frc.org.uk/Our-Work/Publications/Corporate-
Governance/Consultation-Document-revisions-to-the-UK-Corporat-File.pdf. (accessed
02.10.15).
Financial Reporting Council (FRC), 2013a. Standard. The independent auditor’s report on financial
statements. International Standard on Auditing (UK and Ireland) 700. London, England: The
Financial Reporting Council Limited. https://www.frc.org.uk/Our-Work/Publications/Audit-and-
Assurance-Team/ISA-700-(UK-and-Ireland)-700-(Revised)-File.pdf. (accessed 02.10.15).
Financial Reporting Council (FRC), 2013b. Feedback Statement. Feedback Statement on consultation on
proposed revision to ISA (UK and Ireland) 700. London, England: The Financial Reporting
Council Limited. https://www.frc.org.uk/Our-Work/Publications/Audit-and-Assurance-
Team/Consultation-Paper-Revision-to-ISA-(UK-and-Ireland/Feedback-Statement-proposed-
revision-to-ISA-(UK-an.aspx. (accessed 02.10.15).
Financial Reporting Council (FRC), 2013c. Consultation Paper. Revision to ISA (UK and Ireland) 700.
London, England: The Financial Reporting Council Limited.
https://www.frc.org.uk/getattachment/d24bb652-e319-46a4-add5-793d518a035b/Consultation-
Paper-Revision-to-ISA-(UK-and-Ireland.aspx. (accessed 02.10.15).
Financial Reporting Council (FRC), 2014. The U.K. Corporate Governance Code. London, England: The
Financial Reporting Council Limited. https://www.frc.org.uk/Our-Work/Publications/Corporate-
Governance/UK-Corporate-Governance-Code-2014.pdf. (accessed 02.10.15).
Financial Reporting Council (FRC), 2015. Discussion paper. Extended auditor’s report: a review of
experience in the first year. London, England: The Financial Reporting Council Limited.
https://www.frc.org.uk/Our-Work/Publications/Audit-and-Assurance-Team/Extended-auditor-s-
reports-A-review-of-experience.pdf. (accessed 02.10.15).
Financial Reporting Council (FRC). 2016. Discussion paper. Extended auditor’s reports: a further review
of experience. London, England: The Financial Reporting Council Limited, (January).
https://frc.org.uk/Extended-auditors-reports.pdf. (accessed 02.10.15).
Garfinkel, J.A., Sokobin, J., 2006. Volume, opinion divergence, and returns: A study of post-earnings
announcement drift. Journal of Accounting Research 44, 85–112.
Glosten, L.R., Milgrom, P.R., 1985. Bid, ask and transaction prices in a specialist market with
heterogeneously informed traders. Journal of Financial Economics 14, 71–100.

37
Hay, D. C., Knechel, W. R., Wong, N., 2006. Audit fees: A meta-analysis of the effect of supply and
demand attributes. Contemporary Accounting Research 23, 141–191.
Houston, R.W., Peters, M. F., Pratt, J. H., 1999. The audit risk model, business risk and audit-planning
decisions. The Accounting Review 74, 281–298.
International Auditing and Assurance Standards Board (IAASB), 2013. IAASB proposes standards to
fundamentally transform the auditor's report; focuses on communicative value to users. IAASB
Press Release New York, N.Y. (July 25). https://www.ifac.org/news-events/2013-07/iaasb-
proposes-standards-fundamentally-transform-auditors-report-focuses-communi. (accessed
02.10.15).
International Auditing and Assurance Standards Board (IAASB), 2015. International Standard on
Auditing (ISA) 700 (Revised), forming an opinion and reporting on financial statements. New
York, N.Y. https://www.ifac.org/publications-resources/international-standard-auditing-isa-700-
revised-forming-opinion-and-reporting. (accessed 02.10.15).
International Auditing and Assurance Standards Board (IAASB). 2015b. International Standard on
Auditing (ISA) 701, Communicating Key Audit Matters in the Independent Auditor’s Report. New
York, N.Y. http://www.ifac.org/system/files/publications/files/ISA-701_2.pdf. (accessed
02.10.15).
International Auditing and Assurance Standards Board (IAASB). 2015c. IAASB Issues Final Standards
to Improve Auditor’s Report. IAASB Press Release New York, N.Y. (January 15).
http://www.ifac.org/news-events/2015-01/iaasb-issues-final-standards-improve-auditors-report.
(accessed 02.10.15).
Kandel, E., Pearson, N., 1995. Differential interpretation of public signals and trade in speculative
markets. Journal of Public Economy 103, 831–872.
Kothari, S.P., Leone, A., Wasley, C., 2005. Performance matched discretionary accrual measures. Journal
of Accounting and Economics 39, 163–197.
KPMG, 2014. Audit committees’ and auditors’ reports: A survey of December year-ends under the new
reporting. http://www.kpmg.com/UK/en/IssuesAndInsights/ArticlesPublications/Documents/PDF
/Issues%20and%20Insights/audit-comm-audit-reports.pdf. (accessed 02.10.15).
Kravet, T., Muslu, V., 2013. Textual risk disclosures and investors’ risk perceptions. Review of
Accounting Studies 18, 1088–1122.
Lawrence, M., Minutti-Meza, M., Zhang, P., 2011. Can Big 4 versus Non-Big 4 differences in audit-
quality proxies be attributed to client characteristics? The Accounting Review 86, 259–286.
Lee, C., Mucklow, B., Ready, M.J., 1993. Spreads, depths, and the impact of earnings information: An
intraday analysis. Review of Financial Studies 6, 345–374.

38
Lennox, C.S., Schmidt, J. J., Thompson, A.M., 2015. Is the expanded model of audit reporting
informative to investors? Evidence from the UK. Working paper. University of Southern
California, University of Texas at Austin, and University of Illinois at Urbana-Champaign.
Leuz, C., Verrecchia, R. E., 2000. The economic consequences of increased disclosure. Journal of
Accounting Research 38, 91–124.
Menon, K., Williams, D. D., 2010. Investor reaction to going concern audit reports. The Accounting
Review 85: 2075–2105.
Miller, B.P., 2010. The effects of reporting complexity on small and large investor trading. The
Accounting Review 85, 2107–2143.
Norris, F., 2014. Holding auditors accountable on reports. The New York Times (May 8).
Public Company Accounting Oversight Board (PCAOB), 2013. Proposed Auditing Standards – The
Auditor’s report on an audit of financial statements when the auditor expresses an unqualified
opinion; The auditor’s responsibilities regarding other information in certain documents
containing audited financial statements and the related auditor’s report; and related amendments
to PCAOB Standards. PCAOB Release No. 2013-005. Washington, D.C..
http://pcaobus.org/Rules/Rulemaking/Docket034/Release_2013-005_ARM.pdf. (accessed
02.10.2015).
Pratt, J., Stice, J. D., 1994. The effects of client characteristics on auditor litigation risk judgments,
required audit evidence, and recommended audit fees. The Accounting Review 69, 639–656.
PwC. 2015. The revolution in audit reports. (August). http://www.pwc.co.uk/assets/pdf/revolution-in-
audit-reports.pdf (accessed 02.10.15).
Reichelt, K. J., Wang, D., 2010. National and office-specific measures of auditor industry expertise and
effects on audit quality. Journal of Accounting Research 48, 647–686.
Reid, L., Carcello, J.V., Li, C., Neil, T, 2015a. Are auditor and audit committee report changes useful to
investors? Evidence from the United Kingdom. Working paper. University of Pittsburgh and
University of Tennessee.
Reid, L., Carcello, J.V., Li, C., Neil, T., 2015b. Impact of auditor and audit committee report changes on
audit quality and costs: Evidence from the United Kingdom. Working paper. University of
Pittsburgh and University of Tennessee.
Seetharaman, A., Gul, F. A., Lynn, S.G., 2002. Litigation risk and audit fees: Evidence from U.K. firms
cross-listed on US markets. Journal of Accounting and Economics 33, 91–115.
Taffler, R.J., Lu, J., Kausar, A., 2004. In denial? Stock market underreaction to going-concern audit
report disclosures. Journal of Accounting and Economics 38, 263–296.

39
 Appendix A
Example of 2013 new auditor’s report Hilton Food Group plc (SEDOL B1V9NW5)

Independent auditors’ report to the members of Hilton Food Group plc

Report on the financial statements

The financial statements, defined below, give a true and fair view of the state of the Group’ and of the
Company’ affairs as at 29 December 2013 and of the Group’ profit and of the Group’ and Company’ cash
flows for the year then ended;

Our opinion
The Group financial statements have been properly prepared in accordance with International Financial
Reporting Standards (IFRSs) as adopted by the European Union; The Parent Company financial
statements have been properly prepared in accordance with IFRSs as adopted by the European Union and
as applied in accordance with the provisions of the Companies Act 2006; and, The financial statements
have been prepared in accordance with the requirements of the Companies Act 2006 and, as regards the
Group financial statements, Article 4 of the IAS Regulation. This opinion is to be read in the context of
what we say below.

What we have audited

The Group financial statements and Parent Company financial statements (the financial statements),
which are prepared by Hilton Food Group plc, comprise:
• the Group and Company balance sheets as at 29 December 2013;
• the Group income statement and statement of comprehensive income for the year then ended;
• the Group and Company statements of changes in equity and statements of cash flows for the year
then ended; and,
• the notes to the financial statements, which include a summary of significant accounting policies and
other explanatory information.
The financial reporting framework that has been applied in their preparation comprises applicable law and
IFRSs as adopted by the European Union and, as regards the Parent Company, as applied in accordance
with the provisions of the Companies Act 2006.

What an audit of financial statements involves

We conducted our audit in accordance with International Standards on Auditing (U.K. and Ireland) (ISAs
(U.K. & Ireland)). An audit involves obtaining evidence about the amounts and disclosures in the
financial statements sufficient to give reasonable assurance that the financial statements are free from
material misstatement, whether caused by fraud or error. This includes an assessment of:
• whether the accounting policies are appropriate to the Group’ and Company’ circumstances and have
been consistently applied and adequately disclosed;
• the reasonableness of significant accounting estimates made by the Directors; and
• the overall presentation of the financial statements.
In addition, we read all the financial and non-financial information in the Annual Report and financial
statements (the Annual Report) to identify material inconsistencies with the audited financial statements
and to identify any information that is apparently materially incorrect based on, or materially inconsistent
with, the knowledge acquired by us in the course of performing the audit. If we become aware of any
apparent material misstatements or inconsistencies we consider the implications for our report.

Overview of our audit approach

Materiality

40
We set certain thresholds for materiality. These helped us to determine the nature, timing and extent of
our audit procedures and to evaluate the effect of misstatements, both individually and on the financial
statements as a whole. Based on our professional judgement, we determined materiality for the Group
financial statements as a whole to be £1.3 million which is approximately 5% of pre-tax profits. Although
our audit was designed to identify material misstatements, we agreed with the Audit Committee that we
would report to them misstatements identified during our audit above £62,000 as well as misstatements
below that amount that, in our view, warranted reporting for qualitative reasons.

Overview of the scope of our audit

The Group is structured as a parent company with 11 subsidiary undertakings, of which six are trading
subsidiaries, and a joint venture company. The Group financial statements are a consolidation of these
entities, which, together, comprise the Group’ operating businesses, centralised functions and
intermediary holding companies. In establishing the overall approach to the Group audit, we determined
the type of work that needed to be performed at these entities by us, as the Group engagement team, or
component auditors within other PwC network firms and other firms operating under our instruction.
Where the work was performed by component auditors, we determined the level of involvement we
needed to have in the audit work at those entities to be able to conclude whether sufficient appropriate
audit evidence had been obtained as a basis for our opinion on the Group financial statements as a whole.
We determined that all 11 companies within the Group required an audit of their complete financial
information.

Areas of particular audit focus

In preparing the financial statements, the Directors made a number of subjective judgements, for example
in respect of significant accounting estimates that involved making assumptions and considering future
events that are inherently uncertain. We primarily focused our work in these areas by assessing the
Directors’ judgements against available evidence, forming our own judgements, and evaluating the
disclosures in the financial statements. In our audit, we tested and examined information, using sampling
and other auditing techniques, to the extent we considered necessary to provide a reasonable basis for us
to draw conclusions. We obtained audit evidence through testing the effectiveness of controls, substantive
procedures or a combination of both. We considered the following areas to be those that required
particular focus in the current year. This is not a complete list of all risks or areas of audit focus identified
by our audit. We discussed these areas of focus with the Audit Committee. Their report on those matters
that they considered to be significant issues in relation to the financial statements is included within the
risk management and internal controls section of the Report of the Audit Committee on page 39.

1) Fraud in revenue recognition. ISAs (U.K. & Ireland) presume there is a risk of fraud in revenue
recognition because of the pressure management may feel to achieve the planned results. We focused
on the amount and timing of the recognition of revenue, particularly where the contractual customer
terms provide for cost plus pricing or discounts, which are calculated by management. We tested
revenue by agreeing it to supporting documentation, including customer contracts, discounts and
incentives, and to cash receipts. We also tested journal entries posted to revenue accounts by
identifying and challenging unusual or irregular items.

2) Risk of management override of internal controls. ISAs (U.K. & Ireland) require that we consider
this. We assessed the overall control environment of the Group, including the arrangements for staff
to whistle-blow inappropriate actions, and interviewed senior management and the Group’ internal
audit function. We examined the significant accounting estimates and judgements relevant to the
financial statements for evidence of bias by the Directors that may represent a risk of material
misstatement due to fraud. In particular, we challenged the estimates in respect of the cost plus
revenue adjustments and customer rebates. We also tested journal entries.

41
Going Concern
Under the Listing Rules we are required to review the Directors’ statement, set out on page 21, in relation
to going concern. We have nothing to report having performed our review. As noted in the Directors’
statement, the Directors have concluded that it is appropriate to prepare the Group’ and Company’
financial statements using the going concern basis of accounting. The going concern basis presumes that
the Group and Company have adequate resources to remain in operation, and that the Directors intend
them to do so, for at least one year from the date the financial statements were signed. As part of our audit
we have concluded that the Directors’ use of the going concern basis is appropriate. However, because
not all future events or conditions can be predicted, these statements are not a guarantee as to the Group’
and the Company’ ability to continue as a going concern.

Opinions on other matters prescribed by the Companies Act 2006

In our opinion:
• the information given in the Strategic report and the Directors’ report for the financial year for which
the financial statements are prepared is consistent with the financial statements;
• the part of the Directors’ remuneration report to be audited has been properly prepared in accordance
with the Companies Act 2006; and
• the information given in the Corporate governance statement set out on pages 36 to 38 in the Annual
Report with respect to internal control and risk management systems and about share capital
structures is consistent with the financial statements.

Other matters on which we are required to report by exception

Adequacy of accounting records and information and explanations received
Under the Companies Act 2006 we are required to report to you if, in our opinion:
• we have not received all the information and explanations we require for our audit; or
• adequate accounting records have not been kept by the Company, or returns adequate for our audit
have not been received from branches not visited by us; or
• the Company financial statements and the part of the Directors’ remuneration report to be audited are
not in agreement with the accounting records and returns.
We have no exceptions to report arising from this responsibility.

Directors’ remuneration
Under the Companies Act 2006, we are required to report to you if, in our opinion, certain disclosures of
Directors’ remuneration specified by law have not been made. We have no exceptions to report arising
from this responsibility.

Corporate Governance Statement

Under the Companies Act 2006, we are required to report to you if, in our opinion a corporate governance
statement has not been prepared by the Company. We have no exceptions to report arising from this
responsibility. Under the Listing Rules we are required to review the part of the Corporate Governance
Statement relating to the Company’ compliance with nine provisions of the U.K. Corporate Governance
Code (the Code). We have nothing to report having performed our review. On page 53 of the Annual
Report, as required by the Code Provision C.1.1, the directors state that they consider the Annual Report
taken as a whole to be fair, balanced and understandable and provides the information necessary for
members to assess the Group’ performance, business model and strategy. On page 39, as required by
C.3.8 of the Code, the Audit Committee has set out the significant issues that it considered in relation to
the financial statements, and how they were addressed. Under ISAs (U.K. & Ireland) we are required to
report to you if, in our opinion:
• the statement given by the Directors is materially inconsistent with our knowledge of the Group
acquired in the course of performing our audit; or

42
• the section of the Annual Report describing the work of the Audit Committee does not appropriately
address matters communicated by us to the Audit Committee.
We have no exceptions to report arising from this responsibility.

Other information in the Annual Report

Under ISAs (U.K. & Ireland), we are required to report to you if, in our opinion, information in the
Annual Report is:
• materially inconsistent with the information in the audited financial statements; or
• apparently materially incorrect based on, or materially inconsistent with, our knowledge of the Group
and Parent Company acquired in the course of performing our audit; or
• is otherwise misleading.
We have no exceptions to report arising from this responsibility.

Responsibilities for the financial statements and the audit

Our responsibilities and those of the Directors
As explained more fully in the Directors’ responsibilities statement on page 53, the Directors are
responsible for the preparation of the Group and Company financial statements and for being satisfied
that they give a true and fair view. Our responsibility is to audit and express an opinion on the Group and
Parent Company financial statements in accordance with applicable law and ISAs (U.K. & Ireland).
Those standards require us to comply with the Auditing Practices Board’ Ethical Standards for Auditors.
This report, including the opinions, has been prepared for and only for the Company’ members as a body
in accordance with Chapter 3 of Part 16 of the Companies Act 2006 and for no other purpose. We do not,
in giving these opinions, accept or assume responsibility for any other purpose or to any other person to
whom this report is shown or into whose hands it may come save where expressly agreed by our prior
consent in writing.

Martin Pitt (Senior Statutory Auditor)

for and on behalf of PricewaterhouseCoopers LLP
Chartered Accountants and Statutory Auditors
Belfast 26 March 2014

43
 Appendix B
Variable definitions

Variable Definition Source

AFEES natural logarithm of audit fees; Annual financial reports
POST time indicator variable, equal to one for the two years Annual financial reports
following the new report requirements (starting on
September 2013), and zero otherwise;
YR2013 time indicator variable, equal to one for companies with Annual financial reports
fiscal year-ends in the six months around the adoption date
in September 2013 (March 2013 to February 2014), and
zero for the same companies in the year before (March
2012 to February 2013);
ADOPTER indicator variable for the September cut-off of the new Annual financial reports
report requirements, equal to one for companies with fiscal
year-ends in the six months September to February, and
zero in the six months from March to August;
SIZE natural logarithm of total assets; Worldscope
ROA net income before extraordinary items/total assets; Worldscope
LEV long-term debt/total assets; Worldscope
CURR current assets/total assets; Worldscope
NSEG natural logarithm of the number of business segments; Worldscope
FORSALES sales outside the U.K./total sales; Worldscope
BIG4 one if the company is audited by a Big-4 auditor (Deloitte, Annual financial reports
EY, KPMG, and PwC), and zero otherwise;
CHAUD one if the company had an auditor change in the current Annual financial reports
fiscal year, and zero otherwise;
GC one if the audit report includes a going concern Annual financial reports
qualification, and zero otherwise;
DECYE one if the company has a fiscal year-end in December, and Worldscope and annual
zero otherwise; financial reports
USLIST one if the company is cross-listed on the NYSE, AMEX, ADR database of Bank of
or NASDAQ stock exchanges, and zero otherwise; NY Mellon
IndustryFE 1-digit SIC Codes industry fixed effects; Worldscope
TOTAL_NWR natural logarithm of the total number of words in the Annual financial reports
DS auditor’s report;
RISK_WRDS natural logarithm of the number of words in the risks Annual financial reports
section of the auditor’s report;
NRISKS number of risks disclosed in the auditor’s report; Annual financial reports

44
 Variable Definition Source
PERCMAT materiality amount/total assets; Annual financial
reports/Worldscope
LOGMAT natural logarithm of the materiality amount; Annual financial reports
DACCR absolute discretionary accruals, estimated using the Jones Worldscope
model including ROA, scaling all variables by average
assets;
ABRET the sum of the three-day abnormal returns by company, Datastream
calculated each day as the company returns = (Price Closet
– Price Closet-1)/Price Closet-1 minus same-day returns for
the LSE value-weighted portfolio;
ABVOL the natural logarithm of the ratio of the company’s mean Datastream
event-period volume divided by the company’s mean
estimation-period volume. Event-period volume is daily
volume on the two-day event window beginning on the
report filing date, scaled by shares outstanding.
Estimation-period volume is measured over the trading
period beginning 61 days before the earnings
announcement date and ending 40 days later (i.e., 21 days
before the earnings announcement date);
IOWN Natural logarithm of the number of the company's shares ThomsonOne Banker’s
held by institutional investors/total number of shares Ownership Module
outstanding as of the current fiscal year-end.
CHNI (NI before extraordinary itemst – NI before extraordinary Worldscope
itemst-1)/Total Assetst-1;
STDRET standard deviation of daily stock returns in year t; Datastream
BETA slope coefficient from regressing the daily return on the Datastream
company's stock on daily returns of the value-weighed
portfolio, over 220 day period (-250, -21) relative to the
filing date of the current year financial statements; and,
MTB equity market value/book value Worldscope

45
 Appendix C
Example of daily returns around earnings announcement and annual report’s public
release dates (Ashtead Group plc SEDOL 0053673)

Figure 1: Daily returns graph including the annual earnings announcement on June 20, 2013, and
the subsequent public release of the annual report on July 25, 2013. The company’s fiscal year-
end was April 30, 2013 and this was the year before the adoption of the new rules.
.06
.04
.02
Daily stock return
0
-.02
-.04
-.06
-.08

Jun1

Jun20

Jul25

Aug15

Ashtead Group plc returns in the period from June 1 to August 15, 2013

Figure 2: Daily returns graph including the annual earnings announcement on June 17, 2014, and
the subsequent public release of the annual report on July 31, 2014. The company’s fiscal year-
end was April 30, 2014 and this was the first year after the adoption of the new rules.
.04
.02
Daily stock return
-.02 0 -.04
-.06
Jun1

Jun17

Jul31

Aug15

Ashtead Group plc returns in the period from June 1 to August 15, 2014

46
Figure 3: Example of LSE RNS output showing public disclosures made by the company,
including the annual earnings announcement on June 17, 2014 “Final Results”, and the public
release of the company’s annual report on July 31, 2014 “Annual Financial Report”, on the first
year of adoption of the new auditor’s report rules

This appendix shows an example of how we search dates in order to examine investors’ reaction to the annual
report’s public release. The auditor’s report is included in the release of each company’s full annual report. The
release of the full annual report follows an earlier earnings announcement disclosure. The U.K. Companies Act
mandates that annual reports must be publicly available. Specifically, DTR 6.3 requires a public announcement
when the report is submitted to the National Storage Mechanism (NSM) and available for viewing on each
company’s website. We search announcements of annual filings using the Regulatory News Service (RNS) of the
LSE, as well as the NSM. For the companies in our sample the earnings announcement is a very complete form of
disclosure, containing full financial statements and mentioning whether the auditor’s opinion was unqualified. The
LSE RNS website can be accessed searching each company at: http:// www.londonstockexchange.com/prices-and-
markets/markets/prices.htm. An alternative source is the website http://www.investegate.co.uk.

47
 Figure 1-A: Pre-post analyses

Two years before the adoption of new rules Two years after the adoption of new rules
September 2011 - August 2013 September 2013 - August 2015

September 2013
Regulatory cut-off date

Figure 1-B: Difference-in-differences analyses

Adopters
First year of new reports
Non-adopters September 2013 to
March to August 2013 February 2014

September 2013

Adopters
Non-adopters one-year prior
one-year prior September 2012 to
March to August 2012 February 2013

September 2012

Figure 1-C: Cross-sectional analyses of new report’s disclosures

Two years after the adoption of new rules

September 2013 - August 2015

September 2013

These figures illustrate the timelines for the three types of analyses in our study: pre-post analyses, using company
data for four years around the regulatory cut-off; difference-in-differences analyses, using company data for six
months before and after the regulatory cut-off date, as well as a year prior; and, cross-sectional analyses using
company data for the two years after the regulatory cut-off date.

48
 Table 1: Sample selection

U.K. Companies on the LSE main market with premium listing as of December 31, 2013 691
Exclude financial companies and companies without SEDOL identifiers in Worldscope (351)
Non-Financial companies with SEDOL identifiers in Worldscope 340

Audit fee analyses Number of observations

Pre-post analyses Year t-2 Year t-1 Year t Year t+1 Total
Data available by fiscal year 327 333 337 323 1,320
Data available for full panel 312 312 312 312 1,248
Risks analyses 336 315 651
Materiality analyses 334 323 657
Difference-in-differences analyses Non-adopters Adopters
Year t-1 Year t Year t-1 Year t
126 126 206 206 664

Analyses of audit quality Number of observations

Pre-post analyses Year t-2 Year t-1 Year t Year t+1 Total
Data available by fiscal year 279 289 295 245 1,108
Data available for full panel 228 228 228 228 912
Risks analyses 294 239 533
Materiality analyses 292 245 537
Difference-in-differences analyses Non-adopters Adopters
Year t-1 Year t Year t-1 Year t
110 110 176 176 572

Analyses of investors' reaction Number of observations

Pre-post analyses Year t-2 Year t-1 Year t Year t+1 Total
Data available by fiscal year 308 307 317 304 1,236
Data available for full panel 278 278 278 278 1,112
Risks analyses 315 290 605
Materiality analyses 313 298 611
Difference-in-differences analyses Non-adopters Adopters
Year t-1 Year t Year t-1 Year t
111 111 194 194 610

This table shows the samples for the analyses of audit fees, audit quality, and investor’s reaction: pre-post analyses,
using company data for four years around the regulatory cut-off; difference-in-differences analyses, using company
data for six months before and after the regulatory cut-off date, as well as a year prior; and, cross-sectional analyses
using company data for the two years after the regulatory cut-off date.

49
 Table 2: Audit fee analyses

Panel A: Descriptive statistics

POSTi,t =0 POSTi,t =1 Diff. in
Variables Mean Median Std. Dev. Mean Median Std. Dev. means
AFEESi,t 13.014 12.899 1.294 13.095 12.997 1.287
SIZEi,t 13.488 13.481 1.825 13.554 13.500 1.814
ROAi,t 0.056 0.057 0.088 0.045 0.054 0.102 **
LEVi,t 0.555 0.550 0.235 0.557 0.545 0.229
CURRi,t 0.420 0.402 0.224 0.421 0.394 0.221
NSEGi,t 0.942 1.099 0.650 0.934 1.099 0.683
FORSALESi,t 0.390 0.311 0.366 0.387 0.310 0.361
BIG4i,t 0.935 1.000 0.247 0.936 1.000 0.244
CHAUDi,t 0.023 0.000 0.149 0.056 0.000 0.230 ***
GCi,t 0.011 0.000 0.103 0.030 0.000 0.172 **
DECYEi,t 0.476 0.000 0.500 0.494 0.000 0.500
USLISTi,t 0.052 0.000 0.221 0.052 0.000 0.221
TOTAL_NWRDSi,t 7.785 7.769 0.296
RISK_NWRDSi,t 6.500 6.551 0.740
NRISKSi,t 3.928 4.000 1.465
PERCMATi,t 0.652 0.502 0.535
LOGMATi,t 8.311 8.281 1.687

Panel B: Pre-post adoption and difference-in-differences analyses

Dependent variable = AFEESi,t
Variables Coeff. t-stat Coeff. t-stat
POSTi,t 0.038*** [2.64]
ADOPTERi,t 0.049 [0.53]
YR2013i,t 0.007 [0.33]
ADOPTER*YR2013i,t 0.021 [0.77]
SIZEi,t 0.565*** [24.65] 0.567*** [24.35]
ROAi,t 0.007 [0.02] -0.093 [-0.27]
LEVi,t 0.505*** [3.54] 0.554*** [3.77]
CURRi,t 0.468*** [2.87] 0.476*** [2.93]
NSEGi,t 0.159*** [3.67] 0.161*** [3.65]
FORSALESi,t 0.718*** [7.56] 0.760*** [7.90]
BIG4i,t 0.031 [0.26] -0.008 [-0.07]
CHAUDi,t -0.106 [-1.65] -0.057 [-0.59]
GCi,t 0.263* [1.69] -0.057 [-0.29]
DECYEi,t 0.214*** [3.64] 0.200** [2.20]
USLISTi,t 0.351*** [2.78] 0.277** [2.04]
Intercept 3.564*** [13.15] 3.709*** [13.41]
Industry F.E. Included Included
N. Observations 1,320 664
Adj. R2 0.838 0.841

50
Panel C: Cross-sectional analyses of audit fees and length of the report and risks’ section
Dependent variable = AFEESi,t
Variables Coeff. t-stat Coeff. t-stat Coeff. t-stat
TOTAL_NWRDSi,t 0.420*** [4.58]
RISK_NWRDSi,t 0.082*** [2.76]
NRISKSi,t 0.074*** [3.68]
SIZEi,t 0.548*** [23.07] 0.555*** [23.01] 0.546*** [22.59]
ROAi,t 0.177 [0.50] 0.055 [0.16] 0.127 [0.36]
LEVi,t 0.407*** [2.76] 0.448*** [2.96] 0.386** [2.53]
CURRi,t 0.527*** [3.23] 0.500*** [2.97] 0.512*** [3.13]
NSEGi,t 0.139*** [3.00] 0.139*** [2.92] 0.140*** [2.98]
FORSALESi,t 0.617*** [5.99] 0.661*** [6.40] 0.631*** [6.15]
BIG4i,t 0.015 [0.11] 0.046 [0.34] 0.121 [0.90]
CHAUDi,t -0.177** [-2.49] -0.170** [-2.23] -0.156** [-2.07]
GCi,t 0.364** [2.51] 0.402*** [2.60] 0.374** [2.44]
DECYEi,t 0.248*** [4.03] 0.239*** [3.80] 0.230*** [3.74]
USLISTi,t 0.361*** [2.71] 0.418*** [3.04] 0.361*** [2.61]
Intercept 0.692 [1.00] 3.387*** [11.52] 3.647*** [12.75]
Industry F.E. Included Included Included
N. Observations 651 651 651
Adj. R2 0.837 0.832 0.835

Panel D: Cross-sectional analyses of audit fees and materiality

Dependent variable = AFEESi,t
Variables Coeff. t-stat Coeff. t-stat
PERCMATi,t 0.128 [1.13]
LOGMATi,t 0.117 [1.56]
SIZEi,t 0.574*** [22.39] 0.459*** [6.47]
ROAi,t -0.137 [-0.49] -0.197 [-0.68]
LEVi,t 0.494*** [3.33] 0.521*** [3.57]
CURRi,t 0.452** [2.50] 0.445** [2.53]
NSEGi,t 0.148*** [3.24] 0.148*** [3.19]
FORSALESi,t 0.678*** [6.73] 0.681*** [6.71]
BIG4i,t 0.117 [0.80] 0.090 [0.65]
CHAUDi,t -0.163** [-2.14] -0.155** [-2.04]
GCi,t 0.312* [1.80] 0.352** [2.12]
DECYEi,t 0.221*** [3.70] 0.220*** [3.64]
USLISTi,t 0.385*** [2.77] 0.379*** [2.75]
Intercept 3.536*** [10.53] 4.219*** [9.87]
Industry F.E. Included Included
N. Observations 657 657
Adj. R2 0.831 0.831

Panel A of this table includes descriptive statistics for the dependent and independent variables for the analyses of
audit fees. Panel B Columns 1-2 show coefficients and t-statistics for our pre-post adoption analyses of audit fees,
using data from two years before and two years after the regulatory cut-off date (Model 1a). Panel B Columns 3-4
show coefficients and t-statistics for our differences-in-differences analyses, using firm-year observations with fiscal
year-end six months before (March to August 2013) and after (September 2013 to February 2014) the regulatory
cut-off in September 2013, as well as the same companies in the prior year (Model 1b, see Section 3 for additional
details). Panel C shows the coefficients and t-statistics from our analyses of the association between audit fees and

51
the total length of the report, the length of the risks’ discussion, and the number of risks (Model 1c). Panel D shows
the coefficients and t-statistics from our analyses of the association between audit fees and materiality thresholds
(Model 1c). In all panels A-D the continuous variables are winsorized at 1 and 99 percent. All variables are as
described in Appendix B. ***, **, * indicate statistical significance from two-tailed tests at 0.01, 0.05, and 0.1,
respectively. All t-statistics are clustered by company.

52
 Table 3: Audit quality analyses
Panel A: Descriptive statistics
POSTi,t =0 POSTi,t =1 Diff. in
Variables Mean Median Std. Dev. Mean Median Std. Dev. Means
DACCRi,t 0.048 0.033 0.049 0.048 0.034 0.046
SIZEi,t 13.418 13.349 1.815 13.453 13.370 1.796
ROAi,t 0.055 0.057 0.091 0.044 0.054 0.106 *
LEVi,t 0.544 0.536 0.234 0.546 0.529 0.228
BIG4i,t 0.933 1.000 0.250 0.937 1.000 0.243
USLISTi,t 0.053 0.000 0.224 0.052 0.000 0.222
TOTAL_NWRDSi,t 7.767 7.749 0.294
RISK_NWRDSi,t 6.478 6.529 0.730
NRISKSi,t 3.886 4.000 1.465
PERCMATi,t 0.667 0.514 0.547
LOGMATi,t 8.233 8.243 1.682

Panel B: Pre-post adoption and difference-in-differences analyses

Dependent variable = DACCRi,t
Variables Coeff. t-stat Coeff. t-stat
POSTi,t -0.001 [-0.27]
ADOPTERi,t 0.009 [1.64]
YR2013i,t 0.004 [0.95]
ADOPTER*YR2013i,t -0.007 [-1.11]
SIZEi,t -0.004*** [-2.86] -0.005*** [-2.90]
ROAi,t -0.104*** [-3.21] -0.136*** [-3.43]
LEVi,t -0.003 [-0.19] -0.014 [-0.99]
BIG4i,t -0.003 [-0.28] 0.004 [0.32]
USLISTi,t 0.010 [1.54] 0.012 [1.49]
Intercept 0.082*** [4.72] 0.077*** [3.58]
N. Observations 1,108 572
Adj. R2 0.115 0.137

Panel C: Cross-sectional analyses of audit quality and length of the report and risks’
section
Dependent variable = DACCRi,t
Variables Coeff. t-stat Coeff. t-stat Coeff. t-stat
TOTAL_NWRDSi,t -0.008 [-0.90]
RISK_NWRDSi,t -0.003 [-0.68]
NRISKSi,t -0.001 [-0.29]
SIZEi,t -0.005** [-2.55] -0.005*** [-2.63] -0.005*** [-2.81]
ROAi,t -0.099*** [-2.73] -0.097*** [-2.72] -0.097*** [-2.68]
LEVi,t 0.005 [0.30] 0.005 [0.28] 0.005 [0.25]
BIG4i,t -0.006 [-0.46] -0.006 [-0.50] -0.007 [-0.52]
USLISTi,t 0.015 [1.49] 0.013 [1.38] 0.014 [1.36]
Intercept 0.152** [2.25] 0.107*** [3.57] 0.097*** [4.17]
N. Observations 533 533 533
Adj. R2 0.110 0.110 0.109

53
Panel D: Cross-sectional analyses of audit quality and materiality
Dependent variable = DACCRi,t
Variables Coeff. t-stat Coeff. t-stat
PERCMATi,t 0.034*** [4.91]
LOGMATi,t 0.021*** [3.70]
SIZEi,t -0.002 [-1.16] -0.024*** [-4.49]
ROAi,t -0.117*** [-3.92] -0.125*** [-4.01]
LEVi,t 0.003 [0.24] 0.009 [0.55]
BIG4i,t 0.014 [1.22] 0.002 [0.13]
USLISTi,t 0.007 [0.81] 0.010 [0.97]
Intercept 0.032 [1.17] 0.179*** [5.25]
N. Observations 537 537
Adj. R2 0.234 0.167

Panel A of this table includes descriptive statistics for the dependent and independent variables for the analyses of
audit quality. Panel B Columns 1-2 show coefficients and t-statistics for our pre-post adoption analyses of audit
quality, using data from two years before and two years after the regulatory cut-off date (Model 2a). Panel B
Columns 3-4 show coefficients and t-statistics for our differences-in-differences analyses, using firm-year
observations with fiscal year-end six months before (March to August 2013) and after (September 2013 to February
2014) the regulatory cut-off in September 2013, as well as the same companies in the prior year (Model 2b, see
Section 3 for additional details). Panel C shows the coefficients and t-statistics from our analyses of the association
between audit fees and the total length of the report, the length of the risks’ discussion, and the number of risks
(Model 2c). Panel D shows the coefficients and t-statistics from our analyses of the association between audit fees
and materiality thresholds (Model 2c). In all panels A-D the continuous variables are winsorized at 1 and 99 percent.
All variables are as described in Appendix B. ***, **, * indicate statistical significance from two-tailed tests at 0.01,
0.05, and 0.1, respectively. All t-statistics are clustered by company.

54
 Table 4: Analyses of investors’ reaction

Panel A: Descriptive statistics

POSTi,t =0 POSTi,t =1 Diff. in
Variables Mean Median Std. Dev. Mean Median Std. Dev. means
ABRETi,t 0.003 0.000 0.046 0.001 0.000 0.033
ABVOLi,t 0.002 0.085 0.994 -0.029 0.003 0.888
SIZEi,t 13.562 13.563 1.803 13.663 13.600 1.758
ROAi,t 0.056 0.056 0.085 0.049 0.054 0.092
LEVi,t 0.530 0.543 0.198 0.539 0.543 0.197
USLISTi,t 0.057 0.000 0.232 0.056 0.000 0.231
IOWNi,t 4.643 4.681 1.264 4.747 4.833 1.287
CHNIi,t 0.003 0.005 0.070 0.003 0.005 0.070
STDRETi,t 0.021 0.019 0.009 0.018 0.016 0.008 ***
MTBi,t 2.712 1.799 3.009 3.544 2.419 3.881 ***
BETAi,t 1.041 0.950 0.735 1.043 1.038 0.697
TOTAL_NWRDSi,t 7.790 7.791 7.772
RISK_NWRDSi,t 6.517 6.554 0.707
NRISKSi,t 3.932 4.000 1.449
PERCMATi,t 0.614 0.494 0.439
LOGMATi,t 8.388 8.294 1.635

Panel B: Pre-post adoption and difference-in-differences analyses

Dependent variable = ABRETi,t Dependent variable = ABVOLi,t
Variables Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat
POSTi,t -0.001 [-0.33] -0.022 [-0.39]
ADOPTERi,t 0.008 [1.19] 0.110 [1.02]
YR2013i,t 0.003 [0.59] 0.246* [1.73]
ADOPTER*YR2013i,t -0.010 [-1.25] -0.173 [-1.03]
SIZEi,t -0.001 [-0.86] -0.002 [-1.01] 0.123*** [2.76] 0.140** [2.37]
ROAi,t 0.002 [0.10] -0.004 [-0.11] 1.811*** [2.98] 1.228 [1.64]
LEVi,t 0.010 [1.04] 0.015 [0.93] -0.036 [-0.17] -0.546* [-1.86]
USLISTi,t 0.002 [0.58] 0.002 [0.43] -0.222* [-1.90] -0.259 [-1.55]
IOWNi,t -0.001 [-0.82] 0.001 [0.24] -0.052 [-1.02] -0.056 [-0.77]
CHNIi,t 0.050* [1.68] 0.076 [1.44] -0.574 [-1.00] -0.159 [-0.19]
STDRETi,t 0.097 [0.32] 0.269 [0.39] 2.204 [0.42] 1.984 [0.24]
MTBi,t 0.000 [0.18] -0.001 [-1.23] 0.002 [0.27] 0.021* [1.69]
BETAi,t 0.001 [0.33] 0.001 [0.26] 0.011 [0.18] 0.018 [0.24]
Intercept -0.003 [-0.17] 0.006 [0.20] -1.881** [-2.53] -1.688* [-1.82]
N. Observations 1,236 610 1,236 610
Adj. R2 0.005 -0.004 0.037 0.027

55
Panel C: Cross-sectional analyses of investors’ reaction and length of the report and risks’ section
Dependent variable = ABRETi,t Dependent variable = ABVOLi,t
Variables Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat
TOTAL_NWRDSi,t -0.000 [-0.07] -0.079 [-0.59]
RISK_NWRDSi,t 0.001 [0.40] -0.056 [-1.09]
NRISKSi,t -0.001 [-0.65] 0.025 [0.96]
SIZEi,t -0.001 [-0.43] -0.001 [-0.52] -0.001 [-0.40] 0.077 [1.46] 0.079 [1.47] 0.066 [1.22]
ROAi,t 0.024 [0.83] 0.025 [0.87] 0.023 [0.78] 2.513*** [2.75] 2.488*** [2.76] 2.623*** [2.90]
LEVi,t 0.012 [1.23] 0.012 [1.20] 0.013 [1.33] 0.307 [1.12] 0.312 [1.15] 0.262 [0.95]
USLISTi,t 0.003 [0.47] 0.003 [0.47] 0.003 [0.55] 0.009 [0.06] -0.003 [-0.02] -0.017 [-0.12]
IOWNi,t -0.004 [-1.53] -0.004 [-1.53] -0.004 [-1.48] -0.039 [-0.60] -0.039 [-0.59] -0.041 [-0.61]
CHNIi,t 0.029 [0.74] 0.029 [0.73] 0.029 [0.75] -0.666 [-0.87] -0.643 [-0.83] -0.681 [-0.89]
STDRETi,t -0.028 [-0.10] -0.032 [-0.12] -0.027 [-0.10] 0.761 [0.10] 0.770 [0.11] 0.487 [0.07]
MTBi,t -0.000 [-0.09] -0.000 [-0.11] -0.000 [-0.10] -0.007 [-0.63] -0.007 [-0.60] -0.007 [-0.67]
BETAi,t 0.006 [1.51] 0.006 [1.53] 0.005 [1.48] 0.018 [0.24] 0.016 [0.22] 0.025 [0.33]
Intercept 0.011 [0.28] 0.005 [0.23] 0.009 [0.39] -0.855 [-0.66] -1.184 [-1.38] -1.426* [-1.73]
N. Observations 605 605 605 605 605 605
Adj. R2 0.009 0.009 0.009 0.046 0.047 0.046

56
Panel D: Cross-sectional analyses of investors’ reaction and materiality
Dependent variable = ABRETi,t Dependent variable = ABVOLi,t
Variables Coeff. t-stat Coeff. t-stat Coeff. t-stat Coeff. t-stat
PERCMATi,t 0.006 [1.42] 0.129 [1.08]
LOGMATi,t 0.007* [1.96] 0.127 [1.43]
SIZEi,t -0.000 [-0.10] -0.006* [-1.80] 0.075 [1.43] -0.044 [-0.44]
ROAi,t 0.022 [0.76] 0.017 [0.59] 2.550*** [2.83] 2.454*** [2.62]
LEVi,t 0.015 [1.49] 0.017* [1.67] 0.419 [1.51] 0.453 [1.61]
USLISTi,t 0.002 [0.27] 0.001 [0.18] -0.004 [-0.03] -0.010 [-0.07]
IOWNi,t -0.004* [-1.70] -0.005* [-1.84] -0.033 [-0.51] -0.041 [-0.63]
CHNIi,t 0.025 [0.64] 0.027 [0.71] -0.765 [-0.98] -0.714 [-0.91]
STDRETi,t -0.061 [-0.22] -0.052 [-0.19] 0.062 [0.01] 0.362 [0.05]
MTBi,t -0.000 [-0.46] -0.000 [-0.68] -0.014 [-1.23] -0.016 [-1.38]
BETAi,t 0.005 [1.51] 0.005 [1.46] 0.018 [0.24] 0.016 [0.21]
Intercept -0.003 [-0.13] 0.029 [1.12] -1.549* [-1.89] -0.892 [-0.94]
N.
Observations 611 611 611 611
Adj. R2 0.011 0.015 0.047 0.0485

Panel A of this table includes descriptive statistics for the dependent and independent variables for the analyses of
investors’ reaction. Panel B Columns 1-2 examining abnormal returns and Columns 5-6 examining abnormal
volume on the report filing date, show coefficients and t-statistics for our pre-post adoption analyses of audit quality,
using data from two years before and two years after the regulatory cut-off date (Model 3a). Panel B Columns 3-4
examining abnormal returns and Columns 7-8 examining abnormal volume on the report filing date, show
coefficients and t-statistics for our differences-in-differences analyses, using firm-year observations with fiscal year-
end six months before (March to August 2013) and after (September 2013 to February 2014) the regulatory cut-off
in September 2013, as well as the same companies in the prior year (Model 3b, see Section 3 for additional details).
Panel C shows the coefficients and t-statistics from our analyses of the association between investors’ reaction and
the total length of the report, the length of the risks’ discussion, and the number of risks (Model 3c). Panel D shows
the coefficients and t-statistics from our analyses of the association between audit fees and materiality thresholds
(Model 3c). In all panels A-D the continuous independent variables are winsorized at 1 and 99 percent. All variables
are as described in Appendix B. ***, **, * indicate statistical significance from two-tailed tests at 0.01, 0.05, and
0.1, respectively. All t-statistics are clustered by company.

57
Table 5: Risk classification for the first and second year of adoption of the new rules

Percentage of companies reporting each type of risk

Year t (First year of adoption) Year t+1 (Second year of adoption)
Revenue recognition 71% Revenue recognition 61%
Impairment of goodwill and long-term assets 36% Business combinations 39%
Internal controls 35% Impairment of goodwill and long-term assets 39%
Business combinations 34% Tax accounting 36%
Tax accounting 33% Receivables and provisions 30%
Receivables and provisions 28% Pensions 26%
Pensions 25% Inventory 25%
Inventory 23% Intangibles and long-lived assets 24%
Intangibles and long-lived assets 23% Other presentation and disclosure issues 19%
Other presentation and disclosure issues 15% Other assets and liabilities valuation 18%
Other assets and liabilities valuation 14% Going concern considerations 11%
Going concern considerations 9% Litigation provisions and disclosure 8%
Discontinued operations and restructuring 7% Discontinued operations and restructuring 7%
Investments valuation 7% Internal controls 6%
Litigation provisions and disclosure 7% Investments valuation 6%
Other issues 6% Other issues 6%

This table shows the percentage of companies reporting each risk type in years t and t+1 after the adoption of the new rules. We classify individual risks in 16
categories, based on our interpretation of the main issue by reading the risk headings and disclosures.

58