Accepted Manuscript

Title: The Value Relevance of Risk Disclosure in Annual

Reports: Evidence from MENA Emerging Markets (RIBAF
350)

Author: Néjia Moumen Hakim Ben Othman Khaled

Hussainey

PII: S0275-5319(15)00009-4
DOI: http://dx.doi.org/doi:10.1016/j.ribaf.2015.02.004
Reference: RIBAF 350

To appear in: Research in International Business and Finance

Received date: 30-1-2015

Accepted date: 2-2-2015

Please cite this article as: Moumen, N., Othman, H.B., Hussainey, K.,The Value
Relevance of Risk Disclosure in Annual Reports: Evidence from MENA Emerging
Markets (RIBAF 350), Research in International Business and Finance (2015),
http://dx.doi.org/10.1016/j.ribaf.2015.02.004

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 Value Relevance of Risk Disclosure in Annual Reports: Evidence from MENA
Emerging Markets

Abstract

t
ip
Our paper aims to examine whether voluntary risk disclosure in the annual report contains
value-relevant information for investors to predict future earnings. We used a large-scale

cr
sample firms from MENA emerging markets. Our sample includes 809 year-observations for
the three year period, 2007 to 2009. Our study offers two significant contributions. First, we

us
found a positive relationship between voluntary risk information and the market’s ability to
anticipate two-years ahead future earnings change. The positive association provides us with
the first empirical evidence of the usefulness of risk disclosure in annual reports. Second, we

an
found that the level of proprietary costs tends to moderate the perceived relevance of risk
information, thereby making investors rely on another source of information in forecasting
M
future earnings change.

Keywords: Risk Disclosure, Value Relevance, Future Earnings, Proprietary Costs, MENA
emerging markets.
d
te

JEL Classification: G15, G32, G170, M41

p
ce
Ac

1
Page 1 of 46
1. Introduction

Accounting information has been considered as one of the fundamental pillars contributing to
good corporate governance (Healy & Palepu, 2001). As owners delegate most of their
decision making responsibilities to managers, agency conflicts arise between insiders and
outsiders. Basically, managers are more informed about corporate business than investors.

t
The availability of information is essential to reduce the information asymmetry between both

ip
sides. It plays an important role in the individual and corporate decision making process
(Bogdan et al., 2009). Firm’s reporting activity is expected to provide users with a great

cr
insight into the amounts, timing, and uncertainty of its future cash flows (FASB, 2010) and
adequate disclosure helps to ensure the efficiency of capital markets.

us
With the increasing complexity of business operations, stakeholders’ needs of information
have become more sophisticated. Society and standard setters, now, require a higher level of

an
information, specifically, from publicly traded companies. The unexpected successive
collapses of large firms with a biased image of being low risk and highly predictable (e.g.,
M
Enron) put into question the scope and the usefulness of corporate disclosure. This situation
compelled accounting regulators to rethink the set of requirements for financial reporting.
Meanwhile, there were considerable calls by academics, accountancy bodies and regulatory
d

agencies for enhanced and understandable information about corporate prospects and
te

business uncertainties (Deumes, 2008). As a result, narrative risk disclosure became

increasingly required in periodic reports by national GAAPS or formal codes of best practice
p

in corporate governance worldwide. Many developed countries such as USA, Canada,

Germany, UK, Austria and Finland… etc, mandate financial risk communication, however its
ce

current regulatory framework reveals a piecemeal approach, focuses, predominantly, on

market risk associated with the use of derivatives (e.g., FAS 119, FAS 133, IAS 32, and IAS
Ac

39). In this regard, a wider set of risk reporting is hitherto not driven by rules and is offered
on a voluntary basis.

The regulatory reforms supported by the stead debate among the professional accountancy
bodies (e.g. ICAEW) and standard setters over the lack of transparency and clarity in
corporate risk reporting triggered a research stream in accounting and related fields about
firms’ risk exposure, risk management, and their monitoring processes around the world
(Dobler et al., 2011).

2
Page 2 of 46
The reason for this interest is that information on the risks and rewards of a firm’s business is
becoming crucial for well-informed investors and for an accurate corporate valuation and
investment decision (Fuller & Jensen, 2002). Narrative risk disclosure is expected to narrow
the information gap between management and stakeholders about business uncertainties and
opportunities. It may decrease the firm’s perceived risk because enhanced information on
corporate risk should result in a better assessment of the firm’s future performance. Managers

t
ip
also benefit from their transparency about the underlying risks in their strategic goals. They
may signal their quality in identifying and managing risk, hence differentiating themselves

cr
from other corporate managers who may be perceived to measure and report on risk less
effectively (Elshandidy et al., 2013).

us
Despite the recent significant focus on examining risk disclosure, information on firms’ risk
exposure is still one of the most ambiguous areas of corporate disclosure strategy. In

an
particular, we know too little about the value relevance of risk disclosure, though the recent
debate about its inadequacy. Some scholars argue that, as it is, current risk disclosure does not
contain reliable information, for it is, in most cases, left to managers’ discretion, (Schrand &
M
Elliot, 1998), others point to its lack of progress (e.g. Abraham & Shrives, 2014), and still
others consider it unhelpful and devoid of real meaning (e.g. Campbell & Slack, 2008;
d

Davies, Moxey, & Welch, 2010; Moxey & Berendt, 2008). These criticisms may be argued to
be groundless, for they are not supported by any empirical evidence. The few existing studies,
te

investigate, mainly, the information content of mandatory risk disclosures provided in line
with the Securities and Exchange Commission's requirement FRR No.48. Firms are required
p

to provide quantitative and qualitative information about exposure to market risk and to
ce

disclose how they account for derivatives (e.g. Rajgopal, 1999; Linsmeier, 2002; Jorion,
2002; Liu et al., 2004; Lim & Tan, 2007; Pérignon & Smith, 2010). Kravet & Muslu (2013)
Ac

fill this gap in the literature by addressing how users’ risk perceptions change around the
filing dates in response to changes in textual risk disclosures. Their study provides evidence
on the informativeness of narrative risk disclosures. Following Kravet & Muslu (2013),
Campbell et al. (2014) show that risk factor disclosures are not boilerplate, but instead
meaningfully reflect the risks a firm faces. They decrease the information asymmetry among
firms’ shareholders and are incorporated into stock prices. Conversely, Bao & Data (2014)
find that around two-thirds of risk types lack of informativeness and have no significant
influence.

3
Page 3 of 46
It is important then to take these mitigated results one step further and to document more
evidence on whether narrative risk disclosures convey useful information to investors. We
explore the following research questions: First, does voluntary risk disclosure in the annual
reports impact the return future earnings relationship? Second, does the level of proprietary
costs moderate investors’ ability to anticipate future earnings change?

t
We address these research questions for several reasons. First, recent evidence (e.g. Lehavy et

ip
al., 2010) on the informativeness of risk disclosure are based on US data where specific
information on risk factors (market risk) is mandated. Voluntary risk information covers,

cr
though, a broader spectrum of risk factors (e.g. operational and strategic risks). Their
informational content has been recently emphasized by Kravet & Muslu (2013) and Bao &

us
Data (2014). The differences in the firms’ sample and in the research design do not allow for
a direct comparison between our study and their findings. Unlike these event studies, which

an
concentrate on the market’s reaction to mandated risk information over a short time interval,
we conduct an association study, which allows investors and analysts to interpret the reported
risk exposures over a long time interval. We also explore whether risk disclosure is credible
M
within a quite different context (heavily concentrated shareholding and a weak investor
protection) with a growing importance to foreign investors. Our international investigation of
d

risk disclosure practice of MENA firms is justified by the relatively heterogeneous but
increasingly globalized environment. Second, our paper answers the call of Dobler et al.
te

(2011) regarding the current gaps in the body of risk reporting literature in countries with
weak risk reporting legislation. The approach towards risk disclosure in MENA emerging
p

markets relies largely on encouraging rather than requiring firms to reveal risk information
ce

(risk reporting regulations is limited to the financial risk factor) differing from other contexts
where the usefulness of risk disclosure has been tested recently. This specific institutional
Ac

milieu suggests that there might be little chances for risk information to alleviate the
information asymmetry between managers and market participants. Third, given that market
forces and management incentives are likely to play a major role in shaping the amount of
risk disclosures and their informativeness (Dobler, 2008), firms may consider risk disclosure
harmful to their competitive position. Consequently, proprietary costs are likely to moderate
the value relevance of risk information (Verrechia, 1983).

4
Page 4 of 46
Our study contributes then to the extant risk disclosure literature along several dimensions.
Given the recent evidence on the usefulness of mandatory risk factors, it is argued that the
content of risk information shall satisfy an early-warning function and more specifically a
forecasting function for outsiders. Our study is a way forward in seeking an empirical
assessment of the relevance of risk information conveyed on a voluntary basis. Making
generalizations, based on the extant evidence on the usefulness of risk disclosure might go

t
ip
against a strong strand of literature that argues that accounting practices (measurements and
disclosures) within a specific context (such as MENA countries) should reflect its underlying

cr
environmental factors (e.g., Nobes, 1998; Dober et al., 2011). Second, contrary to general
disclosure studies, little is known about the impact of competition on the informativeness of

us
risk disclosure. The proprietary costs theory provides a way of explaining and understanding
the current problematic state of risk reporting and notably its reliability. Our study suggests
that narrative risk disclosure in corporate annual reports is an important source of information

an
for market participants in making investment decisions. Our results help regulatory bodies and
market participants, to better understand the extent to which, and in what manner, risk
M
disclosure influences the investors’ ability to anticipate future earnings in an environment
where overall country-level investor protection is relatively poor. Regulators should not also
acknowledge that corporate risk assessments include proprietary information, which
d

companies tend to withhold to reduce competition, when rethinking the set of requirements
te

for financial disclosure.

The paper proceeds as follows. Section 2 discusses the research background, reviews relevant
p

literature and develops our hypotheses. Section 3 discusses the research methodology. Section
ce

4 presents the empirical analysis. Section 5 provides additional robustness tests. Section 6
concludes.
Ac

2. Background, literature review and hypotheses development

2.1. Background

Equity markets in the Middle East and North Africa region (MENA) provide an excellent case
study on the financial reporting environment and the efficiency of emerging stock markets for
the following reasons. Firstly, countries in this region have embraced a series of reforms in
their economic environment, followed by extensive efforts to diversify their economies
(massive privatization plan) and develop their stock exchanges.

5
Page 5 of 46
Most MENA countries managed to revitalize their long-standing markets, such as Egypt,
Saudi Arabia, Kuwait and Amman or to establish new markets, such as Dubai and Abu Dhabi.
Stock exchanges in this region became very meaningful to the world’s economy and their role
in the international financial system experienced a considerable increase. This involved the
establishment of a new legal framework with new financial disclosure requirements being
imposed upon listed companies such as securities exchange laws, international financial

t
ip
reporting standards (IFRS) and corporate governance codes (Ben Naceur et al. 2008; IFC &
Hawkamah, 2008; Al-Akra et al., 2010a,b; Samaha & Dahawy, 2011). The compliance with

cr
these international requirements for best practices in the region was intended to enhance the
levels of disclosure and transparency, to attract more foreign capitals into the market, and to

us
increase investors’ protection through provisions forbidding unfair market practices (Hassaan,
2013). These series of changes, supported by international institutions reports (e.g., IFC and
Hawkamah, 2008; ROSC, 2009) claiming a de jure but not a de facto compliance with the

an
requirements of newly adopted laws and regulations in the region, stimulates empirical
examination of the outcomes of such reforms (Hassaan, 2013). Secondly, most MENA
M
countries have a civil law legal system based on French legal family. They are likely to opt
for conservative and rigid accounting system whereby transparency and the average of
investor protection are lower than common law countries (Gray, 1988; Ben Othman &
d

Zeghal, 2008). The cultural context within the region is by far characterized by a preference
te

for secrecy which is encouraged by weak law enforcement and low non compliance costs, if
any (Al-Akra et al., 2009; Al-Omari, 2010; Ismail et al., 2013; Samaha & Dahawy, 2011;
p

Hassaan, 2013). MENA stock markets exhibit in this respect a weak-form efficiency and a
ce

capital market fragmentation due to the poor-quality of information. There exist relatively
limited corporate disclosure requirements and barriers to efficient information transmission
through these markets. Investors are likely to be “ill-informed” and have often access to
Ac

unreliable financial information (Harrison & Moore, 2012). Thirdly, the pressures from
international organizations such as the World Bank, the International Monetary Fund and
other institutional investors on the governmental agencies in MENA emerging countries led to
mandating laws and governance mechanisms that cope with international best practices
without taking into account the need of enhancing the financial literacy of investors and
spreading sufficient awareness among different stakeholders affecting and being affected by
corporate disclosure activity.

6
Page 6 of 46
Accordingly, there is a need for more research to provide evidence on whether accounting
information is relevant in MENA emerging markets given the possible contradiction between
newly introduced governance notions and the native cultural values such as secrecy within the
MENA society.

2.2. Literature Review

t
ip
A primary motivation of market based accounting research is to seek evidence on how
accounting information is perceived as value relevant by a wide range of users in making

cr
economic decisions. The extent to which market participants are likely to impound
information about corporate prospects into share prices, (e.g., Collins et al. 1994; Kothari &

us
Sloan 1992; Warfield & Wild, 1992) has been of particular interest during the last two
decades. Falls within this framework, a series of accounting research (e.g. Hussainey et al.,

an
2003; Hossain et al, 2006; Schleicher et al, 2007; Hussainey & Walker, 2009; Athanasakou &
Hussainey, 2014) that addressed the cross-sectional variation in price informativeness with
respect to future earnings. These studies identified a number of corporate attributes and
M
incentives (e.g. Profitability, growth, earnings quality and ownership structure) as well as the
level of forward-looking disclosure as the main reason for such variation. They suggest that
forward-looking information enhances investors’ ability to predict future earnings, and helps
d

them make informed investment decisions (Hussainey et al., 2003). Empirical evidence
te

(Hussainey et al., 2003; Hussainey &Walker, 2009; Hussainey & Mouselli, 2010;
Athanasakou & Hussainey, 2014) show that the level of narrative forward-looking
p

information in the management discussion and analysis section of annual reports reduces
ce

market uncertainty about corporate future earnings. They also indicate that any revealed news
that changes market expectation about future performance will be impounded in stock prices.
Ac

Recently, along with the increasing sophistication of the international business environment,
stakeholders have asked listed companies to reveal additional information about their
uncertainties and outlooks. In response, public companies have been improving their
disclosure about the risks faced and their expected impact on corporate profitability (Beretta
& Bozzolan, 2004). Risk disclosure gained accordingly interest in financial reporting activity,
regulation, and international research. Kravet & Muslu (2013) argue that narrative risk
disclosures differ from other corporate reporting and notably forward looking information, for
they inform the users about the range of future performance rather than the level of future
performance.

7
Page 7 of 46
They are likely to elucidate, but do not necessarily unravel corporate uncertainties.
Informative risk disclosures are believed to increase market efficiency (Dietrich et al., 2001)
and to influence (increase/decrease) users’ risk perceptions, i.e., the range of users’ forecast of
future performance as well as users’ confidence in their predictions (Kravit & Muslu, 2013).

2.2.1 The effect of voluntary risk disclosure on share price anticipation of future earnings

t
ip
Notwithstanding recent regulatory efforts (IFRS7, Item 1A & Item7 US-GAAP, CICA
guidance in Canada, FRS7 and FRS32 in Malaysia and Singapore respectively… etc..) to

cr
require or encourage corporations to reveal their assessments about risks and uncertainties in
their annual reports, risk disclosures remains largely discretionary and the value relevance of

us
such disclosures, to our knowledge, is largely unexplored. Li (2008) focuses on the
implications of risk sentiment in annual reports (i.e., its emphasis on risk and uncertainty) for

an
future earnings and stock returns. He documents that an increase in risk sentiment (as
captured by count of words ‘risk’ and ‘uncertainty’) is associated with lower future stock
returns and poor future earnings, suggesting that stock returns do not fully reflect the
M
information about future profitability in the narrative sections of annual reports.

Kothari et al. (2009) examine some economic consequences of risk disclosure disseminated
d

through three main channels. In particular, they analyze the impact of risk information
te

provided by corporations, financial analysts, and business press, on the firms’ capital market
environment. They show that favorable total risk disclosures are associated with a significant
p

decrease in the firm’s cost of capital, stock return volatility, and dispersion in analysts’
earnings forecasts. In contrast, unfavorable total risk information is accompanied by a
ce

significant rise in the cost of capital, stock return volatility, and analysts’ earnings forecast
dispersion. Kothari et al. (2009) emphasize also that management’s favorable disclosure does
Ac

not significantly impact the firm’s cost of capital, which suggests that companies’
communications with investors are not credible.

Kravet & Muslu (2013) examine whether annual changes in narrative risk disclosures in 10-K
filings have an impact on users’ risk perceptions, as measured by investor and analyst
activities within the immediate two months before and after the 10-K filings. Results reveal
that annual changes in textual risk disclosures are significantly and positively associated with
changes in the daily stock return volatility, the trading volume, the dispersion of outstanding
forecasts and the volatility of forecast revisions.

8
Page 8 of 46
Kravet & Muslu (2013) provide hence evidence that risk disclosure reveals ‘new’ information
about corporate risk and uncertainty; corporate risk disclosures appear, however, to introduce
unknown future contingencies and risk factors rather than only update information about
known risks. Risk information is likely to widen the range of users’ predictions of future
performance and to decrease investors ‘confidence levels in their predictions.

t
Campbell et al. (2014) investigate similarly whether mandatory risk disclosures are

ip
informative about corporate risk factors, and whether investors impound this public

cr
information into firm value. Their findings suggest that risk reporting informs about the risks
facing the business, reduces the information asymmetry, and is timely incorporated into stock

us
prices. Particularly, they find that qualitative risk factor disclosures in the 10-K filings are not
boilerplate as they are significantly associated with pre-disclosure firm specific risks. They
also show that textual risk information is associated with low bid-ask spreads, high beta and

an
stock return volatility following the disclosure of such information. These results suggest that
managers provide useful risk factor disclosures and investors impound this information into
market values.
M
Finally, Bao & Datta (2014) propose a new method to identify the types of risk information in
section 1A of the 10-K form and examine whether the extent of these textual risk disclosures
d

affects users' risk perceptions. Inconsistent with Kravet & Muslu (2013) and Campbell et al
te

(2014), they find that around two-thirds of risk types (22 out of 30 risk types) are lacking of
informativeness and have no significant influence on investors’ risk perception. Such result
p

provides support to earlier claims that risk disclosures in 10-K forms are by and large
ce

boilerplate. Managers tend to reveal generic risk factor disclosures that are not useful in
assessing corporate uncertainties and future cash flows. Bao & Datta (2014) show indeed that
few types of risk information are likely to be informative. In particular, the disclosure of three
Ac

types of systematic and liquidity risks (systematic, funding and credit risks) increases the risk
perceptions of investors, whereas the other five types of firm/industry specific risks (i.e.
human resources, infrastructure, and regulation changes) decrease the stock return volatility.
Bao & Datta (2014) explain that these contradictory findings with recent researches are due to
their method that seeks to discover one risk type specific impact on stock return volatility
rather than mix them together.

9
Page 9 of 46
Overall, these recent studies lend support to the informational content of risk disclosure in
annual reports. Based on the above discussion, we hypothesize that:

H1. Share price anticipation of future earnings is positively associated with the level of firm’s
voluntary risk disclosure.

t
2.2.2. The effect of proprietary costs and voluntary risk disclosure on share price anticipation

ip
of future earnings.

cr
Marshall & Weetman, (2007) suggest that while most firms are likely to own or to implement
risk management systems, they may be discouraged to disclose related information when they

us
feel it is commercially sensitive. This is mainly because “outside parties” such as corporate
opponent, may use the information in ways that it damages firms’ competitive position
(Cormier et al., 2005). Managers should trade off the benefits of expanded disclosure against

an
the costs of disclosing potentially damaging information (Abraham & Shrives, 2014).
However, Skinner (1994) and Healy et al. (1999) believe that managers should incur some
‘upfront’ costs by disclosing some proprietary information in order to enhance their reputation
M
as a credible discloser. To face this dilemma, Abraham & Shrives (2014) suggest that
corporations tend to shape their disclosure activity (e.g. restrict or provide boilerplate) in
d

order to reduce any proprietary costs. Solomon et al. (2011) note also that managers may
prefer to disclose sensitive information to large shareholders or potential investors through
te

other, more private, reporting mechanisms such as special meetings.

p

Prior financial disclosure literature demonstrates that managers’ decision to voluntarily

ce

disclose (non) financial information (Dedman & Lenox, 2009; Nichols, 2009), segment
information (Harris, 1998; Botosan & Harris, 2000; Piotroski, 2001; Leuz, 2004; Prencipe,
2004), and earnings forecasts (Bamber & Cheon, 1998; Clarkson et al, 1994; King &Wallin,
Ac

1995; Li, 2010) as well as corporate disclosure rankings such as the AIMR (Shin, 2001) are a
negative function of competitive costs.

Within the risk disclosure literature, Abraham & Shrives (2014) find that managers do not
signal good risk management systems via comprehensive disclosures. Managers prefer
general nonspecific information, which could apply to any company within the same industry
consistent with the proprietary costs theory. The authors also show that companies tend to
manipulate disclosures to arrange themselves and do not wish to disclose sensitive
information that might be exploited by others including industry competitors.

10
Page 10 of 46
Corporations consider risk information as something, which is useful for internal management
only. As an alternative, they may choose to disclose information privately and to provide
irrelevant risk reporting to fill the disclosure gap. Some decision usefulness studies came to
the same conclusion. Hussain et al. (2006) suggest that the product market competition is
likely to reduce the value relevance of voluntary disclosure. Findings indicate that firms with
higher proprietary costs choose to disclose less future earnings information suggesting that

t
ip
share price informativeness with respect to future earnings is lower for firms with high
proprietary cost because the information environment is less rich. Ali et al. (2014) argue that

cr
companies in more concentrated industries face a greater risk of revealing company
secrets to their rivals through discretionary disclosures. Such disclosure would be harmful

us
for the corporate competitive position because rivals would take actions to enhance their
profits at the expense of the firm. Results indicate that the firms’ quality disclosure ratings are
negatively related to industry concentration. These results suggest that firms in more

an
concentrated industries (i.e. higher proprietary costs) disclose less reliable voluntary
information.
M
Based on the above discussion, we formulate the following hypothesis:

H2. The association between voluntary risk disclosure and share price anticipation of future
d

earnings is weaker for firms with high proprietary costs.

te

3. Methodology
p

In this section we discuss our approach in measuring the level of voluntary risk disclosure and
proprietary costs, the sample selection and our regression model.
ce

3.1 Measuring voluntary risk disclosure

Ac

To assess the level of risk disclosure across different countries, we use a traditional approach
in disclosure studies that is content analysis. Krippendorf (1980) believes that content analysis
ensure repeatability and valid inferences from data according to their contexts. We collect
data from corporate annual reports which, despite the reported decline in the value relevance
of some financial information (Francis & Schipper, 1999; Collins et al, 1994), still provide
investors with information that explains accounting numbers, sketches and present firms’
future prospects (Beattie et al., 2002).

11
Page 11 of 46
We perform the risk disclosures’ analysis of the sample companies in all the narrative sections
in the annual reports and notably the management discussion and analysis (MD&A) section
and board reports (BR). BR and the MD&A section are meant to provide investors with a
greater understanding of corporate business, corporate strategy and performance, as well as
how it manages risk and capital resources. MD&A section serves as a mean for managers to
reveal “their perspectives of the firm to investors, such as why earnings have changed, what

t
ip
liquidity needs the firm faces, what capital resources have been or are planned to be used,
what material market risks the firm is exposed to” (Brown & Tucker, 2011, p. 310) and what

cr
are the future trends that may affect future operations. Such information would help investors
to assess firms’ past performance and current financial condition as well as to predict future

us
cash flows (Brown & Tucker, 2011).

Researchers have at times employed a variety of different approaches to coding and

an
measurement units, namely words, sentences and proportion of pages. Consistent with recent
risk disclosure literature (Beretta & Bozzolan 2004; Linsley & Shrives 2006; Amran et al.
2009; Dobler et al, 2011; Elzahar & Hussainey, 2012) we use the number of risk-related
M
sentences as a measure of risk disclosure levels. Even though Unerman (2000) argues that
words can be counted with a high degree of accuracy, they cannot be coded into different risk
d

categories without reference to the sentence. That is “the meaning of a word typically
te

depends on its syntactical role within a sentence (Krippendorff, 2004 p. 101). Particularly
in our international study, annual reports are disclosed and/or available in different languages
p

(Arabic, French and English) depending on the country of investigation. For instance, less
bias can be expected when referring to sentences compared to words.
ce

A further requirement from a content analyst is to specify how to construct the coding
instrument. There are few coding grids for risk disclosure activity due to the limited academic
Ac

research (e.g Beretta & Bozzolan, 2004; Linsley & Shrives, 2006). In our study, we rely on
the coding instrument used by Linsley & Shrives (2006) and Kajüter (2001) which is
developed by one of the professional accountancy firms. Since our focus is on corporate
voluntary risk disclosure, we remove mandatory risk categories (financial risk factor) from the
initial coding grid. Most MENA emerging markets adopted the international financial
reporting standards (IFRS) and are providing the mandatory information about their market
risk (currency, liquidity and credit risks).

12
Page 12 of 46
The disclosure index (appendix A1.) reflects five categories1 of voluntary risk information:
operations risk, empowerment risk, information processing and technology risk, integrity risk
and strategic risk whereby 32 items were identified.

Following Linsley & Shrives (2006), we adopt a broad definition of risk to identify risk
disclosures. We code risk disclosures any sentence that informs the reader about “any

t
opportunity or prospect, or of any hazard, danger, harm, threat or exposure, that has already

ip
impacted or may impact upon the company, as well as the management of any such

cr
opportunity, prospect, hazard, harm, threat, or exposure”. Similar to Linsley & Shrives (2006)
and Elzahar & Hussainey (2012), we have to adhere to some decision rules. First, because the

us
definition of risk is broad, disclosures have to be explicitly mentioned and not merely implied.
Moreover, we code risk disclosure sentence any disclosure that is repeated each time it is
discussed. Any sentence with more than one possible classification is classified into the

an
category most emphasized within the sentence. We then generate an aggregated score for risk
disclosure for each firm by counting the number of risk-related sentences in corporate annual
reports. It should be noted that the instrument we use captures risk disclosure quantity and not
M
necessarily its quality (Botosan 2004).

Since content analysis is inevitably subjective, the coding method has to be reliable and
d

agreement procedures should be adopted to draw valid conclusions (Bowman, 1984).

te

Krippendorff (2004) contends that reproducibility is the most important interpretation of

reliability in content analysis. To ensure reproducibility, we use one single coder to perform
p

content analysis. Bowman (1984) argues that to increase confidence that the interpretation of
ce

a written document correspond to objective reality, there should be more than one person to
read and code the written document. Accordingly, we use an experienced researcher familiar
with the technique of content analysis to code, independently, a sub sample of 5 firms. The
Ac

subsamples are selected randomly from the yearly pool of observations from 2007 to 2009
for a total of 15 observations. Before we begin the analysis, we discuss risk disclosure and
research objective with the coder to familiarize him with relevant literature. Additionally, we
provide the coder with a set of rules to replicate the pretest coding and in the process, we
clarify and refine the rules as needed. As the coder and the researcher independently code the
initial sample, we use tests of inter rater reliability to check for consistency in coding, a proxy
for accuracy.

1
We identify the sentences related to these categories of risk based on the Institute of Chartered Accountants in
England and Wales (ICAEW) (2002) definition of risk factors (further information is provided in appendex A2.).

13
Page 13 of 46
There's no general consensus in the literature on how to do and report inter-coder reliability
tests. In our study, consistent with Hayes & Krippendorff (2007) and Krippendorff (2010), we
rely on Krippendorff’s alpha test, which is considered to be the most appropriate test of inter
rater reliability. The test generates a Kalpha of 0.889, a satisfactory level of inter-rater
reliability for this intra-class agreement coefficient. It is customary to require Kalpha= 0.80 as
the cut off point for a good reliability test, with a minimum of 0.67 (Krippendorff, 2004).

t
ip
3.2. Measuring proprietary costs level

cr
A reliable measure of proprietary costs is still challenging scholars. Although a number of
proxies of proprietary costs were proposed, no proxy has gained general acceptance. Prior

us
analytical models estimate competitive costs, conditioned on how competition is defined. One
group of studies model competition in the context of an entry game where a firm decides

an
whether to enter a particular new product market (e.g., Darrough & Stoughton, 1990) while
the others model competition in the context of a post-entry game (e.g., Darrough, 1993).
According to Monk (2011), the type of competition has an impact on the disclosure
M
equilibrium outcomes. The difference between product market competition and the threat of
entry has been shown to be enough to change the effect of competition on voluntary
disclosures.
d
te

The industry concentration ratio is the most widely used proxy of the product market
competition in the accounting literature (Verrecchia, 1983; Stiglitz, 1987; Sutton, 1991;
p

Harris, 1998; Botosan & Stanford, 2005; Verrecchia & Weber, 2006; Ali et al., 2014). In
these studies, it is believed that highly concentrated industries are less competitive. This involves a
ce

higher barrier to entry and a lower proprietary information costs, which leads to a higher level of
disclosure. Ali et al. (2014) and Dedman & Lennox, (2009) suggest however that relying on
Ac

the industry concentration variable is problematic. They infer that high industry concentration
does not necessarily correspond with low competition intensity and that an industry may have
relatively few participants, but low barriers to entry. The use of industry concentration ratio
may lead then to incorrect conclusions and inferences, especially that it assumes that all
companies within a given industry face the same level of competition.

For the purpose of our research we estimate the proprietary costs variable (PCit) using a firm
specific measure that is the “Profit Margin ratio” consistent with Cohen (2004), Dedman &
Lennox (2009) and Tang (2010).

14
Page 14 of 46
According to these studies, corporations with high profit margins attract future competition
and exhibit higher forewarning of potential entrants. In fact, high profit margin reflects a
corporate successful strategy that yields a considerable competitive advantage compared to
other existing rivals. Such competitive advantage can be only maintained if competitors
cannot reproduce the company’s activities. Consequently, the presence of a competitive
advantage should discourage disclosure of business opportunities and expansion plans. Highly

t
ip
profitable company has then an incentive to hide proprietary information in order to prevent
adverse reactions from less successful rivals.

cr
3.3. Sample and Data

us
The aim of this study is to investigate the usefulness of voluntary risk disclosure in a number
of Middle Eastern and North African countries. As far as we know, the countries studied in
this research have not been examined extensively, despite the growing importance of the

an
region with respect to commerce and foreign direct investment.

Our original sample covers twelve MENA emerging markets. We apply some filtering rules
M
to ensure data availability and sample homogeneity. Consequently, we drop Israel from our
initial sample because it is considered as the only country with a common law system in the
d

region while the other countries have mainly civil law system. We also drop Bahrain and
Qatar because their capital markets include mostly financial and investment corporations.
te

Our final filtered sample has only nine MENA emerging capital markets: Egypt, Jordan,
Kuwait, Morocco, Oman, Saudi Arabia, Tunisia, Turkey and UAE. The companies in our
p

countries’ sample are periodically listed from 2007 to 2012.

ce

We choose this period of analysis for the following main reasons. One, MENA emerging
countries witnessed a growing GDP per capita during the last decade, which remained
Ac

relatively notable and close over the years before and after the international financial crisis. In
fact, except for 2009 (-0.048%) and 2012 (2.229%) the GDP per capita growth is on average
above the 3 % for 2007, 2008, 2010 and 2011, respectively. Foreign direct investments
jumped also from $18.36 billion in 2005 to more than $33 billion in 2007 and 2008 before
falling slightly by nearly 16% and 13% during the next years respectively (2009, 2010 and
2011). Two, Market capitalization exhibited a considerable growth reaching 91.32% of GDP
in 2007 and represented on average 55.29 % of MENA region GDP between 2007 and 2012.
This growth is considered the most important rise during the last decade compared to Latin
America (46.61% of GDP) emerging markets.

15
Page 15 of 46
Three, the notable growth of the capitalization ratio coincided with a sizable number of listed
companies reaching 2163 by 2012 and representing approximately two times the number of
listed companies in Latin America. Four, the total value traded as a share of GDP (which
gives the value of stock transactions relative to the size of the economy) is on average 41.5%,
representing about 4.394% of stocks traded in East Asia and The Pacific and nearly 82.21% of
stock traded in Latin America (WDI, World Bank, 2012).

t
ip
The firms included in our sample, have to satisfy the three conditions: First, it has to belong to
a non-financial sector. Financial firms such as banks, insurance firms and investment firms

cr
are excluded because their reports are not comparable to those of non-financial firms. Second,
this study focuses on annual reports and not other media of financial communications such as

us
interim reports. Third, the non-financial firm, has to have at least one annual report, from
2007-2009. We choose 2009 as the end year for the study because the level of corporate risk

an
disclosure is linked to share price anticipation of earnings and accounting and return data are
required for at least three years ahead (Year 2012).
M
We match the selected companies with the Thomson one database codes from which we
gather financial information such as stock prices, earnings per share, asset growth, and the
profit margin ratio. Firms that do not have Thomson one code are excluded from our sample
d

because they have no accounting and return data. After performing these series of sample-
te

filtering we ended up with a sample of 809 firm-year observations for the period 2007 to
2009.
p

The following table 1 summarizes the composition of our sample by country

ce

Table1: Sample composition by Country

Country/Year 2007 2008 2009 Panel
Ac

Egypt 19 24 25 68
Jordan 38 34 50 122
Kuwait 17 18 22 57
Morocco 16 15 17 48
Oman 47 50 49 146
Saudi Arabia 27 29 31 87
Tunisia 24 23 9 56
Turkey 50 52 52 154
UAE 19 23 29 71
Total 257 268 284 809

16
Page 16 of 46
3.4. Regression model specification

Our study follows the recent trend in the accounting studies (e.g. Hussain et al, 2006; Gelb &
Zarowin, 2002; Hussainey & Walker, 2009; Lundholm & Myers, 2002; Schleicher et al 2007)
to measure the value relevance of voluntary risk disclosure as well as the effect of proprietary
costs on the usefulness of such information. We use Collins et al. (1994) future earnings

t
response coefficient model (FERC) to assess the mix of revealed current and future earnings

ip
news in current stock returns.

cr
Collins et al. (1994) applied the following specification:
N N
Rt  b0  b1 X t   bk 1 X t k   bk  N 1Rt k  b2 N 2 EPt 1 b2 N 3 AGt

us
k 1 k 1

Where: Rt: stock return for year t. Rt+1, Rt+2, Rt+3: stock returns for year t+1, t+2, t+3

an
respectively. Xt, Xt+1, Xt+2, Xt+3: earnings change for year t, t+1, t+2, t+3 respectively. AGt:
the growth rate of the total book value of assets for period t. EPt-1: the period t-1’s earnings
over price at the start of period t.
M
The above mentioned studies suggest that corporate reporting is a significant source of
changing expectations about a firm’s future earnings. They argue that if a firm reports
relevant information, disclosure will ‘bring the future forward’ in that it affects the
d

unanticipated portion of current year’s earnings and the revision in future earnings
te

expectations. It follows that there exists an interaction effect between future earnings and the
current level of a firm’s disclosure labeled by Lundholm & Myers (2002) “revealed future
p

earnings”. Subsequently, and consistent with Lundholm & Myers (2002) and Raonic et al
ce

(2011) approach, we examine the effect of risk disclosure (RDit) on share price anticipation of
future earnings in the presence of proprietary costs (PCit) by interacting all right-hand side
variables in Collins et al (1994) regression model with RDit and PCit variables.
Ac

We made two changes relative to the Collins et al. (1994) regression model. First, we control
for the cross sectional variation among corporate characteristics. A considerable number of
studies on the returns-earnings relationship documented many determinants (past growth, risk,
earnings persistence, firm size, firm profitability and the presence of an accounting loss) of
the earnings response coefficient. In our study, we examine our hypothesis with some of these
determinants. We use firm size, profitability and risk included in the regressions.

17
Page 17 of 46
 We include one control variable in the model at a time due to the limited number of
observations2 (degrees of freedom). Second, when calculating the current and future earnings
growth variables we deflate earnings change by price and not by lagged earnings.

Our regression models are presented as follows:

3 3
(M 1) : Rt  b0  b1 X t   bk 1 X t  k  bk 4 Rt  k  b8 EPt 1 b9 AGt  b10 RD  b11 [ RD * X t ]

t
k 1 k 1

ip
3 3
  bk 11[ RD * X t  k ]  bk 14 [ RD * Rt  k ]  b18 [ RD * EPt 1 ] b19 [ RD * AGt ] 
k 1 k 1

cr
3 3
b 20 Ctl  b21[Ctl * X t ]   bk  21[Ctl * X t  k ]  bk  24 [Ctl * Rt  k ]  b28 [Ctl * EPt 1 ] 
k 1 k 1

b29 [Ctl * AGt ]  et

us
3 3
(M2) : Rt  b0  b1Xt  bk1Xt k bk 4Rt k  b8EPt 1 b9 AGt 

an
k 1 k 1
3 3
b10RD b11[RD* Xt ]  bk11[RD* Xt k ] bk 14[RD* Rt k ]  b18[RD* EPt 1] b19[RD* AGt ] 
k 1 k 1
3 3
b20PC b21[PC* RD* Xt ]  bk21[PC* RD* Xt k ] bk 24[PC* RD* Rt k ]  b28[PC* RD* EPt 1] 
M
k 1 k 1
3 3
b29[PC* RD* AGt ]  b30Ctl b31[Ctl* Xt ]  bk 31[Ctl* Xt k ] bk 34[Ctl* Rt k ]  b38[Ctl* EPt 1] b39[Ctl* AGt ]  et
k 1 k 1
d

Where: Rt : stock return for year t is the buy-and-hold returns from six months before the financial
year-end to six months after the financial year-end. Xt : the earnings change per share deflated by the
te

share price at the start of the return window for period t. Xt+1 , Xt+2, Xt+3: the earnings change per share
for year t+1, t+2, t+3 respectively deflated by the price at the start of the return window for period t.
p

Rt+1, Rt+2, Rt+3: stock return for year t+1, t+2, t+3 are calculated as buy-and-hold returns for the 12-
ce

months period. AGt: The growth rate of total book value of assets for period t. EPt-1: is defined as
period t–1’s earnings over price six months after the financial year-end of period t–1. RD: Risk
disclosure variable defined as the natural logarithm of the number of risk-related sentences,
Ac

respectively for operation risk, empowerment risk, information processing and technology risk,
integrity risk and strategic risk. PC: proprietary cost is measured by the profit margin ratio = net
income/revenue. Ctl: controlling variables are defined as follows: Lev: Firm risk is the book value
of equity scaled by total liabilities. Size: Firm size is the natural logarithm of corporate revenues. Prof
: Firm profitability is the natural logarithm of Return on Equity (ROE).

2
This corresponds to Lundholm and Myers (2002) and Banghoj & Plenborg (2008). By including control
variables we expect to reduce the residuals, although they will also affect our degrees of freedom. Controlling
variables are actually interacted with each predictor in the model in the same way as our variables of interest.

18
Page 18 of 46
Before running the multiple regression analysis based on panel data, we perform several
specification tests and procedures in order to insure that the regression model specification fits
the data. We address three main concerns: the question of whether to pool the data or not, the
tests for individual and time effects and the heteroskedasticity of error terms. First, the Chow
test statistic rejects the null hypothesis of homogeneity among individuals at the level of 5%,
confirming the presence of specific effects in our cross sectional time series data. We then use

t
ip
the Hausman specification test to examine the nature of these specific effects. The test allows
us to choose between a fixed effect model and a random effect model. The results indicate

cr
that random effects estimation is inconsistent and that the fixed effects model is more
appropriate for our panel data. Finally, we run the Wald test for groupwise heteroskedasticity

us
in the residuals for a fixed-effects regression model. The Chi2 statistics reject the null
hypothesis of homoscedastic disturbances. Accordingly and consistent with Baltagi (2005),
we handle the heteroskedasticity problem by computing standard errors that are robust to the
homoscedasticity specification.

an
Applying fixed effects regressions to our empirical models helps us to control for unobserved
M
sources of firm heterogeneity. Fixed effects method resolves “joint determination” problems
in which an unobserved time invariant variable impacts simultaneously share price
d

informativeness and voluntary disclosure. The fixed effects results go a long way towards
dismissing omitted variable explanations as a source of endogeneity (Ferreira et al, 2010).
te

Because only the effect of within changes in risk disclosure is taken into account, we are quite
sure that firm specific omitted variables cannot explain the observed relationship between
p

share price informativeness with respect to future earnings growth and the level of risk
ce

information.

4. Empirical analysis
Ac

4.1 Descriptive analysis

Table 2 provides summary statistics of our sample with observations from year 2007 to 2009.
The mean current return is 3.38 percent. The mean current earnings per share change is 0.47
percent of the price. The mean of future earnings change is 0.28, 0.45 and 0.3 percent of the
price for the three periods ahead t+1, t+2 and t+3, respectively. These statistics suggest that
the mean accumulated 3 year future earnings is roughly three times the size of current
earnings3.

3
The paired samples t-tests for difference in means are not significant at the level of 5%

19
Page 19 of 46
Table 2 shows also that the mean of future annual returns is 0.51, 10.8 and 4.47 percent for
t+1, t+2 and t+3 respectively. These statistics indicate a somewhat lower mean future return in
t+1, period compared to current return. In contrast, there is a reverse and an increase in the
mean stock returns for period t+2. This implies changes in the returns over the sample time
period4. The mean of corporate asset growth rate is about 12.84 percent and the standard
deviation is about 0.1796 suggesting that there is little variation in the corporate asset growth

t
ip
rate in our sample. Proprietary costs level ranges from -34.454 percent to 95.2 percent.
Corporate profit margin has a mean of 0. 35 and a standard deviation of 5.744 indicate

cr
relatively high competition among firms in the MENA region.

us
Table 3 contains descriptive statistics for the risk disclosure level (RDit). The mean (median)
of risk disclosure level is increasing over time from 23.41 (19) sentences in 2007 to 29.76
(24) sentences in 20095. This trend is considerably different among firms with an increasing

an
standard deviation from 19.61 in 2007 to 22.27 in 2009. On a longitudinal data basis, the
level of risk disclosure is on average 27.26 sentences reflecting a fairly low score over time.
M
Regarding control variables, the mean of financial leverage, as measured by debt to equity
ratio, is 41.52 percent and the standard deviation is about 43.16 This risk level average
suggests that our sample includes moderately geared firms despite the high dispersion among
d

observations. Firm sizes vary considerably among our sample. Corporate revenue for the first
te

quartile is $22.19 million and more than $437 in the third quartile indicating that our
observations contain both small and large corporations. The mean of corporate profitability as
p

measured by the return on equity (ROE) is 25.96 and the standard deviation is about 149. This
ce

indicates that firms in our sample are relatively well performing during the period of analysis,
though the substantial dispersion among them in term of performance.
Ac

4
The paired samples t-tests for difference in means between the current period and the t+2 future period as well
as the cumulated 3 year future returns are significant at the level of 5%
5
The paired samples t-tests for difference in means between the level of risk disclosure in 2007, 2008 and 2009
are significant at the level of 5%

20
Page 20 of 46
 Table.2 Summary descriptive statistics

Variables Mean. S.D Min. 25% Median 75% Max. Obs.

Rit 0.0338 0.464 -0. 597 -0.340 -0.044 0.317 0.945 809
Xit 0.0047 0.046 -0.078 -0.0173 0.002 0.027 0.0925 809
Xit+1 0.0028 0.042 -0.073 -0.018 0.0001 0.023 0.079 809
Xit+2 0.0045 0.043 -0.070 -0.0179 0.0002 0.024 0.0875 809

t
Xit+3 0.003 0.034 -0.0617 -0.0107 0.000 0.025 0.064 809

ip
Rit+1 0.0051 0.4421 -0.604 -0.3402 -0.0679 0.2816 0.8514 809
Rit+2 0.1083 0.3631 -0.3538 -0.1718 0.0350 0.3125 0.8139 809
Rit+3 0.0447 0.2850 -0.3478 -0.1854 0 0.2519 0.5532 809

cr
AGit 0.1284 0.1796 -0.0894 -0.0108 0.082 0.2327 0.4919 809
Epit-1 0.8855 1.3427 -0.02 0.03 0.17 1.18 4.05 809
RDit 27.265 21.181 0 13 22 35 145 809

us
PCit 0. 350 5.744 -34.454 3.3 0.093 0.198 95.2 809
Levit 41.527 43.16 0 3.164 25.892 71.209 127.23 809
Sizeit (M$) 660.56 1930. 0 22.19 117.07 437.39 25101.7 809
9

an
Profit 25.964 149.13 0.05 0.12 14.59 184.3 3216.18 809
Table 2 reports the summary statistics for the sample firms using data pooled across the three year sample
period. The earnings per share measure is a Reuters’ fundamental item, calculated by dividing ‘earnings for
ordinary-full tax’ by the number of shares outstanding. Xit, Xit+1, Xit+2 and Xit+3 are defined as earnings change
M
deflated by price. Both current and future earnings changes are deflated by the price at the start of the return
window for period t. Rit, Rit+1, Rit+2 and Rit+3 are calculated as buy-and-hold returns (inclusive of dividends) over
a 12-month period, starting six months after the end of the previous financial year. EPit–1 is defined as period t–
1’s earnings over price six months after the financial year-end of period t–1. AGt is the growth rate of the total
d

book value of assets for period t. RDit is the total number of risk related sentences,. Unlogged values are
reported. In subsequent regressions the natural logarithm is used. Proprietary cost PCit is the profit margin
te

defined as Net Income/Revenue. Leverage Levit is defined as book value of equity scaled by total liabilities (data
are winsorized at the level of 10%). Sizeit is measured by corporate revenues. Profit is measured by the Return on
Equity (ROE). Likewise, unlogged values are reported and in subsequent regressions the natural logarithm is
used.
p

Table 3. Descriptive statistics for Risk Disclosure level from 2007-2009

ce

N Mean Median Min Max S.D

2007 254 23.411 19 0 140 19.616
2008 264 27.640 22 1 145 21.092
Ac

2009 270 29.762 24 1 144 22.276

Table 4 presents pairwise Pearson correlations for all regression variables. We estimate
correlations using data pooled across the three-year sample period. As documented in
previous studies, the correlation between current Returns (Rit) and current earnings growth
(Xit) is strong and significant at the 1% level suggesting that current earnings are perceived as
value relevant. Besides, the correlation between Rit and Xit+1 is weaker, but still significant at
the 5% level. The correlation between Rit and Xit+2 and Xit+3 is however not significant. This
provides evidence of prices leading earnings by only one period.

21
Page 21 of 46
Current return is also correlated with future returns regarding period t+1 while it is
uncorrelated with Rit+2 and Rit+3. Future returns (Rit+1, Rit+2 and Rit+3) are significantly
correlated with future earnings growth (Xit+1, Xit+2 and Xit+3), consistent with Collins et al.
(1994). These correlations indicate that future returns should not influence the results
except through their role as a proxy for the measurement error in future earnings. We report
furthermore a significant negative correlation between current earnings (Xit) and future

t
ip
earnings growth for period t+1 and t+2. Similarly, Xit+1 is significantly and negatively related
to Xit+2 and Xit+3. These correlations suggest potential multi-collinearity problems within the

cr
independent variables. The variable inflation factor (VIF) did not raise significant problems
among the explanatory variables. The mean VIF is about 1.19 and the computed VIF for

us
each variable of interest is under 106. There is additionally some evidence that AGit as well
as Epit-1 are a good estimation error proxy. In theory, an errors-in-variables proxy should be
highly correlated with the measurement error but uncorrelated with the dependent variable.

an
This is the case for these two control variables. The correlation coefficients between Rt in one
hand and AGit as well as Epit-1 in the other hand, are respectively 0.035 (0.32) and -0.024
M
(0.489) both insignificant. Risk disclosure level is positively and significantly associated with
current stock returns at the level of 10 %. This significant relation indicates that risk
disclosure practice of firms in MENA emerging markets is driven by stock price performance.
d

It is worth noting that Lundholm & Myers (2002) find in the same way a positive and
te

significant correlation between current returns and the disclosure score. Finally, The level of
proprietary costs (PCit) as measured by the profit margin ratio is positively and significantly
p

correlated with the earnings yield of prior period (t-1). This suggests that the proprietary costs
ce

are increasing in function of the earnings to price ratio. As the earnings yield serves for many
investment managers to determine optimal asset allocations, a high level is likely to attract
more competitors into the product market. Pairwise testing reveals also that proprietary costs’
Ac

level is significantly and negatively correlated with risk disclosure pointing out that
corporations tend to decrease their risk disclosure when the proprietary costs are high.

6
More details about the tabulated VIF and the tolerance values are provided in appendix B. It is worth noting
that high VIF is commonly noticed in interaction models. Recent studies (e.g. Brambor et al, 2006; Echambadi
& Hess, 2007) argue that this issue has been overstated since the coefficients in interaction models no longer
indicate the average effect of a variable as they do in an additive model. consequently, they are almost certain to
change with the inclusion of an interaction term and this should not be interpreted as a sign of multicollinearity.
Alisson (1999) suggests also that we can safely ignore multicollinearity as long as the variables of interest and
the overall mean VIF in interaction models are below the critic threshold. Recall that Echambadi & Hess, (2007)
showed the shortness of the main applied method (mean centering) in addressing the collinearity issue in
moderated multiple regression models.

22
Page 22 of 46
 t
ip
cr
Table .4 Pearson correlation matrix

Rit Xit Xit+1 Xit+2 Xit+3 Rit+1 Rit+2 Rit+ 3 EPit–1 AGit RDit PCit

us
Rit 1.000
Xit 0.143*** 1.000
(0.000)
Xit+1 0.0843** -0.200*** 1.000

an
(0.016) (0.000)
Xit+2 0.0132 -0.059* -0.144*** 1.000
(0.706) (0.093) (0.000)
Xit+3 0.0132 -0.035 -0.062* -0.138*** 1.000
(0.708) (0.308)

M
(0.077) (0.000)
Rit+1 -0.249*** -0.036 0.170*** 0.153*** -0.061* 1.000
(0.000) (0.301) (0.000) (0.000) (0.081)
Rit+2 0.0175 0.032 0.004 0.067* 0.132*** -0.169*** 1.000
(0.619) (0.358) (0.907) (0.054) (0.000) (0.000)

ed
Rit+3 0.054 0.018 -0.002 0.116*** 0.067* 0.041 0.039 1.000
(0.123) (0.599) (0.974) (0.001) (0.053) (0.171) (0.256)
EPit–1 -0.024 -0.073** -0.040 -0.060* -0.038 0.014 0.003 -0.072** 1.000
(0.489) (0.036) (0.244) (0.087) (0.273) (0.686) (0.914) (0.040)
AGit 0.035 0.073** -0.034 -0.002 0.020 -0.056 0.099*** -0.083** 0.063* 1.000
pt
(0.320) (0.036) (0.322) (0.951) (0.558) (0.110) (0.004) (0.017) (0.070)
RDit 0.0622* -0.021 -0.005 0.027 -0.020 0.085** 0.064* 0.002 0.162*** 0.030 1.000
(0.076) (0.544) (0.880) (0.431) (0.569) (0.015) (0.065) (0.955) (0.000) (0.385)
0.0341 0.022 -0.016 -0.0087 -0.025 -0.027 0.001 -0.044 0.043 1.000
ce

PCit 0.162*** -0.092***

(0.332) (0.518) (0.645) (0.805) (0.467) (0.437) (0.978) (0.211) (0.000) (0.221) (0.008)
Table4 presents Pearson correlations for all regression variables using panel data across the three year sample period. P-values are given in parentheses. The number of
observations is 809. The earnings per share measure is a Reuters’ fundamental item, calculated by dividing ‘earnings for ordinary-full tax’ by the number of shares
Ac

outstanding. Xit, Xit+1, Xit+2 and Xit+3 are defined as earnings change deflated by price. Both current and future earnings changes are deflated by the price at the start of the
return window for period t. Rit, Rit+1, Rit+2 and Rit+3 are calculated as buy-and-hold returns (inclusive of dividends) over a 12-month period, starting six months after the end of
the previous financial year. EPit–1 is defined as period t–1’s earnings over price six months after the financial year-end of period t–1. AGit is the growth rate of the total book
value of assets for period t. RDit is the natural logarithm of the total number of risk related sentences. Proprietary costs PCit is the profit margin ratio defined as Net
Income/Revenue.

23

Page 23 of 46
4.2. Results

4.2.1. The effect of voluntary risk disclosure on share price anticipation of future earnings

The regression results in table 5 show that the coefficients on Xit are, though positive,
insignificant in the three regression models with each control variable. This suggests that
investors do not consider current earnings as a good prior estimation of future earnings7.

t
ip
There is also strong evidence of share price anticipation of future earnings for one period
ahead. The fact that the current return–earnings relationship is weak, but future earnings

cr
growth variables and measurements error proxies are significantly more important to return
variation is consistent with earnings lacking timeliness in measuring value creation. This

us
phenomenon applies to the three regressions and is independent of the amount of voluntary
risk information in annual report narratives. All the three coefficients on Xit+1 are positive and
these coefficients are significant at the level of 1%. The significant estimates for this

an
coefficient range from 2.516 to 2.961. There is also some evidence on the ability of stock
returns to anticipate future earnings changes for three years ahead. The coefficients on Xit+3
M
extend from 1.054 to 1.395 and are significantly different from zero at 5% and 10% level
when controlling for financial leverage and corporate profitability respectively. The
coefficients on AGit are insignificant and have the wrong sign. This suggests that AGit is not a
d

satisfactory measurement error proxy. Findings show also that the coefficients on Epit-1 are
te

negative but insignificant. The negative sign on Epit-1 suggests that earnings follow a random
walk. The coefficients on the future stock return variables (Rit+1, Ritt+2 and Rit+3) are expected
p

to be negative. Negative coefficients on future stock returns should confirm that realized
ce

future earnings contain a measurement error that future returns remove (Collins et al., 1994).
This is the case for all period future stock returns. Particularly, future stock returns are
significantly different from zero at respectively 1%, 5% and 10% levels for t+1 as well as t+2
Ac

periods.

The coefficients of primary interest in the current study are the coefficients on RDit*Xit, RDit*
Xit+1, RDit*Xit+2, and RDit* Xit+3. We anticipate positive coefficients on the interaction terms.
We find a strong and significant impact of risk disclosure on the return current earnings
association at the 1% level when controlling for corporate profitability.

7
When interpreting the regression coefficient on current earnings, it must be noted that when future earnings are
also in the regression, the current earnings variable proxies for the market’s prior expectation about future
earnings. This may confound the traditional interpretation of an earnings response coefficient (Lundholm and
Myers, 2002).

24
Page 24 of 46
This finding suggests that voluntary risk disclosure makes financial data more credible; a
price reaction to the unanticipated portion of current earnings is documented mainly when the
market is informed about firm profitability. The coefficients on the interaction terms RDit*
Xit+1 are negative (with one exception) and insignificant at conventional levels for the three
regression models (for each control variable apart). The coefficient on RDit* Xit+2 is positive
and significant at 10% and 5%. These conflicting coefficients on RDit* Xit+1 and RDit* Xit+2

t
ip
indicate that investors take more time to assess the amount of risk information prepared at the
end of t period and available to the financial market at the start of the following year.

cr
Investors are not able to incorporate such information in the firm value estimates in t+1
period. As a result, the benefits mainly accrue in the following t+2’s financial year. Voluntary

us
risk information “brings the future forward” in that it impacts positively share price
informativeness with respect to future earnings for t+2 period, particularly when information
about the financial leverage level and corporate profitability are made available to the market.

an
The panel regression analysis provides evidence on the usefulness of risk information in
MENA emerging markets. The results reflect the role of voluntary risk disclosure in reducing
M
the information asymmetry regarding firm future performance between insiders and outsiders.
This is consistent with the explanation that in a low-regulation environment, managers
voluntarily report on risk and uncertainties to reduce the agency problems and to allow
d

investors to make better-informed decisions (Deumes & Knechel, 2008). At the same time,
te

investors react positively to this signal and incorporate such relevant information into stock
prices. Our results highlight then a trend toward improving corporate disclosure activity and
p

ensuring the well functioning of financial markets within the scrutinized context. Firms’
ce

transparency about their risk profile is somehow the outcome of the recent corporate
governance reforms in the region. Regulators and professional agencies targeted to enhance
monitoring of management, to gain the trust of foreign investors and to enforce their
Ac

protection by encouraging the adoption of international recommended practices. Investors’

financial literacy seems to be, on the other side, in progress as they are aware about the
informational needs despite the unique cultural context in the region. Their accurate
understanding of business risks, improves their ability to predict future earnings.

25
Page 25 of 46
 Table 5. Voluntary risk disclosure and share price informativenesss

Model Exp. Sign Leverage Size Profitability

Intercept (?) -0.107 (0.257) 0.073 (0.607) -0.018 (0.859)
Xit (+) 0.370 (0.546) 0.257 (0.672) 0.430 (0.488)
Xit+1 (+) 2.961*** (0.000) 2.634*** (0.000) 2.516*** (0.000)
Xit+2 (+) 0.695 (0.339) 1.005 (0.156) 0.833 (0.238)

t
Xit+3 (+) 1.054 (0.166) 1.550** (0.038) 1.395* (0.076)

ip
Rit+1 (–) -0.579*** (0.000) -0.607*** (0.000) -0.551*** (0.000)
Rit+2 (–) -0.319** (0.020) -0.313** (0.019) -0.264* (0.071)
Rit+3 (–) -0.191 (0.168) -0.187 (0.180) -0.219 (0.121)

cr
AGit (–) 0.194 (0.134) 0.168 (0.179) 0.160 (0.200)
Epit-1 (+) -0.018 (0.630) -0.009 (0.823) -0.011 (0.774)
RDit (?) 0.041 (0.155) 0.061** (0.037) 0.024 (0.412)

us
RDit*Xit (?) 0.114 (0.243) -0.011 (0.914) 0.271*** (0.005)
RDit*Xitt+1 (+) -0.129 (0.178) -0.163 (0.199) 0.118 (0.352)
RDit*Xit+2 (+) 0.191* (0.064) 0.091 (0.524) 0.302** (0.014)
RDit*Xitt+3 (+) 0.035 (0.746) 0.054 (0.713) 0.104 (0.468)

an
RDit*Rit+1 (?) -0.020 (0.423) -0.003 (0.904) 0.023 (0.427)
RDit*Rit+2 (?) -0.064* (0.076) -0.030 (0.468) -0.047 (0.291)
RDit*Rit+3 (?) -0.031 (0.324) -0.023 (0.465) -0.020 (0.583)
RDit*AGit (?) 0.002 (0.666) 0.002 (0.587) 0.003 (0.469)
M
RDit*Epit-1 (?) -0.000 (0.445) 0.000 (0.741) 0.002 (0.114)
Year Yes Yes Yes
Dummies
Adj R2
d

39.31% 40.14% 39.96%

Observations 809 809 809
te

Table 5 presents fixed effects regression results for panel data. Heteroscedasticity-consistent p-values are given
in parentheses. The dependent variable is current period return, Rit. Rit, Rit+1, Rit+2 and Rit+3 are calculated as buy-
and-hold returns (inclusive of dividends) over a 12-month period, starting six months after the end of the
previous financial year. Xit, Xit+1, Xit+2 and Xit+3 are defined as earnings change deflated by price. Both current
p

and future earnings changes are deflated by the price at the start of the return window for period t. EPit–1 is
ce

defined as period t–1’s earnings over price six months after the financial year-end of period t–1. AGit is the
growth rate of the total book value of assets for period t. RDit is the natural logarithm of the total number of risk
related sentences. Leverage, size and profitability are control variables (their regression estimates are not
tabulated). The significance levels (two-tail test) are: *= 10 %, ** =5 % and *** = 1 %.
Ac

4.2.2. The effect of proprietary costs and voluntary risk disclosure on share price anticipation of future
earnings

Notwithstanding the above mentioned evidence on the usefulness of risk disclosure in

anticipating future earnings growth, higher disclosure informativeness seems to be associated
also with more intense competitive business environment (Irvine & Pontiff, 2009). In such
situation, corporations may incur proprietary costs. A decrease in corporate future cash flows
is likely to occur following the disclosure of risk information. Typically, actual as well as
potential competitors would use such information in a way it harms corporate market position.

26
Page 26 of 46
Corporations look to reduce their risk disclosure and balance their desire to convey relevant
information under a tradeoff between costs and benefits. Investors may consider this practice
as a weakness in analyzing and monitoring their business risk. The lack of transparency may
lead to a lower perceived credibility of risk disclosure.

Table 6 reports the main findings of H2 testing that predicts a weaker value relevance of

t
voluntary risk disclosure when interacted with proprietary cost level. First, the coefficient on

ip
Xit is as predicted significantly positive at the 10% level for the regression when we control
for financial leverage variable. This result provides weak evidence that current earnings

cr
reflect prior anticipation of future earnings. There is also strong evidence of prices leading
earnings by three periods ahead regardless of the extent of corporate risk disclosure and

us
proprietary cost level. The future earnings response coefficients on Xit+1, Xit+2 and Xit+3 are,
with one exception, positive and significant in the three regression models. When we control

an
for corporate risk, the coefficient on Xit+3 is insignificant suggesting that share price
informativeness with respect to future earnings is limited to two years ahead. Future earnings
response coefficients extend from 1.062 to 2.831 and are considerably higher than the current
M
earnings coefficients. The coefficients on future stock returns (Rit+1, Rit+2 and Rit+3) have as
predicted the right sign and are significantly different from zero mainly for t+1 and t+2
d

periods. These findings indicate that realized future earnings contain measurement error that
future returns remove. Table 6 also shows that the coefficients on RDit*Xit as well as on
te

RDit*Xit+k are mostly insignificant in the three regression models. There is only a positive and
significant association between current returns and current and future earnings for period t+2
p

when the market is informed about corporate profitability. There is, accordingly, little
ce

evidence on share price informativeness with respect to current and future earnings when
interacted with risk disclosure level.
Ac

The coefficients of our primary interest in this study indicate that proprietary costs affect
investors’ ability to impound more future earnings news into the current return. Consistent
with Darrough & Stoughton (1990), Gelb (2000) and Hussain et al (2006) ‘s findings that
when proprietary costs are high, firms tend to restrain value relevant information about future
cash flows, we find that there is a significant and negative relation between PCit*RDit*Xit+k
and current returns. The negative coefficients on the interaction terms indicate that the amount
of voluntary risk information impounded in the current stock price is less for firms with high
proprietary costs because investors perceive it as less relevant.

27
Page 27 of 46
Investors will reconsider the perceived relevance of risk information as long as they are aware
that firms’ management would restrain useful information about their business uncertainties.
Capital market participants are likely to rely on other sources of risk information other than
corporate annual reports rather than hold a short-term view. One possible interpretation of
these results is that a significant profit margin (the mean is about 0.35) may attract
future competition and reveal higher forewarning of potential entrants (Cohen, 2004).

t
ip
Investors may believe that corporations need to protect their future opportunities and are
likely to provide relevant risk information about future prospects via private channels.

cr
Since the content of risk information is not explicitly ruled through norms or laws for it is

us
subject to managerial discretion within the scrutinized context, investors perceive risk
disclosure as a matter of biased selected information. Most profitable companies in an
increasingly competitive setting like MENA emerging markets have potentially more to lose

an
by revealing commercially sensitive information into the public domain. Actually, the
development of private sector and the improvement of economic freedom through trade
liberalization and property rights promotion within the MENA region reduced the barriers to
M
market entry and created new business opportunities for international investors. As risk
management and disclosure strategies reflect the underlying environmental influences (e.g.
d

economic incentives, the regulatory environment, competition.. etc..), it seems natural that
te

successful MENA companies tend to withhold information about their opportunities and prospects.
Meanwhile, market participants seem to make allowances for management discretion over
p

risk information and tend to diversify their sources of information when making their
economic decision.
ce
Ac

28
Page 28 of 46
 Table 6. Proprietary costs and share price informativeness
Model Exp. Leverage Size Profitability
Sign
Intercept (?) 0.208** (0.038) 0.389 (0.124) 0.236** (0.046)
Xit (+) 0.043* (0.060) -0.167 (0.792) 0.181 (0.842)
Xit+1 (+) 2.831** (0.048) 2.565** (0.038) 2.292* (0.083)
Xit+2 (+) 1.062* (0.093) 1.106** (0.040) 1.256* (0.061)

t
Xit+3 (+) 1.364 (0.112) 2.050** (0.025) 1.706* (0.088)

ip
Rit+1 (–) -0.545* (0.057) -0.542* (0.059) -0.531* (0.061)
Rit+2 (–) -0.423*** (0.007) -0.369** (0.013) -0.366** (0.013)
Rit+3 (–) -0.184 (0.165) -0.169 (0.170) -0.221 (0.118)

cr
AGit (–) 0.112** (0.026) 0.104** (0.022) 0.076** (0.028)
Epit-1 (+) -0.023* (0.084) -0.010 (0.414) -0.011 (0.347)
RDit (?) 0.035 (0.214) 0.057 (0.169) 0.021 (0.309)

us
RDit*Xit (?) 0. 149 (0.176) 0.024 (0.764) 0.288* (0.098)
RDit*Xit+1 (+) -0.146 (0.222) -0.176** (0.049) 0.180 (0.272)
RDit*Xit+2 (+) 0.168 (0.130) 0.095 (0.537) 0.347* (0.051)
RDit*Xit+3 (+) -0.001 (0.988) -0.048 (0.543) 0.135 (0.385)

an
RDit*Rit+1 (?) -0.023 (0.249) -0.006 (0.626) -0.031 (0.405)
RDit*Rit+2 (?) -0.035* (0.097) -0.020 (0.386) -0.029 (0.217)
RDit*Rit+3 (?) -0.031 (0.291) -0.016 (0.693) -0.028 (0.177)
RDit*AGit (?) 0.007 (0.167) 0.007 (0.258) 0.008 (0.111)
M
RDit*Epit-1 (?) 0.000 (0.986) 0.001 (0.116) 0.002*** (0.003)
PCit (?) 0.079 (0.116) 0.085 (0.158) 0.079** (0.050)
PCit*Xit (?) 0.036** (0.021) 0.029* (0.084) 0.032** (0.018)
PCit*Xit+1 (?) 0. 029* (0.076) 0.022* (0.072) 0.026* (0.068)
d

PCit*Xit+2 (?) -0.005 (0.225) -0.003* (0.070) 0.002 (0.347)

PCit*Xit+3 (?) -0.010 (0.118) -0.015 (0.145)
te

-0.008** (0.030)
PCit*Rit+1 (?) -0.000 (0.128) -0.000 (0.184) -0.000** (0.033)
PCit*Rit+2 (?) 0.002 (0.125) 0.001 (0.226) 0.002 (0.185)
(?) 0.001 (0.398) 0.001 (0.452) 0.001 (0.368)
p

PCit*Rit+3
PCit*AGit (?) -0.001* (0.051) -0.001* (0.098) -0.001* (0.051)
(?) -0.000 (0.180) -0.000 (0.109)
ce

PCit*Epit-1 -0.000** (0.037)

PCit*RDit*Xit (?) -0.001 (0.228) -0.000 (0.447) -0.001 (0.347)
PCit*RDit*Xit+1 (–) -0.002** (0.037) -0.002* (0.073) -0.002* (0.072)
PCit*RDit*Xit+2 (–) -0.001** (0.015) -0.001** (0.042) -0.002** (0.011)
Ac

PCit*RDit*Xit+3 (–) -0.001** (0.022) -0.001** (0.019) -0.002*** (0.002)

PCit*RDit*Rit+1 (?) -0.000 (0.513) -0.000 (0.509) -0.000 (0.903)
PCit*RDitit*Rit+2 (?) -0.001 (0.118) -0.000 (0.162) -0.000 (0.134)
PCit*RDit*Rit+3 (?) -0.000 (0.347) -0.000 (0.396) -0.000 (0.270)
PCit*RDit*AGit (?) 0.000* (0.072) 0.000 (0.133) 0.000** (0.021)
PCit*RDit*EPit-1 (?) -0.000** (0.049) -0.000 (0.315) 0.000 (0.316)
Year Dummies Yes Yes Yes
Adj R2 40.41% 40.95% 41.01%
Observations 809 809 809
Table 6 presents fixed effects regression results for panel data. Heteroscedasticity-consistent p-values are given
in parentheses. The dependent variable is current period return, Rit. Rit, Rit+1, Rit+2 and Rit+3 are calculated as buy-
and-hold returns (inclusive of dividends) over a 12-month period, starting six months after the end of the
previous financial year. Xit, Xit+1, Xit+2 and Xt+3 are defined as earnings change deflated by price. Both current and
future earnings changes are deflated by the price at the start of the return window for period t. EPt–1 is defined as

29
Page 29 of 46
period t–1’s earnings over price six months after the financial year-end of period t–1. AGt is the growth rate of
the total book value of assets for period t. RDit is the natural logarithm of the total number of risk related
sentences,. Proprietary cost PCit is the profit margin defined as Net Income/Revenue. Leverage, size and
profitability are control variables (their regression estimates are not tabulated). The significance levels (two-tail
test) are: *= 10 %, ** =5 % and *** = 1 %.

5. Robustness checks: Controlling for potential endogeneity problems

Accounting literature addressed an important question of whether corporate disclosure has a

t
ip
first order effect on some outcome variables (e.g. Firm value, cost of capital, share price
informativeness… etc..). It follows that the association between capital market variables and

cr
corporate reporting may be driven by firm performance rather than information per se. It is
also possible that disclosure policy changes are not random events: they may coincide with

us
changes in firm economics and governance characteristics (Healy & Palepu, 2001). According
to Beyer et al. (2010) given the endogenous nature of the corporate information environment,
governance structure and some observed outcomes, it is hard to make an assessment of the

an
causal connection and recognize the exact impact that one mechanism would have on another
one.
M
As a sensitivity test, we acknowledge in this section two potential sources of endogeneity
between share price informativeness and risk disclosure. Intuitively, our reasoning is that
d

differences exist in the share price informativeness that are correlated with the firm’s risk
disclosure level, but that are not necessarily caused by this activity. Instead, these differences
te

are caused either by: (1) unobservable heterogeneity among firms in a cross-sectional sample;
or (2) observable determinants of the return-earnings relationship which are correlated with
p

disclosure but omitted from the analysis. All of these problems imply a correlation between
ce

the error term and one or some of the independent variables (e.g. risk disclosure and the
interaction terms) in the regression model. Controlling for time-invariant, unobserved
heterogeneity by adding firm fixed effects and some appropriate control variables can help to
Ac

mitigate this concern. However, this method implies a strong exogeneity hypothesis. It
assumes that current observations of the explanatory variable (e.g. voluntary risk disclosure)
are independent of past values of the dependent variable (e.g. stock returns). As firms choose
their level of risk disclosure based on their perceived cost-benefit analysis, this assumption is
not likely to hold in the specific context of voluntary disclosure.

30
Page 30 of 46
Following Larcker & Rusticus (2010), we use an instrumental regression as an alternative
approach to deal with the above mentioned concerns. These authors consider this method as
the most promising econometric approach to addressing the endogeneity problem in social
sciences. Instrumental variables are used to cut correlations between the error term and
independent variables. When more instrumental variables are available, two stage least
squares regression (2SLS) provides a way of obtaining the optimal linear combination of

t
ip
instruments (Woodridge, 2010).

Given that previous studies documented that the extent and the quality of corporate disclosure

cr
are associated with ownership structure, size, leverage, profitability and industry sector, we
estimate the relationship between the level of voluntary risk disclosure and these firms’

us
features in the first stage of the regression. The endogenous dependent variable (the level of
voluntary risk disclosure) is as such explained by these instruments and the rest of

an
exogenous variables of our model are also included in this first equation.

In the first stage we estimate:

M
M (1)

RD it  b0  b1OwC it  b 2 MOw it  b3 INsOw it  b4 Size it  b5 Lev it  b6 Pr of it 

d

3 3
b7 Sect i  b8 X it   bk  8 X it  k   bk 11 Rit  k  b15 EPt 1 b16 AG it  eit
te

k 1 k 1

In the second stage, we estimate the following regressions using the fitted values (estimates)
of risk disclosure (RDit) derived from the first stage:
p
ce

M(2a):
3 3
Rit  b0  b1 X it   bk 1 X it k  bk 4 Rit k  b8 EPit 1 b9 AGit  b10 Fit ( RDit )  b11[ Fit ( RDit ) * X it ] 
k 1 k 1
Ac

3 3
 bk 11[ Fit( RDit ) * X it k ]  bk 14 [ Fit( RDit ) * Rit k ]  b18 [ Fit(RDit ) * EPit 1 ] b19 [ Fit(RDit ) * AGit ]  eit
k 1 k 1

M(2b):

31
Page 31 of 46
 3 3
Rit  b0  b1 X it   bk 1 X it  k   bk  4 Rit  k  b8 EPit 1 b9 AG it  b 10 Fit ( RD it )  b11 [ Fit ( RD it ) * X it ] 
k 1 k 1
3 3
 bk 11 [ Fit ( RDit ) * X it  k ]   bk 14 [ Fit ( RD it ) * Rit  k ]  b18 [ Fit ( RD it ) * EPit 1 ] b19 [ Fit ( RDit ) * AG it ] 
k 1 k 1
3 3
b20 PC it  b21 [ PC it * Fit ( RD it ) * X it ]   bk  21 [ PC it * Fit ( RD it ) * X it  k ]   bk  24 [ PC it * Fit ( RD it ) * Rit  k ] 
k 1 K 1

b28 [ PC it * Fit ( RD it ) * EPit 1 ]  b 29 [PC it * Fit(RD it ) * AG it ]  eit

t
ip
Where: Rit : stock return for year t is the buy-and-hold returns from six months before the financial year-end to
six months after the financial year-end. Xit : earnings change per share deflated by the share price at the start of
the return window for period t. Xit+1 , Xit+2 , Xit+3: the earnings change per share for year t+1, t+2, t+3 respectively

cr
deflated by the price at the start of the return window for period t. Rit+1, Rit+2, Rit+3: stock return for year t+1, t+2,
t+3 are the buy-and-hold returns for the 12-months period. AGit: The growth rate of total book value of assets for
period t. EPt-1: the period t–1’s earnings over price six months after the financial year-end of period t–1. RDit: is

us
the natural logarithm of the number of risk-related sentences. PCit: proprietary cost is the profit margin ratio =
net income/revenue. Levit : Firm risk is the book value of equity scaled by total liabilities. Sizeit: Firm size is the
natural logarithm of corporate revenues. Profit : Firm profitability is the natural logarithm of Return on Equity
(ROE). Secti: Industry sector 1 if the firm is classified into one of the nine commonly defined industry sectors

an
and 0 otherwise. OwCit : the proportion of equity held by the top-five largest shareholders, including financial
institutions, firms’ inside shareholders (directors and executives) and other outside block shareholders. MOwit :
the proportion of equity held by managers and executive directors. InsOwit : the total number of shares held by
institutions to the total number of shares outstanding per sample firm.
M
According to Larcker & Rusticus (2010), it is common for accounting researchers to report
the classic Hausman (1978) test for endogeneity. As we are dealing with panel data, we rely
d

on the “endog” test (known as “difference-in-Sargan-Hansen" statistic) which is robust to

various violations of conditional homoscedasticity. If the probability of the test is below 0.05,
te

then we reject the hypothesis of the exogeneity of the instrumented variable (related-risk
disclosure, RDit). The “endog” test distributed as chi(2) statistic with 10 degrees of freedom
p

reports a value of 19.067 and a probability of 0.0394. Thus, it is necessary to apply an

ce

instrumental variable approach to correct for endogeneity problem. 2sls procedure is used in
the form of Baltagi's EC2SLS random-effects estimator8.
Ac

Table 7. reports the main findings for our two regression models. The instrumental approach
yielded slightly different results compared to our initial analysis.

8
The test of overidentifying restrictions (orthogonality conditions) failed to reject the the null hypothesis that the
excluded instruments are valid instruments, i.e., uncorrelated with the error term. Accordingly a RE estimation is
preferred over a FE regression (Wooldridge, 2010).

32
Page 32 of 46
5.1. The effect of voluntary risk disclosure on share price anticipation of future earnings

The coefficients on Xit and Xit+k become insignificant suggesting that current stock price is
not associated with current and future earnings change. So there is no evidence of prices
leading earnings regardless of corporate risk disclosure. There is also a weak positive
association between risk disclosure and current stock returns at the 10% level. This indicates

t
that risk reporting activity for our sample firms is driven by stock price performance. Finding

ip
revealed furthermore a positive and significant association between current stock returns and
the interacted term of current earnings with the risk disclosure at the level of 5%. This

cr
provides evidence that information about risk and uncertainties increase the credibility of
current earnings. These results indicate that if earnings are not complemented with risk

us
information, they may not be believed by stock market participants. Finally, unlike the initial
findings, instrumental analysis showed that firm risk disclosure level is incrementally relevant

an
for anticipating future earnings growth only for one year ahead. The coefficient on the
interaction term RDit*Xit+1 is positive and significant at the level of 5%. There is accordingly
evidence that corporate risk disclosures are value relevant to market participants.
M
5.2. The effect of proprietary costs and voluntary risk disclosure on share price anticipation of future
earnings
d

Results report evidence on share price informativeness with respect to future earnings for one
te

year and three years ahead when interacted with risk disclosure level. The coefficients on the
interaction terms RDit*Xit+1 and RDit*Xit+3 are positive and significant at the levels of 1% and
p

10% respectively. These findings suggest that investors perceive risk reporting activity as
ce

informative about future earnings growth in the absence of information about the level of
competition in the product market. Somewhat similar to the early finding, instrumental
Ac

regression confirms that there is less amount of risk information (that are related to future
earnings) impounded into the current stock price of highly profitable firms. The negative and
significant coefficients on the interaction terms PCit*RDit*Xit, and PCit*RDit*Xit+k suggest that
the product market competition is likely to moderate the perceived relevance of narrative risk
information in annual reports. As a result of which, we should conclude that risk discolosures
of high profit margin firms are not reliable for market participants in their valuation process.
Investment decision is likely to be jeopardized by the increasing difficulty of making reliable
forecasts in a context with high proprietary costs, especially when it is associated with low
regulation and weak and ineffective corporate governance mechanisms.

33
Page 33 of 46
 Table 7. Instrumental multiple regression analysis
Model Exp. Sign M2(a) M2(b)
Intercept (?) -0.788 (0.132)
-0.196 (0.405)
Xit (+) -3.796 (0.108)
-0.378 (0.850)
Xit+1 (+) -6.894 (0.111)
-8.233*** (0.003)
Xit+2 (+) 4.294 (0.349)
-0.525 (0.867)
Xit+3 (+) 4.625 (0.393)
-6.194 (0.116)
Rit+1 (–) 0.429 (0.470)
0.096 (0.860)

t
Rit+2 (–) -0.056 (0.946)
-0.286 (0.710)

ip
Rit+3 (–) -0.450 (0.556)
-0.794 (0.237)
AGit (–) 0.416* (0.093)
0.960*** (0.001)

cr
Epit-1 (+) 0.009 (0.875)
0.0642 (0.292)
RDit (?) 0.311* 0.113
(0.062) (0.119)
RDit*Xit (?) 0.904** 0.308
(0.018) (0.300)

us
RDit*Xit+1 (+) 1.533** (0.019)
1.332*** (0.000)
RDit*Xit+2 (+) -0.393 (0.613)
0.358 (0.469)
RDit*Xit+3 (+) -0.512 (0.582)
1.095* (0.056)
RDit*Rit+1 (?) -0.249* -0.081
(0.094) (0.534)

an
RDit*Rit+2 (?) -0.072 (0.726)
0.014 (0.938)
RDit*Rit+3 (?) 0.078 (0.656)
0.202 (0.178)
RDit*AGit (?) 0.000 (0.991)
-0.011 (0.235)
RDit*Epit-1 (?) -0.002 (0.458)
-0.004 (0.301)
M
PCit (?) 0.007** (0.026)
PCit*Xit (?) 0.372*** (0.010)
PCit*Xit+1 (?) 0.711*** (0.000)
PCit*Xit+2 (?) 0.206 (0.272)
d

PCit*Xit+3 (?) 0.438* (0.077)

PCit*Rit+1 (?) -0.030 (0.280)
te

PCit*Rit+2 (?) 0.026 (0.451)

PCit*Rit+3 (?) 0.026 (0.557)
PCit*AGit (?) -0.051*** (0.001)
p

PCit*Epit-1 (?) -0.004 (0.154)

PCit*RDit*Xit (?) -0.063** (0.016)
ce

PCit*RDit*Xit+1 (–) -0.101*** (0.000)

PCit*RDit*Xit+2 (–) -0.045* (0.098)
PCit*RDit*Xit+3 (–) -0.071* (0.064)
(?) 0.006 (0.381)
Ac

PCit*RDit*Rit+1
PCit*RDit*Rit+2 (?) -0.007 (0.393)
PCit*RDit*Rit+3 (?) -0.008 (0.434)
PCit*RDit*AGit (?) 0.001** (0.019)
PCit*RDit*EPit-1 (?) 0.000 (0.337)
Year Dummies Yes Yes
Wald chi2 240.81(p-value<0.05) 221.86 (p-value<0.05)
Observations 785 785
Table 7 presents IV regression results for panel data. Heteroscedasticity-consistent p-values are
given in parentheses. The dependent variable is current period return, Rit. Rit, Rit+1, Rit+2 and Rit+3
are calculated as buy-and-hold returns (inclusive of dividends) over a 12-month period, starting
six months after the end of the previous financial year. Xit, Xit+1, Xit+2 and Xt+3 are defined as
earnings change deflated by price. Both current and future earnings changes are deflated by the
price at the start of the return window for period t. EPit–1 is defined as period t–1’s earnings
over price six months after the financial year-end of period t–1. AGit is the growth rate of the

34
Page 34 of 46
 total book value of assets for period t. RDit is the natural logarithm of the total number of risk
related sentences. Proprietary cost PCit is the profit margin ratio defined as Net
Income/Revenue. The significance levels (two-tail test) are: *= 10 %, ** =5 % and *** = 1 %.

6. Conclusion

Our paper addresses the question of whether voluntary risk disclosure in the annual reports
improves share price anticipation of future earnings in the context of MENA emerging

t
ip
markets. Alike the frontier financial markets, MENA stock exchanges are often characterized
by a weak flow of financial reporting and ineffective corporate governance mechanisms. It is

cr
suggested that corporate disclosure policy does not reflect the true underlying economic
conditions of firms in these markets. Investors are likely to be discouraged from capitalizing

us
on firm-specific information, thereby making stock prices inefficient (Farooq, 2013). It is
worthwhile then to analyze the credibility of disclosed information. The main hypotheses
predict that there is a positive association between voluntary risk disclosures and share price

an
anticipation of future earnings. Such association might be lower when proprietary costs exist.
Our empirical findings show that the ability of the market to anticipate future earnings
M
changes is positively related to the level of risk information. This evidences the usefulness of
risk disclosure in annual reports. Nevertheless, the level of proprietary costs moderates the
informational content of risk disclosure and investors are likely to rely on other sources of
d

information in anticipating future earnings growth.

te

We generate our estimates from panel regressions whereby we control for the cross-sectional
effects of some determinants of voluntary disclosure as well as the earnings response
p

coefficients such as risk, firm size and profitability. As a sensitivity test we re-estimate our
ce

regression models based on an instrumental variables approach. The results confirm to some
extent our initial findings in that risk disclosure seems to be useful to investors in anticipating
Ac

future earnings growth for one year ahead while product market competition reduces the
informativeness of such sensitive information.

The findings are of interest because they emphasize the informational content of risk
disclosure. Our results challenge the claims made by some recent studies, that risk
information is ambiguous, with limited use (Schrand & Elliot, 1998) and do not contain
reliable information (Campbell & Slack, 2008; Davies, Moxey, & Welch, 2010; Moxey &
Berendt, 2008).

35
Page 35 of 46
As regulators tighten corporate reporting standards all over the word in response to the recent
economic crisis, we document evidence that narrative risk disclosure reveals new information
about corporate risks and opportunities. Regulators and standard setters within the MENA
emerging markets should in this respect assure an increased focus on the informational needs
of investors in their accounting regulation. They should require firms to provide details about
the specific risks they face, given their importance to investors in forecasting firms’ future

t
ip
earnings. The findings in this study are also expected to provide regulatory bodies with useful
information about factors that influences the perceived relevance of risk disclosure. As such,

cr
regulators should not acknowledge the proprietary nature of risk disclosure when asking for
more corporate transparency. Wang & Hussainey (2013) assert that corporate governance

us
plays a more significant role in increasing transparency and improving investors’ ability to
predict future earnings. We think that their assertion should be the topic of future research in
risk disclosure literature. Moreover, we think that future studies should differentiate risk

an
disclosure that provides favorable from those that provide unfavorable earnings news when
analyzing the effect of the amount of risk information on share price anticipation of future
M
earnings.

References
d

Abraham S. and Shrives P.J. (2014) “Improving the relevance of risk factor disclosure in
corporate annual reports” The British Accounting Review, 46 (1), 91–107
te

Al-Akra, M., Eddie, I., and Ali, M. J. (2010a). The influence of the introduction of accounting
regulation on mandatory disclosure compliance: evidence from Jordan. British
Accounting Review, 42, 170–186.
p

Al-Akra, M., Eddie, I., and Ali, M. J. (2010b). The association between privatisation and
ce

voluntary disclosure: evidence from Jordan. Accounting and Business Research, 40(1),
55–74.
Al-Akra, M., Jahangir Ali, M., and Marashdeh, O. (2009). « Development of accounting
regulation in Jordan”. The International Journal of Accounting, 44(2), 163-186.
Ac

Ali A. Klasa S. and Yeung E. (2014) “Industry concentration and corporate disclosure
policy.” Journal of Accounting & Economics , 58(2-3), 240-264.
Allison, P. D. (1999). « Multiple regression: A primer. » Pine Forge Press.
Al-Omari, AM (2010), “The Institutional Framework of Financial Reporting in Jordan”,
European Journal of Economics, 22, 32-50.
Amran, A., A. M. R. Bin, and B. C. H. M. Hassan. (2009). “Risk reporting-An exploratory
study on risk management disclosure in Malaysian annual reports”. Managerial
Auditing Journal, 24 (1), 39–57.
Athanasakou, Vasiliki E. and Hussainey, Khaled (2014) “The perceived credibility of forward
looking performance disclosures”. Accounting and Business Research, 44 (3). pp. 227-
259.

36
Page 36 of 46
Baltagi (2005) “Econometric analysis of panel data”. John Wiley & Sons Ltd. The Atrium,
Southern Gate, Chichester
Bamber L. and Cheon Y. (1998). “Discretionary Management Earnings Forecast Disclosures:
Antecedents and Consequences of Forecast Venue and Forecast Specificity
Choices”. Journal of Accounting Research: 36(2): 167-190.
Bao Y. and A. Datta (2014). “Simultaneously Discovering and Quantifying Risk Types from
Textual Risk Disclosures”. Management Science, 60 (6), 1371-1391

t
Beattie, V. A., McInnes, B., and Fearnley, S. (2002) “Through the eyes of management: A

ip
study of narrative disclosure”. An interim report. London: Centre for Business
Performance, The Institute of Chartered Accountants in England and Wales.
Ben Naceur S, S. Ghazouani and M. Omran (2008) “Does stock market liberalization spur

cr
financial and economic development in the MENA region?” Journal of Comparative
Economics 36, 673–693.
Ben Othman, H., and D. Zeghal, (2008) “A study of corporate governance disclosure and its

us
country-level determinants in the emerging markets”. Research in Accounting in
Emerging Economies, 8, 125-155.
Beretta, S., and S. Bozzolan. (2004). “A framework for the analysis of firm risk

an
communication” The International Journal of Accounting 39 (3): 265–288.
Beyer, A, D. Cohen, T. Z. Lys, and Walther B. R., (2010) “The financial reporting
environment: review of the recent literature”. Journal of Accounting and Economics, 50,
296-343.
M
Bogdan V, P. C. Madalina and P. D. Nicoleta (2009) “Voluntary Internet Financial Reporting
and Disclosure – A New Challenge for Romanian Companies” Annals of the University
of Oradea, Economic Science Series; 18(3), 770-778
d

Botosan, C. A. (2004). “Discussion of‘‘A framework for the analysis of firm risk
communication”. The International Journal of Accounting, 39 (3): 289–295.
te

Botosan, C., and M. Harris. (2000). “Motivations for changes in disclosure frequency
and its consequences: An examination of voluntary quarterly segment
disclosures”. Journal of Accounting Research, 38, 329-354.
p

Botosan C. A., and M. Stanford, (2005) “Managers' Motives to Withhold Segment

Disclosures and the Effect of SFAS no. 131 on Analysts' Information Environment”.
ce

The Accounting Review, 8(3), 751-771.

Bowman, E. H. (1984). “Content analysis of annual reports for corporate strategy and risk”.
Interfaces, 14, 61-71.
Ac

Brambor, T., Clark, W. R., and Golder, M. (2006). “Understanding interaction models:
Improving empirical analyses”. Political analysis, 14(1), 63-82.
Brown, S. V., and Tucker, J. W. (2011). “Large-sample evidence on firms’ year-over-year
MD&A modifications”. Journal of Accounting Research, 49, 309-346
Campbell, D., and Slack, R. (2008). “Narrative reporting: Analysts’ perceptions of its value
and relevance”. London: The Association of Chartered Certified Accountants. RR104.
Campbell, J. L., Chen, H., Dhaliwal, D. S., Lu, M., and Steele, L. B. (2014). “The information
content of mandatory risk factor disclosures in corporate filings”. Review Accounting
Study,19, 396–455
Clarkson, P., J.L. Kao and G. Richardson (1994). “The Voluntary Inclusion of Forecasts in
the MD&A Section of Annual Reports.” Contemporary Accounting Research, 11,423-
450.

37
Page 37 of 46
Cohen, D., (2004). “Quality of Financial Reporting Choice: Determinants and Economic
Consequence”. Working paper, Northwestern University California
Collins, D. W., Kothari, S., Shanken, J., and Sloan, R., (1994) “Lack of timeliness versus
noise as explanations for low contemporaneous return–earnings association”. Journal of
Accounting and Economics 18, 289–324.
Cormier, D., Magnan, M., and van Velthoven, B. (2005). “Environmental disclosure quality
in large German companies: economic incentives, public pressures or institutional
conditions? “ European Accounting Review, 14(1), 3–39

t
ip
Darrough, M. (1993). “Disclosure policy and competition: Cournot vs.Bertrand”. Accounting
Review. 68 534–561.
Darrough, M., and Stoughton, N., (1990) “Financial disclosure policy in an entry

cr
game”. Journal of Accounting and Economics 12, 219–244.
Davies, J., Moxey, P., and Welch, I. (2010). “Risk and reward: Tempering the pursuit of
profit”. London: ACCA.

us
Dedman, E., and Lennox, C. (2009), “Perceived competition, profitability and the withholding
of information about sales and the cost of sales”. Journal of Accounting and Economics,
48(2-3), 210–230

an
Deumes R., (2008) “Corporate risk reporting: A content analysis of narrative risk disclosures
in Prospectuses”. Journal of Business Communication, 45(2), 120-157
Deumes R., and W. R. Knechel, (2008) “Economic Incentives for Voluntary Reporting on
Internal Risk Management and Control Systems” Auditing: A Journal of Practice &
M
Theory, 27(1), 35–66
Dietrich, J.R., Kachelmeier, S.J., Kleinmuntz, D.N., and Linsmeier, T.J., (2001). “Market
efficiency, bounded rationality, and supplemental business reporting disclosures”.
d

Journal of Accounting Research 39 (2), 213–268.

Dobler, M. (2008). “Incentives for risk reporting-A discretionary disclosure and cheap talk
te

approach”. The International Journal of Accounting, 43(2), 184–206.

Dobler, M., Lajili, K., and Zeghal, D. (2011). “Attributes of corporate risk disclosure: An
international investigation in the manufacturing sector”. Journal of International
p

Accounting Research, 10(2), 1–22.

Echambadi, R., and Hess, J. D. (2007). “Mean-centering does not alleviate collinearity
ce

problems in moderated multiple regression models”. Marketing Science, 26(3), 438-

445.
Elshandidy T., I. Fraser and K. Hussainey (2013). “Aggregated, voluntary, and mandatory
Ac

risk disclosure incentives: Evidence from UK FTSE all-share companies”. International

Review of Financial Analysis, 30, 320–333.
Elzahar H and Hussainey K, (2012) “Determinants of narrative risk disclosures in UK interim
reports”. The Journal of Risk Finance, 13 (2), 133 – 147.
Farooq, O. (2013). Does analyst following improve informativeness of earnings? Evidence
from the MENA region. International Journal of Business Governance and Ethics, 8(2),
181-194.
Financial Accounting Standards Board. (2010). Conceptual framework of financial reporting.
Chapter 1, The objective of general purpose financial reporting, and Chapter 3,
Qualitative characteristics of useful financial information. Statement of Financial
Accounting Concepts No. 8. Norwalk, CT.

38
Page 38 of 46
Ferreira, M., M. Massa, and P. Matos, (2010). “Shareholders at the gate? Institutional
investors and cross-border mergers and acquisitions”. Review of Financial Studies, 23,
601–644.
Francis, J., K. Schipper, (1999). “Have financial statements lost their relevance?” Journal of
Accounting Research,37, 319–353.
Fuller, J., and Jensen, M. C. (2002). “Just say no to Wall Street”. Journal of Applied
Corporate Finance, 14(4), 41-46.

t
Gelb, D. S. and Zarowin, P. (2002). “Corporate disclosure policy and the informativeness of

ip
stock prices”. Review of Accounting Studies, 7 (1): 33–52.
Gelb, S.D., (2000). “Corporate signaling with dividends, stock repurchases, and accounting
disclosures: an empirical study”. Journal of Accounting, Auditing & Finance, 99–121

cr
Gray, S. J. (1988). Towards a Theory of Cultural Influence on the Development of
Accounting Systems Internationally. Abacus, 24(1), 1-15.

us
Harris, M.S. (1998) ‘The association between competition and managers’ business segment
reporting decisions’, Journal of Accounting Research, 36(1): 111-128.
Harrison, B., and Moore, W. R. (2012). Stock Market Efficiency, Non-Linearity, Thin
Trading and Asymmetric Information in MENA Stock Markets. Economic Issues

an
Journal Articles, 17(1), 77-93.
Hassaan M. (2013) “Corporate Governance and Compliance with IFRSs: MENA Evidence”.
Cambridge Scholars Publishing
M
Hausman, J. A. (1978). “Specification tests in econometrics”. Econometrica: Journal of the
Econometric Society, 1251-1271.
Hayes, A. F., and Krippendorff, K.(2007).“Answering the call for a standard reliability
measure for coding data.” Communication Methods and Measures, 1, 77-89.
d

Healy, P., Palepu, K., (2001). “Information asymmetry, corporate disclosure and the capital
markets: A review of the empirical disclosure literature”. Journal of Accounting &
te

Economics 31, 405–440.

Healy, P., Hutton, A., and Palepu, K., (1999). “Stock performance and intermediation changes
p

surrounding sustained increases in disclosure”. Contemporary Accounting Research, 16,

485- 520.
ce

Hossain M, S. M. Courtenay., and S. Luo (2006), “The effect of voluntary disclosure,

ownership Structure and proprietary cost on the return–future earnings relation”.
Pacific-Basin Finance Journal, 14,501–521
Hussainey, K. and Mouselli, S. (2010) “Disclosure quality and stock returns in the UK”.
Ac

Journal of Applied Accounting Research, 11(2), 154-174

Hussainey, K. and Walker, M. (2009) “The effects of voluntary disclosure and dividend
propensity on prices leading earnings.” Accounting and Business Research, 39(1), 37-
55.
Hussainey, K., Schleicher, T. and Walker, M. (2003). ‘Undertaking large-scale disclosure
studies when AIMR-FAF ratings are not available: the case of prices leading earnings’.
Accounting and Business Research, 33 (4): 275–294.
Institute of Chartered Accountants In England and Wales (ICAEW). (2002). “Financial
reporting of risk: Proposals for a Statement of Business Risk”. London.
IFC (International Financial Corporation) and Hawkamah (2008), “A Corporate Governance
Survey of Listed Companies and Banks Across MENA”. Washington DC: IFC; and
Dubai: Hawkamah.

39
Page 39 of 46
Irvine, P. J., and Pontiff, J. (2009). Idiosyncratic Return Volatility, Cash Flows, and Product
Market Competition. Review of Financial Studies, 22(3), 1149-1177.
Ismail, W. A. W., Kamarudin, K. A., Van Zijl, T., and Dunstan, K. (2013). Earnings quality
and the adoption of IFRS-based accounting standards: Evidence from an emerging
market. Asian Review of Accounting, 21(1), 53-73.
Jorion, P. (2002). How informative are value-at-risk disclosures? The Accounting Review, 77,
911-931.

t
Kaju¨ter, P., (2001). “Risikoberichterstattung: Empirische Befunde und der Entwurf des DRS

ip
5”. Der Betrieb 54 (3), 105–111.
King R. R. and D. E. Wallin (1995) “Experimental tests of disclosure with an opponent”.
Journal of Accounting and Economics 19, 139-167.

cr
Kothari S. P., X. Li, and J. E. Short (2009) “The Effect of Disclosures by Management,
Analysts, and Business Press on Cost of Capital, Return Volatility, and Analyst
Forecasts: A Study Using Content Analysis”. The Accounting Review, 84(5), 1639-1670

us
Kravet T.D and Muslu V. (2013) “Textual Risk Disclosures and Investors’ Risk Perceptions”.
Review of Accounting Studies, 18(4), 1088-1122
Krippendorff, K. (2010). “Krippendorff's alpha”. In N. Salkind (Ed.), Encyclopedia of

an
research design. (pp. 669-674). Thousand Oaks, CA: SAGE Publications, Inc. doi:
http://dx.doi.org/10.4135/9781412961288.n206
Krippendorff, K., (1980). “Content Analysis: An Introduction to Its Methodology”. Sage,
Beverly Hills.
M
Krippendorff, K., (2004). “Content Analysis: An Introduction to Its Methodology”, second ed.
Sage, Beverly Hills.
Larcker, D. F., Rusticus, T. O. (2010). On the use of instrumental variables in accounting
d

research. Journal of Accounting and Economics, 49(3), 186-205.

Lehavy, R., Li, F., and Merkley, K. (2011). “The effect of annual report readability on analyst
te

following and the properties of their earnings forecasts”. The Accounting Review, 86,
1087-1115.
p

Leuz, C. 2004 “Proprietary versus Nonproprietary Disclosures: Evidence from Germany”

in C. Leuz, D. Pfaff and A. G. Hopwood: The Economics and Politics of Accounting:
ce

International Perspectives on Research, Trends, Policy, and Practice. Oxford

University Press, pp. 164-97
Li, F. (2008). “Do stock market investors understand the risk sentiment of corporate annual
reports?” Unpublished paper, University of Michigan.
Ac

Li, F. (2010). “The information content of forward-looking statements in corporate filings-A

naïve Bayesian machine learning approach”. Journal of Accounting Research, 48, 1049-
1102.
Lim, C. Y. and Tan, P. M. (2007) “Value relevance of value-at-risk disclosure”. Review of
Quantitative Finance and Accounting, 29(4), 353–370.
Linsley, P., and Shrives, P. (2006) “Risk reporting: A study of risk disclosure in the annual
reports of UK companies”. The British Accounting Review, 38(4), 387–404.
Linsmeier, T. J., Thornton, D. B., Venkatachalam, M., and Welker, M. (2002). “The effect of
mandated market risk disclosures on trading volume sensitivity to interest rate,
exchange rate, and commodity price movements”. The Accounting Review 77, 343-377.

40
Page 40 of 46
Liu, C., S. Ryan, and H. Tan (2004): “How Banks’ Value-at-Risk Disclosures Predict their
Total and Priced Risk: Effects of Bank Technical Sophistication and Learning over
Time.” Review of Accounting Studies, 9, 265-294
Lundholm, R. J. and Myers, L. A. (2002) “Bringing the future forward: the effect of
disclosure on the return-earnings relation”. Journal of Accounting Research, 40 (3):
809–39.
Marshall, A. P., and Weetman, P. (2007). “Modelling transparency in disclosure: the case of
foreign exchange risk management”. Journal of Business Finance and Accounting, 34(5

t
& 6), 705–739.

ip
Monk D. (2011) “Corporate Form and Proprietary Costs of Voluntary Disclosure”Working
paper, Rutgers University available at:

cr
http://accounting.rutgers.edu/docs/seminars/spring11/Monk_JobMarket_VoluntaryDisc
losure.pdf
Moxey, P., and Berendt, A. (2008)”Corporate governance and the credit crunch” London:

us
ACCA.
Nichols D.C. (2009) “Proprietary Costs and Other Determinants of Nonfinancial Disclosures”
AAA 2010 Financial Accounting and Reporting Section (FARS) Paper. Available at

an
http://dx.doi.org/10.2139/ssrn.1465768
Nobes, C. (1998), Towards a General Model of the Reasons for International Differences in
Financial Reporting. Abacus, 34: 162–187.
Pérignon C. and D.R. Smith, (2010) “The level and quality of Value-at-Risk disclosure by
M
commercial banks”. Journal of Banking & Finance, 34(2), 362–377
Piotroski, J., (2001). “Discretionary segment reporting decisions and the precision of investor
beliefs”. Working paper, University of Chicago
d

Prencipe A. (2004). “Proprietary costs and voluntary segment disclosure: evidence from
Italian listed companies”. European Accounting Review13(2), 309-340.
te

Rajgopal, S. (1999) “Early evidence on the informativeness of the SEC’s market risk
disclosures: the case of commodity price risk exposure of oil and gas producers”, The
Accounting Review, 74(3), 251–280.
p

Raonic A., Dargenidou C., and Mc. Stuart (2011) “Accruals, Disclosure and the Pricing of
Future Earnings in the European Market”. Journal of Business Finance & Accounting,
ce

38(5) & (6), 473–504

ROSC (Report on the observance of standards and codes ), (2009) “A Corporate Governance
Country Assessment for The Arab Republic of Egypt”. WB: Available at
Ac

http://www.worldbank.org/ifa/rosc_cg_egy.pdf
Samaha, K., & Dahawy, K. (2011). An empirical analysis of corporate governance structures
and voluntary corporate disclosure in volatile capital markets: the Egyptian
experience. International Journal of Accounting, Auditing and Performance
Evaluation, 7(1), 61-93.
Schleicher T., Hussainey K., Walker M. (2007) “Loss firms’ annual report narratives and
share price anticipation of earnings”. The British Accounting Review 39,153 – 171
Schrand C. and J. Elliot, (1998) “Risk and financial reporting: A summary of the discussion
at the 1997 AAA/FASB conference”. Accounting Horizons, 12(3), 271-282.
Shin, Y. (2001) “The Effect of Product Market Competition on Corporate Voluntary
Disclosure Decisions.” Working paper, Tulane University

41
Page 41 of 46
Skinner, D. J. (1994). “Why firms voluntarily disclose bad news”. Journal of Accounting
Research, 32(1), 38–60
Solomon, J., Solomon, A., Norton, S., & Joseph, N. (2011). “Private climate change
reporting: an emerging discourse of risk and opportunity?” Accounting, Auditing and
Accountability Journal, 24(8), 1119–1148
Stiglitz, J. E. (1987). “Competition and the number of firms in a market: Are duopolies more
competitive than atomistic markets?”. The Journal of Political Economy, 1041-1061.

t
Sutton, J., (1991). “Sunk Costs and Market Structure: Price Competition, Advertising and the

ip
Evolution of Concentration”. The MIT Press, Cambridge, MA.
Tang V. W. (2010) “Proprietary Costs of Mandatory Disclosure and the Decision to First
Access the Public Market”. Working Paper, Georgetown University.

cr
The World Bank (2012) “Doing Business” Available online at
http://www.doingbusiness.org/~/media/GIAWB/Doing%20Business/Documents/Annual-
Reports/English/DB12-FullReport.pdf

us
Unerman (2000) "Methodological issues - Reflections on quantification in corporate social
reporting content analysis", Accounting, Auditing & Accountability Journal, 13(5), 667
- 681

an
Verrecchia, R. E. (1983). “Discretionary disclosure”. Journal of Accounting and Economics,
5, 179–194.
Verrecchia, R. E. and Weber, J. (2006), “Redacted Disclosure”. Journal of Accounting
Research, 44: 791–814.
M
Wang, M. and Hussainey, K. (2013). ‘'Voluntary forward-looking statements driven by
corporate governance and their value relevance'. Journal of Accounting and Public
Policy,32 (3), 26-49.
d

Wooldridge, J. M. (2010). “Econometric analysis of cross section and panel data”. MIT
press.
p te
ce
Ac

42
Page 42 of 46
Appendix A1. Risk disclosure categories adopted from Linsley & Shrives (2006)
Operations Risk Customer Satisfaction
Product Development
Efficiency And Performance
Sourcing
Stock Obsolescence And Shrinkage
Product And Service Failure

t
Environmental

ip
Health And Safety
Brand Name Erosion
Empowerment Risk Leadership And Management

cr
Outsourcing
Performance Incentives
Change Readiness

us
Communications
Information Processing And Technology Risk Integrity
Access
Availability

an
Infrastructure
Integrity Risk Management And Employee Fraud
Ilegal Acts
Reputation
M
Strategic Risk Environmental Scan
Industry
Business Portfolio
Competitors
d

Pricing
Valuation
te

Planning
Life Cycle
Performance Measurement
p

Regulatory
Sovereign And Political
ce

Appendix A2. Type of Risk and Its Definition

Ac

Type of Risk Definition (in accordance to ICAEW)

Operations risk The risk that operations are inefficient and ineffective in
executing the firm’s business model, satisfying customers and
achieving the firm’s quality, cost and time performance
objectives.

Empowerment risk The risk that managers and employees (1) are not properly led, (2)
do not know what to do when they need to do it, (3) exceed the
boundaries of their assigned authorities, or (4) are given
incentives to do the wrong thing.

43
Page 43 of 46
Information technology The risk that the information technologies used in the firm (1) are
risk not operating as intended, (2) are compromising the integrity and
reliability of data and information, (3) are exposing significant
assets to potential loss or misuse, or (4) are exposing the firm’s
ability to sustain the operation of critical processes.

Integrity risk The risk of management fraud, employee fraud, illegal acts and

t
unauthorised acts, any or all of which could lead to reputation loss

ip
in the marketplace.

Strategic risk The risk associated with future business plans and strategies,

cr
including plans for entering new business lines, expanding
existing services through mergers and acquisitions, enhancing
infrastructure, etc.

us
Appendix B:
Collinearity Diagnostics

an
SQRT R-
Variable VIF VIF Tolerance Squared
----------------------------------------------------
Rt 1.19 1.09 0.8410 0.1590
Xt 2.66 1.63 0.3761 0.6239
Xt1 2.45 1.57 0.4083 0.5917
M
Xt2 2.85 1.69 0.3503 0.6497
Xt3 2.52 1.59 0.3966 0.6034
Rt1 6.89 2.62 0.1452 0.8548
Rt2 6.00 2.45 0.1666 0.8334
Rt3 4.24 2.06 0.2360 0.7640
Agt 1.23 1.11 0.8139 0.1861
Ept 1.48 1.22 0.6761 0.3239
d

riskdisct 1.13 1.06 0.8870 0.1130

RdXt 2.91 1.71 0.3432 0.6568
RdXt1 2.73 1.65 0.3658 0.6342
te

RdXt2 3.08 1.75 0.3248 0.6752

RdXt3 2.62 1.62 0.3815 0.6185
RdRt1 6.80 2.61 0.1471 0.8529
RdRt2 5.95 2.44 0.1682 0.8318
RdRt3 4.25 2.06 0.2351 0.7649
p

RdAgt 1.17 1.08 0.8537 0.1463

RdEpt 1.41 1.19 0.7111 0.2889
Lev 1.06 1.03 0.9427 0.0573
ce

Size 1.08 1.04 0.9293 0.0707

profit 1.22 1.11 0.8168 0.1832
----------------------------------------------------
Mean VIF 2.91
Ac

44
Page 44 of 46
 Collinearity Diagnostics

SQRT R-
Variable VIF VIF Tolerance Squared
----------------------------------------------------
Rt 1.23 1.11 0.8138 0.1862
Xt 3.08 1.75 0.3248 0.6752
Xt1 3.15 1.78 0.3171 0.6829
Xt2 3.44 1.85 0.2908 0.7092
Xt3 3.25 1.80 0.3072 0.6928
Rt1 7.69 2.77 0.1301 0.8699
Rt2 6.67 2.58 0.1500 0.8500
Rt3 4.72 2.17 0.2121 0.7879

t
Agt 1.96 1.40 0.5113 0.4887

ip
Ept 3.94 1.98 0.2539 0.7461
riskdisct 1.15 1.07 0.8675 0.1325
RdXt 3.09 1.76 0.3232 0.6768
RdXt1 2.99 1.73 0.3349 0.6651

cr
RdXt2 3.31 1.82 0.3021 0.6979
RdXt3 2.83 1.68 0.3534 0.6466
RdRt1 6.99 2.64 0.1430 0.8570
RdRt2 6.38 2.53 0.1567 0.8433
RdRt3 4.40 2.10 0.2275 0.7725

us
RdAgt 1.21 1.10 0.8248 0.1752
RdEpt 1.49 1.22 0.6693 0.3307
Pct 1.76 1.33 0.5666 0.4334
PcXt 7.53 2.74 0.1327 0.8673
PcXt1 6.63 2.57 0.1509 0.8491
PcXt2 5.22 2.29 0.1915 0.8085

an
PcXt3 5.82 2.41 0.1719 0.8281
PcRt1 10.59 3.25 0.0945 0.9055
PcRt2 10.00 3.16 0.1000 0.9000
PcRt3 12.33 3.51 0.0811 0.9189
PcAgt 10.40 3.23 0.0961 0.9039
PcEpt 19.24 4.39 0.0520 0.9480
M
PcRdXt 2.04 1.43 0.4894 0.5106
PcRdXt1 2.01 1.42 0.4978 0.5022
PcRdXt2 2.02 1.42 0.4959 0.5041
PcRdXt3 1.65 1.28 0.6076 0.3924
PcRdRt1 2.19 1.48 0.4560 0.5440
PcRdRt2 2.32 1.52 0.4311 0.5689
PcRdRt3 1.74 1.32 0.5750 0.4250
d

PcRdAgt 2.54 1.59 0.3933 0.6067

PcRdEpt 4.92 2.22 0.2032 0.7968
----------------------------------------------------
te

Mean VIF 4.72

Collinearity Diagnostics
p

SQRT R-
ce

Variable VIF VIF Tolerance Squared

----------------------------------------------------
Rt 1.15 1.07 0.8676 0.1324
Xt 1.15 1.07 0.8674 0.1326
Xt1 1.19 1.09 0.8373 0.1627
Ac

Xt2 1.17 1.08 0.8561 0.1439

Xt3 1.07 1.04 0.9325 0.0675
Rt1 1.25 1.12 0.7981 0.2019
Rt2 1.10 1.05 0.9068 0.0932
Rt3 1.05 1.03 0.9518 0.0482
Agt 1.06 1.03 0.9416 0.0584
Ept 1.48 1.22 0.6755 0.3245
riskdisct 1.13 1.06 0.8833 0.1167
Pct 1.27 1.13 0.7883 0.2117
Lev 1.10 1.05 0.9128 0.0872
size 1.50 1.22 0.6679 0.3321
profit 1.22 1.11 0.8187 0.1813
----------------------------------------------------
Mean VIF 1.19

45
Page 45 of 46
[Article title] The Value Relevance of Risk Disclosure in Annual Reports: Evidence from MENA Emerging Markets
(RIBAF 350)

[Author Details]

Author 1 Name: Néjia Moumen

Department: Accounting and Finance
University: Higher Institute of Accounting and Business Administration (ISCAE), University of Manouba
Town/City: Tunis
Country: Tunisia

Author 2 Name: Hakim Ben Othman

Department: Accounting and Finance

t
University: Tunis Business School (TBS), University of Tunis & LIGUE-ISCAE, University of Manouba

ip
Town/City: Tunis
Country: Tunisia

Author 3 Name: Khaled Hussainey

cr
Department: Accounting and Finance
University: School of Management, Plymouth University
Town/City: Plymouth, England
Country: UK

us
Corresponding author: Miss Néjia Moumen
Corresponding Author’s Email: mounej.tn@gmail.com
Corresponding Author’s Mobile: (216) 55388795

Biographical Details:

[Author 1 bio] PhD student in Accounting

an
M
[Author 2 bio] Professor of Accounting

[Author 3 bio] Professor of Accounting

d
p te
ce
Ac

Page 46 of 46