Sie sind auf Seite 1von 36

Incremental Voluntary Disclosure on Corporate Websites, Determinants and Consequences

Samir T r a b e l s i a * , Real L a b e l l e b * and Pascal D u m o n t i e r ' *


"Brock University
b~~~

Montrkal

'University of Grenohle Received September 2007; Accepted May 2008

Abstract

This study analyses the determinants and consequences of internet financial reporting

(IFR). evidence indicates that firms use the internet to report complementary inforOur
mation on firm background, management forecasts, intangible assets and on social and environmental issues. Our results indicate that the decision to provide additional voluntary financial disclosures through corporate websites is mostly influenced by share turnover, the future profitability of the firm and the level of competition in the industry. Last, we find that the extent of voluntary disclosure on corporate websites is related positively to forecast accuracy, and negatively to the dispersion of analysts forecasts, suggesting that such disclosures provide useful information to analysts.
JEL Classifications: G14, M41, M45,033
Keywords: internet financial reporting, voluntary disclosure theory, content analysis

* We acknowledge support from SSHRC's Initiative on the New Economy Program. We would like to thank seminar participants from Brock University; from the University of Waterloo and Rutgers Business School at the ls' Continuous Reporting Symposium; from Paris- I-Sorbonne and Toulouse I University workshops; from the 2008 AJPT/JCAE! Joint Symposium; from the 2005 CAAA annual meeting; and from the 2006ABR (Business) conference sponsored by the Clute Institute for Academic Research. We thank also Michael Alles, Gaktan Breton, Ken Calssen, John Core, Darlene Bay, Patricia O'Brain, Robert Gagne. Steve Huddart, Claude Laurin, James Moore, Sameer Mustapha, Patricia Myers, Steven Salterio, Dan Simunic, Miklos A. Vasarhelyi,Ari Yezegel and especially Gordon Richardson and Gerald Trites for their many helpful comments. Finally we thank Cameron Maccinis and John Hughes at the Ontario Securities Commission, Christine Botosan for supplying us with her disclosure scoring sheet, and Gerry Trites, Steve Bonin and Victoria Neal for helping to validate the disclosure scores. The usual caveat applies.

Samir Trabelsi, Rkal Labelle and Pascal Dumontier 121 Journal o Contemporary Accounting & Economics Vol4,N o 2 (December 2008) 120-155 f

1. Introduction
This paper is aimed at gauging the incremental contribution of the internet to a firms financial reporting policy. It is also aimed at analysing the determinants and consequences of this contribution. As a first step, we isolate the influence of the internet by quantifying the information displayed on corporate websites that is in addition to that conveyed in traditional financial reporting (TFR) sources.TFR sources are defined as annual reports and other documents electronically reproduced on SEDAR and that are otherwise available to investors without regard to a companys website. Second, using the predictions of voluntary disclosure theories, we examine the motives leading some firms to extend disclosure and to broaden access to financial information through their websites. Finally, as a third step, using financial analystsforecasts we investigate the impact of corporate website financial reporting on information asymmetry. The use of the internet to report financial information has been drawing the attention of both standard setters and regulators who recognise that internet financial reporting (IFR) may provide more complete and more timely information to investors, thus transcending the paper paradigm in terms of content and presentation (TSE, 1999). However, despite the well-documented and widespread use of electronic financial communications, the regulatory focus continues to be driven by the TFR model. Even when discussing dissemination of financial information through the internet, regulators (OSC, 2004) are still mainly concerned with the documents which are exact reproductions of the printed ones (FASB, 2000,48). While recognising that IFR poses new regulatory challenges with regard to protecting investors from unreliable information, regulators have yet to make the necessary adjustments. This research extends and contributes to the prior literature on voluntary disclosure and on IFR in several ways. First, because disclosures on corporate websites are truly voluntary, it provides more powerful tests of the determinants of voluntary disclosure than those dedicated to annual report disclosures which, to a large extent, are mandatory. Our analysis of the determinants of IFR indicates that the decision to use a website to enhance financial disclosure is influenced by the firms expected future performance; by its ownership structure; by the informativeness of its accounting data; and by the level of competition in the industry. Second, prior studies devoted to IFR, such as Debreceny et al. (2002) and Ettredge et al. (2002), analyse the information available on corporate websites without considering that the same information may also be available in TFR sources. In contrast, this study appraises the net contribution of financial reporting through the internet by focusing on incremental IFR information. Since, as pointed out by Ashbaugh et al. (1999), IFR allows firms to disclose specific information exclusively available on their websites as well as detailed financial data intended to clarify related TFR disclosures, we quantify both the incremental and the disaggregated IFR information. Incremental

Since January 1997 issuers are required by the Canadian Securities Administrators (CSA) to archive their statutory TFR on an online public system, namely the System of Electronic Document Analysis and Retrieval (SEDAR). Likewise in the US the Securities and Exchange Commission (SEC) requires issuers to file using the Electronic Data Gathering and Retrieval (EDGAR) system. For further information consult SEDAR.coms Rules and Forms page (www.sedar.com/sedar/sedar rules forms en.htm) and www.sec.gov/edgar.shtml.

122

Samir Trabelsi, Rial Labelle and Pascal Durnontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

information refers to the voluntary information disclosed exclusively on the firms website. Disaggregated information refers to the firms explanation of its financial disclosure in TFR. To this end, we compute net disclosure scores for a sample of Canadian firms from separate content analyses performed on each firms website and a set of TFR documents filed on SEDAR.These scores indicate that only half of our sample firms use their websites to report incremental and disaggregated financial information and that they generally take advantage of the internet to report complementary information on the firms background, management forecasts,intangible assets and social and environmentalissues, and to provide non-financial statistics that are not reported in TFR media. Considering that incremental and disaggregated corporate disclosures through the internet are primarily aimed at reducing asymmetric information between managers and outsiders, the third contribution of this paper relies on an examination of the informational consequences of IFR. Using analysts forecasts to proxy for information asymmetry, we show that forecast dispersion and forecast errors both decrease with the magnitude of incremental and disaggregatedinformation disclosed on corporate websites,suggestingthat such disclosures provide useful information to financial analysts and therefore to market participants.The fourth contribution of this study is that we control for potential selection bias resulting from two interrelated decisions faced by managers with regard to IFR: whether to use the firmswebsite to provide market participants with additional information and the nature and extent of the additional information to be disclosed.Prior studies exclude (Xiao et al., 2004) or assign a disclosure score of zero (Debreceny et al., 2002) to firms that did not have a website that was accessible during their examination, which implies a potential selection bias. To correct for this bias, we use the two-stage estimation procedure developed by Heckman (1979) both for the analysis of the determinants of the use of IFR and for the analysis of the IFR impact on information asymmetry. The remainder of the paper is organised as follows. Section 2 briefly reviews prior relevant IFR literature and develops hypotheses. Section 3 describes the research design. Section 4 presents and discusses our empirical findings. Section 5 provides a summary and conclusion. 2. Prior Research and Hypotheses Development In this section we demonstrate how this study builds on prior research devoted to the emergence and evolution of IFR. Then, by adapting voluntary disclosure theories to take the particular nature of IFR into consideration, we develop hypotheses on the determinants of the firms decision to use its website to extend disclosure and to broaden access to financial information. Last, we justify the use of analysts forecasts to examine the impact of IFR on firm information asymmetry. 2.1 The internetjnancial reporting literature According to the FASB (2000,p. vii) more than anything else, the internet has expanded the amount of information available to non-specialist investors and allowed delivery of that information at no cost or very low cost. Speed of delivery and ease of access are

Samir Trabelsi, RLal Labelle and Pascal Dumontier 123 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

important, but, today, any investor can obtain information that was previously available, as a practical matter, only to company officials, professional investment analysts, and the financial press. However, the level of IFR varies among firms. Pirchegger and Wagenhofer (1999) document a positive association between the magnitude of IFR and shareholder dispersal, suggesting that IFR is mainly addressed to shareholders of widely held firms. F Using US data, Ashbaugh et al. (1999) find no relation between the use of I R and firm ownership structure as measured by the percentage of shares held by individual investors. Using an agency framework, Debreceny et al. (2002) hypothesise that managers use IFR to reassure lenders about the firms capacity to meet its obligations. Their results do not support their predictions. These studies suffer from research design limitations that may explain the conflicting results and the weak explanatorypower of models under study.First, by using a dichotomous variable to proxy for IFR, they do not consider the actual cross sectional variability of IFR. Second,they focus on the presentation format, the technology used and the support provided rather than on the extent of disclosure. Third, they ignore most of the variables explaining voluntary disclosure identified in the literature (Lang and Lundholm, 1993; Clarkson et al., 1999; Healy and Palepu, 2001, Trabelsi et al., 2004). Finally, when they exclude firms that do not publish financial information on their websites (Pirchegger and Wagenhofer, 1999; Ettredge et al., 2002), they suffer from a selection bias that could result in inefficient and inconsistent regression estimates (GouriCroux, 1989; Maddala, 1991; Shehata, 1991; Greene, 2003). Our study contributes to the literature by addressing these concerns.

2.2 Voluntary disclosure literature and determinants o voluntary IFR f


The steady increase in the number of firms using websites for financial reporting purposes suggests that IFR is becoming standard practice for most firms. However, during our sample period, several firms did not make available any IFR, while others voluntarily provided various levels of information on their websites over that available in TFR. The decision to use corporate websites to supplement the available information may depend on the manner in which the firm expects to benefit from the more complete and timely business reporting. We propose four main motivations for managers to use IFR to extend and broaden access to financial information: the nature of the firms investor base; the nature or quality of its TFR; the expected level of future profitability; and the level of competition in the industry. In a clienteleeffect scenario,IFR is provided to help users, especially small or foreign investors, who do not have direct and immediate access to financial information. IFR in this case is intended to reduce information asymmetry,therefore increasing the liquidity of the firms stocks and reducing its cost of capital (Verrecchia, 1983,2001; Botosan, 1997; Healy and Palepu, 2001). In a study of open versus closed conference calls, Bushee et al. (2003,178) find that firms provide open conference calls in order to meet non-professional investors demands for more widely available information. Wang (1993) argues that less informed investors demand the broadest reporting of financial information so as to avoid trading against better informed investors. Botosan and Harris (2000) also document that firms respond to investorsdemands for more informationby changing the frequency of their voluntary quarterly segment disclosures. In line with these studies, we postulate that firms

124

Samir Trubelsi, Rial Labelle and Puscul Dumontier Journal of Contemporary Accounting C Economics Vol4, No 2 (December 2008) 120-155 ?

with frequently traded shares are likely to be under greater pressure to disclose additional information on their websites and broaden information dissemination practices to reduce information asymmetry and increase liquidity.Thus, firms with high share turnover (ST), as measured by the ratio of the mean monthly trading volume over the average number of outstanding shares, are more likely to use IFR. This leads to hypothesis 1 (H,) in Table 2. Consistent with the clientele effect scenario, we also posit that firms with concentrated ownership are less concerned with the extent of their financial reporting than widely held ones, as their shareholders can obtain information directly from the firm (Gelb, 2000; Bushee et al., 2003). Furthermore, these firms may want to preserve their informational advantage by not using dissemination media such as IFR. This leads to H, in Table 2, where ownership concentration (OC) is measured by the proportion of shares held by the firms executives or other investors holding more than 10 percent of the firms shares. Finally, in line with the clientele effect hypothesis, we hypothesise that cross-listed firms are more inclined to take advantage of IFR to reach foreign investors (H3,Table 2). Cross listing (LISTING) is a dichotomous variable equal to 1 if the firm is listed on a foreign market and 0 otherwise. Informationquality refers to the extent to which TFR provides informationabout a firms operations. Firms for which value is relatively more difficult to assess using conventional accountingdata (Amir and Lev, 1996) or firms whose value is mainly composed of growth options (Myers, 1977) may be more likely to use IFR to supplement TFR. These firms, especially those not followed by a large number of analysts, may also wish to bypass informational intermediariesby disclosing information directly to all investors via IFR. To proxy for the relevance of the firms TFR, we use its book-to-market ratio (RELEV) as in Bushee et al. (2003), its R&D expenditures as in Huddart and Ke (2007) and the degree of information asymmetry as measured by the standard deviation of the firms adjusted stock returns for the last ten years (ASYM). The lower the book-to-market ratio or the higher the R&D expenditures and information asymmetry, the greater the firms motivation to disclose additional information on its website (H4,H, and H,, Table 2). Verrecchia (1983) argues that managers tend to increase voluntary disclosure if they have good news to report or expect an increase in future profits. With regard to this good news hypothesis, the empirical evidence is mixed. On the one hand, Singhvi and Desai ( 1971),Lev and Penman ( 1990),Lang and Lundholm ( 1993) and Wallace et a1.( 1994) find that profitable firms tend to use more forthcoming disclosure practices. On the other hand, Wallace and Naser (1995) find a negative relationshipbetween expected profitabilityand the extent of disclosures.Ashbaugh et al. (1999) fail to display any relationship. Furthermore, firms with profitable prospects are also more likely to plan security offerings (Gibbins et al., 1990; Lang and Lundholm, 1993; Clarkson et al., 1999). Consequently, since IFR is a voluntary disclosure mechanism, we expect EX firms to be more profitable and more likely to have financing activities (H7 H,, Table 2). As a proxy for these motivations to and extend IFR voluntary disclosures, we measure firm performance as in Lang and Lundholm (1993) and Clarkson et al. (1999) using a dummy variable (PERF) equal to 1 if the firms net income is larger in 2002 than in 2001, and 0 otherwise. Financing activities (FIN) are taken into consideration using a dichotomous variable equal to 1 if the firm has issued shares in 2001 or plans to do so in 2002 or 2003. FIN equals 0 otherwise. Finally, a firms strategic reporting policy may also depend on the likelihood that new

Samir Trabelsi, R6al Labelle and Pascal Dumontier 125 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

competitors enter its market. Hayes and Lundholm (1996) predict that managers mitigate this potential cost of proprietary information through non-disclosure. Clarkson et al. (1999) find a negative association between disclosures in the management discussion and analysis section of annual reports (MD&A) and the level of competition suggesting that firms enjoying greater abnormal profits are not forthcoming to conceal the source of their profitability. If this theoretical prediction holds in the context of IFR, we anticipate a negative relation between the level of competition (COMPET) and the decision to use IFR (H9, Table 2 ) . Competition is measured by the average return on equity (ROE) over the five preceding years as in Lang and Lundholm (1993); Hayes and Lundholm (1996); and Clarkson et al. (1999). Following Piotroski and Roulstone (2004), we control for industry concentration using a quarterly revenue-based Herfindahl index (HERF) which helps approach the level of competition from an industry-wide perspective (HJ. Competition is low when industrys sales are concentrated in a few firms (high HEW). Competition is likely to be more severe when an industrys sales are evenly distributed across firms (low HERF). In addition,we also control for various other factorsknown to be related to the voluntary disclosure channels of firms overall financial reporting package such as firm size, analyst following, audit quality and the industry.

2.3 The impact o IFR on analysts forecasts f


If IFR is effective in improving the information environment of firms, and in lowering information asymmetry, IFR should result in more accurate analysts forecasts. IFR disclosures, such as the ones captured by our disclosure scores, can improve analyst forecasting in many ways. Timely access to trend analyses, background information and summariesof historical results can help analystsassess the financial position and performance of the firm. Management discussion and analysis, projected information about the firms opportunities and risks, data on intangible assets and social or environmental background information help them assess growth opportunities and earnings sustainability. Analysts forecasts are based both on common information shared by all market participants, and on private or idiosyncratic information resulting from specific data collection or from private interpretation of common information. Barron et al. (1998) suggest that mean forecast errors reflect the quality of analystscommon information,while forecast dispersion captures the precision of their private information. Because corporate websites are easily accessible, we expect the quality of analysts common information to increase with the extent of disclosure on corporate websites. With regard to private information,two conflicting outcomes can be expected from increased corporate disclosure through the internet. On the one hand, if analysts idiosyncratic information comes from private communications with the management of firms they follow, IFR should reduce private information relative to common information,resulting in lower forecast dispersion. On the other hand, inasmuch as analysts process publicly available information to infer new private information, divergence of analystsbeliefs, and therefore forecast dispersion, might increase with the magnitudeof additionalinformationprovided on corporate websites. Several prior studies have investigated whether the properties of analysts forecasts, i.e., forecast errors and forecastdispersion are related to the extent of voluntary disclosure.Lang

126

Sumir Trabelsi, R4al Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

and Lundholm (1996) document that firms with more informative disclosure exhibit more accurate and less dispersed forecasts. They also exhibit less volatility in forecast revisions. In an international setting, Hope (2003) shows that forecast accuracy is positively related to the level of firm disclosure. Byard and Shaw (2003) find that higher quality disclosure increases the precision of both the public information and the private information that analysts incorporate into their forecasts. Based on the above discussion, using forecast accuracy as a proxy for the quality of publicly available information, we hypothesise that the extent of incremental and disaggregated disclosures on corporate websites is positively associated with analysts forecast accuracy. Given the competing predictions related to the impact of IFR on analysts private information, using forecast dispersion as a proxy for the quality of private information, we hypothesise that providing supplemental disclosures through the internet can either increase or decrease the dispersion of forecasts. We measure forecast accuracy (ACCURACY) similar to Lang and Lundholm (1996) as the absolute difference between mean earnings forecasts and actual earnings. We measure forecast dispersion as the standard deviation of analysts forecasts. Both measures are scaled by stock price.

3. Research Design
3.1 Sample selection

As shown in Table 1, an initial sample of 180 firms was randomly drawn from the population of firms listed on the Toronto Stock Exchange (TSE) using the StockGuide database. Of the 180 firms selected, 16 firms belonging to the financial sector were excluded as their particular characteristics could induce distortion in the data, leaving a subsample of 164 firms or approximately 15 percent of the TSE population. Ten firms were excluded because they were suspended or removed from the StockGuide database. During the data collection period which lasted four months from September to December 2002, we used a service called InfoMinder to warn us of any website change made by the sample firms between the time we performed the content analysis and the end of the sampling period. InfoMinder allowed us to take all the time necessary to collect and thoroughly compute each firms IFR scores while making sure that the information was collected at the same time for all firms. This led us to exclude ten other firms that could not be followed by InfoMinde3. Finally, 36 firms were excluded because of missing accounting or stock market data. The final sample consists of 108 firms.

InfoMinder is a service for searching and detecting changes and updates on designated web pages. It is mainly used to follow any information edited about firms products, events and news. For more information: www.infominder.com/webminderlIMTechSummary.jsp.

Samir Trabelsi, Rkal Labelle and Pascal Dumontier 127 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155
Table 1 Sampling Procedure Description Number of firms

Population of Toronto Stock Exchange firms in the StockGuide database Initial random sample Exclusion of firms in the financial sector Suspended firms Firms whose website updates could not be tracked by InfoMinder Firms whose accounting or stock market data were not available on the StockGuide or on the Datastream databases Final sample

14 94
180 16 10 10

36 22

3.2 Assessment of IFR quality

There is no direct measure of financial reporting quality. Beyond ratings by the AIMR that are no longer available, some prior studies measure reporting quality from independent evaluations based on the opinion of experts, such as financial analysts (Lang and Lundholm, 1993, 1996; Clarkson et al., 1999; Labelle, 2004). Assuming that the extent of financial reporting is a relevant indicator of its quality, because information quality is expected to be positively related to the quantity of available information, other studies use disclosure scores based on a content analysis of firms annual reports (Lang and Lundholm, 1993; Botosan, 1997; Healy and Palepu, 2001). Since there are no independent ratings gauging IFR quality and no recognised experts in this field, we developed our own disclosure index to evaluate the additional informationreported by Canadian firms on their websites. As we are interested in capturing the information superiorityof IFR over TFR,appraisingreporting quality adequately requires taking into consideration both the incremental content of IFR over TFR and the higher flexibility of IFR over TFR.FASB (2000, chapter 2) identifies two dimensions of corporate website disclosure: the content and the presentation format of the financial information (Figure 1). With regards to content, websites may simply reproduce TFR by presenting mandatory information in compliancewith standards,as well as voluntary disclosures already published in paper format or in SEDAR.These options are presented in quadrant I and quadrant I1 of Figure 1. Companies may also enhance TFR by publishing more disaggregated or more incrementalfinancial information on their websites (quadrants I11 and IV of Figure 1). As regards presentation, websites may include a static presentation of what is available on paper along with static graphs (quadrants I and 111). Alternatively,the firm can take advantage of IFR by opting for a dynamic presentation not possible on paper, notably by means of sound or video (quadrants I1 and IV). Given the specific features of IFR discussed above, we took inspiration from the scoring procedure used by Botosan (1997) because it takes into consideration both the quantity and the precision of the available information. As the relevance of the score depends on the choice of the items included in the disclosure index (Marston and Shrives, 1991), our selection is based on (a) items used in similar studies (e.g., Merton, 1987; Cooke,

I28

Samir Trabelsi, Rial Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics VoI4, No 2 (December 2008) 120-155

Figure 1 Incremental Dimensions of IFR vs TFR

Dynamic

I1 -1FR
Same as in quadrant I

j IV-IFR

$1

!
Dynamic presentation

Same as in quadrant 1 plus more detailed or incremental financial information

Dynamic presentation

Limited

Content

Important

1989; Marston and Shrives, 1991; Healy and Palepu, 1993,2001; Wallace et al., 1994; Botosan, 1997; FASB, 2001), and especially in Botosan (1997); (b) items recommended by the AICPA study on financial reporting (Jenkins Committee Report, 1994); and (c) items recommended in the FASB study on voluntary disclosure (FASB, 2001). This led us to classify the content of financial reporting on corporate websites and in traditional media into seven categories considereduseful by investors, financial analysts and standard setters: background information,summary of historical results, key non-financial statistics, management discussion and analysis, projected information outside of the MD&A, information on intangible assets and social and environmental information. Appendix A and Appendix B describe our disclosure index as well as the scoring procedure. For each firm i, the extent of incremental information disclosures on corporate websites (AIFR,) is measured by the difference between the scores attributed to corporate websites (SIFR,) and the scores attributed to documents available on SEDAR (STFR,):

(am,) - STFR, = SEXi


where SIFR, =
J=l

i IFScore,jand STFR,= i TFScorelJ


j=l

IF(TF)Scoreijdenotes the number of points awarded to the ithfirm for information of category j resulting from two separate content analyses, the first one conducted on the firms website (IFScore) and the second conducted on the documents posted on SEDAR

Samir Trabelsi, Rkal Labelle and Pascal Dumontier 129 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

(TFScore).These scores were estimated over the period September to December 2002. We needed a benchmark to determine whether and to what extent firms exercise the additional discretion provided by IFR over TFR. SEDAR was chosen as a benchmark for TFR for three reasons. First, SEDAR is the most exhaustive source of financial documents about public companies. Second, being essentially a reproduction of TFR, SEDAR is subject to the same regulations and is thus a good proxy for the degree of ritualism or opportunism deployed in TFR. Third, SEDAR is a source of information that is clearly defined and easily available for all listed companies. In order to ascertain the validity of our disclosure scores, following Clarkson et al. (1999) we randomly selected a subset of ten TFR disclosures and ten web disclosures, and asked two experts, the chair of the 2006 CICA corporate reporting award and a manager in the Ontario Securities Commissions corporate finance branch, to subjectively rate the extent of the internet incremental information (AIFX) on a scale of 1 to 10 for each of the seven information categories. Overall, the correlation between our disclosure scores and the practitionersratings comes to 0.85 and is statistically significant at the 1 percent level. To further validate our scores, we adopted the same procedure as Botosan (1997) and Richardson and Welker (2001). Using non-parametric correlation analyses, we examined the relation between the disclosure scores attributed to traditional media (STFR) and the firms size, performance and future financing activities, previous studies showing that the extent of financial information increases along with these three variables. As predicted,each of these variables is positively and significantly associated with the extent of STFR. Two other methods were used to estimate the reliability of both SIFR and STFR scores.The testretest method consists of analysing the content of the documents and websites under study at different time periods. A high correlation between the scores of two successive periods tends to validate their reliability. We obtained highly significant correlation coefficients mostly close to 0.98. The internal coherence method consists in estimating reliability coefficients, such as the alpha developed by Cronbach (1951). For both SIFR and STFR scores we obtained Cronbach coefficientshigher than 0.80, indicating that items included in our disclosure indices would have led independent experts to reach similar conclusions with regard to the quality of the sampled firms financial disclosures.
3.3 Determinants of incremental IFR disclosure

Based on the hypotheses developed in the previous section, Table 2 describes the variables used to analyse the reasons leading frsto use their websites to extend disclosures im and to broaden access to financial information. It also provides the expected relations between these variables and the extent of the information disclosed. In addition to these variables, we also control for variables - such as firm size, extent of analyst following, audit quality and industry - that prior research has found to be relevant to predicting the level of voluntary disclosure. Lang and Lundholm (1993) and Clarkson et al. (1999) show that firm size is positively associated with voluntary disclosure. Because of economies of scale, larger firms are more likely to post voluntary information on their websites than smaller firms. The extent of analyst following is also included in our model as Lang and Lundholm (1993) indicate that more forthcoming firms are followedby a larger number of analysts. Prior research having found that audit quality is positively associated

130

Samir Trabelsi, Real Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

Table 2 Hypotheses and Variables Used for the Analysis of the Determinants of IFR Variable (NAME) Description (Source or database) Exp. Sign

Firms investor base or clientele effect

HI Share turnover (ST)

Average monthly trading volume scaled by the average number of outstanding shares from January to December 2001 (Datastream, item VT) Percentage of shares held by executives and majority shareholders (StockGuide, item # 18) Dummy equals 1 if the firm is also listed on a foreign market, and 0 otherwise (Firms ID on SEDAR)

H, Ownership concentration
(OC)

H,

Cross listed (LISTING)

Nature o TFR f

H, H, H,

Relevance (RELEV) or informativeness Research & development (R&D) Degree of asymmetry (ASYM)

Book value scaled by market value (StockGuide, items 123 & 137) R&D expenditures (StockGuide, item 66)
Standard deviation of adjusted stock returns over the past ten years

+ +

Goodnews hypothesis

H, H,

Fms performance (PERF)


Financing activities (FIN)

Obtained by comparing the 2001 to the 2002 net income (NI). PERF = 1 if NI 2002 > NI 2001.0 otherwise (StockGuide, item 75) Dummy equals 1 if the firm issued shares or debt in 2001,2002 or 2003 and 0 otherwise (Financial Post new issues)

Effect o competition f

H,

Competition (COMPET)

Average returns on equity over the last five years (StockGuide, item 171) Index constructed using quarterly revenues (StockGuide, item 44)

HI, Herfindahl index (HERF)


Control variables Firm size (SIZE) Number of analysts following (NAF) Audit quality (BIG). Sector of activity (SIC)

Natural logarithm of total assets Average number of analysts following the firm in 2000 (International Brokers Estimate System tapes - IBES) Dummy equals 1, if firm audited by one of the Big 4,Ootherwise Dummy variables based on one-digit SIC codes

+ + +
?

Samir Trabelsi, RPal Labelle and Pascal Dumontier 131 Journal of Contemporary Accounting & Economics Vol4,No 2 (December 2008) 120-155

with voluntary disclosure (Xiao et a]., 2004), we take audit quality into consideration as proxied by whether the auditor is a Big 4 firm or not. Finally, we control for industry as there exists a high degree of financial reporting mimicking within industries. The decision to use the internet to provide additional disclosure is a two-stage decision. First, managers decide to use this medium to broaden the access to additional disclosure rather than simply to reproduce TFR in all or part. This broadening decision is followed by a second deliberate managerial decision with regard to the extent of additional information provided on the corporate website. Consequently, we cannot analyse the determinants of the information disclosed on a website in addition to that conveyed in TFR sources by excluding firms that do not provide additional disclosure. This could generate a serious selection bias resulting in inefficient and inconsistent regression estimates. To correct for this potential bias, we adopt the Heckman (1979) procedure. In a first step, we use a Logit model to examine the partial effects of independent variables on the managers decision to provide additional information on the firms website (Model A). Then, in a second step, we use an OLS model (Model B) to investigate the impact of the same independent variables on the extent of the additionalinformation disclosed on websites over that already available in TFR media ( AIFR).

Model A: IFRDecision,, = a, + a , ST, + a, OCi + a, RELEV, + a, R&D, + a, PERF, + as FIN, + a, LISTING, + a, COMPET, + a, HERF, + a,, ASYM, + a,, SIZE, + a,, NAF, + a,, BIG, + a,,,, SIC, + ci, (1)
where IFRDecision equals 1 if firmi provides incremental IFR (SIFR-STFR > 0); IFRDecision equals 0 otherwise.

where AIFR is a measure of the extent of additional IFR voluntary disclosure ( AIFR = SIFR - STFR); ST is the monthly average number of shares traded scaled by the average number o outstanding shares from January to December 2001; OC is the percentage of f shares held by executives and major shareholders; RELEV is the book value divided by market value; R&D is the natural logarithm of current spending on research and development;PERF equals 1 if NI 2002 > NI 2001 and 0 otherwise. PERF therefore results from the comparison of the net income earned in 2002 (NI 2002) and the one achieved in 2001 (NI 2001); FIN equals 1 if the firm has issued equity or debt securities in 2003, 2002, or 2001, and 0 otherwise; LISTING equals 1 if the firm is cross listed on a foreign exchange, 0 otherwise; COMPET is the firms average return on equity for the last five years; HERF denotes the Herfindahl index constructed using quarterly StockGuide data; ASYM is the standard deviation of adjusted stock market returns over the previous ten

132

Samir Trabelsi, Rkal Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-15.5

years; SIZE is the natural logarithm of the firms total assets; NAF is the average number of analysts following the firm during 2000; BIG equals 1 if the firm is audited by one of the Big 4 firms, otherwise 0; SIC is a dummy variable based on the firms one-digit SIC code. As there are six one-digit codes, we introduce five dummies in the model. The first step of the Heckam (1979) estimation procedure is a Logit regression. The dependent variable is qualitative, taking the value of 1 or 0 to indicate whether a firm uses its website to provide market participants with additional information. Results of the Logit model are not presented as they turned out to be mostly redundant with those of the OLS regression analysis. In the OLS model, the dependent variable is the extent of the additional information voluntarily disclosed on websites (AIFR). In our regressions, we winsorise all variables at the 1 percent and 99 percent levels to mitigate the influence of outliers. Finally, we use the heteroskedasticity-consistentstandard errors of White (1980) to compute p-values for statistical significance.

3.4 Consequences of incremental IFR disclosure


Relying on previous studies finding that analysts seek to forecast earnings accurately using firm-provided disclosures (Ashbaugh and Pincus, 2001; Byard et al., 2006), we use regressions to determine whether IFR incremental and disaggregated additional disclosures result in higher earnings forecast accuracy and in lower (or higher) forecast dispersion. We measure analysts forecast accuracy (ACCURACY) similar to Lang and Lundholm (1996). For each firm under study, using the IBES detailed forecasts file, we first calculate the 2002 fourth quarter mean earnings forecast by averaging the first forecasts issued by each analyst during the fourth quarter.We then estimate accuracy as the absolute difference between the mean earnings forecast (FORECAST) and the actual 2002 fourth quarter earnings (EPS) reported by IBES, using the beginning-of-quarter stock price (PRICE) as a deflator. Multiplying this absolute forecast error by (-1) provides a metric that increases with forecast accuracy so that a positive association with ACCURACY indicates that the variable under study improves the quality of forecasts. We therefore expect a positive association between ACCURACY and AIFR. ACCURACY is measured as follows: ACCURACY = (-1)

IFORECAST
PRICE

- EPS 1

Using the IBES detailed forecasts file, we measure DISPERSION similar to Lang and Lundholm (1996) as the inter-analyst standard deviation of the 2002 fourth quarter forecasts deflated by the beginning-of-quarter stock price. According to the above discussion, we expect a positive or a negative association between DISPERSION and AIFR depending on whether increased voluntary disclosure on corporate websites have a positive or a negative impact on the proportion of private information analysts incorporate into forecasts relative to common information. DISPERSION is caculated as follows: DISPERSION = Standard deviation of forecasts PRICE

Numerous studies (Lys and Soo, 1995; Gu and Wu, 2003) examining the determinants

Samir Trabelsi, Rlal Labelle and Pascal Dumontier 133 Journal of Contemporary Accounting & Economics VoE4,No 2 (December 2008) 120-155

of forecast accuracy and forecast dispersion show that accuracy declines and dispersion increases with the complexity of the forecasting task. Research shows that task complexity increases with the size of the firm and with the number of analysts following the firm, larger firms and firms followed by numerous analysts having richer information environments (Collins et al., 1987). Furthermore, Hwang et al. (1996) show that reporting net losses or negative changes in earnings also increases analysts task complexity and results in forecasts that are less accurate. In addition, Lys and So0 (1995) and Gu and Wu (2003) provide evidence suggesting that forecast accuracy is negatively related to forecast dispersion. Consequently, we control for firm characteristics affecting analysts forecast accuracy and dispersion using the natural logarithm of the firm market value (MV) and the number of analysts following the firm (NAF) as a proxy for the quality of the information environment of the firm. Furthermore, we use a dichotomous variable that is equal to 1 if the fourth quarter earnings is a loss (LOSS) and a dichotomous variable equal to 1 if the difference between the current years fourth quarter earnings and previous years fourth quarter earnings is negative (SURPRISE). Table 3 reports descriptive statistics on these variables. Firms under study are on average followed by four analysts.They exhibit a mean forecast error and a mean standard deviation of forecasts coming respectively to 2 percent and 1.1 of their share price. Each of these variables exhibits considerable variation across the sample, as shown by the ranges between their minimum and the maximum values.

Table 3 Descriptive Statistics Pertaining to Analyst Forecast Properties


N
~~

Mean
0.019 0.012 4.070

Minimum
0.056 0.000 1.000

Maximum Std. Deviation


0.365 0.157 19.000 0.257 0.133 4.120

ACCURACY DISPERSION NAF

45 32
45

ACCURACY negative absolute value of the analyst forecast error deflated by stock price; DISPERSION the interim analyst standard deviation of forecasts deflated by stock price; N A F number of analysts following the f r .

As firms not using the internet for financial reporting purposes cannot be ignored without generating a selection bias affecting regression estimates, the model is estimated on the conditional sample where AIFR > 0 to investigate the impact of IFR voluntary disclosure on analysts forecast accuracy (Model C) and forecast dispersion (Model D).

Model C: ACCURACY,i= y o + y , AIFR,i+y,MV,i+y,NAF,i+y,DISP,i+y,LOSS,i+ y6SURPRISE,,+E (3)

134

Samir Trabelsi, Rdal Labelle and Pascal Dumontier Journal of Contemporary Accounting C Economics Vol4,No 2 (December 2008) 120-155 ?

Model D:
DISPERSION,, = 6,

+ 6, AIFR,, + 6, MV,, + 6, NAF,, + 6, LOSS,,+


(4)

6, SURPRISE,, + E

where ACCURACY is the opposite of the absolute value of the analyst forecast error scaled by stock price; DISPERSION is the inter-analyst standard deviation of forecasts deflated by stock price; AIFR is the extent of additional voluntary disclosure on IFR over TFR (SIFR - STFR); MV is the natural logarithm of the firms market value; NAF is the number of analysts following the firm; LOSS equals 1 if the fourth quarter earnings is a loss; SURPRISE equals 1 if the difference between the current years fourth quarter earnings and the preceding years fourth quarter earnings is negative. Only 45 over the 57 firms using IFR to provide additional information have IBES coverage for the fourth quarter of 2002. Moreover,requiring at least three forecasts issued by three distinct analysts to compute the standard deviation of forecasts leads to the exclusion of 13 other firms. Consequently, the accuracy model is estimated for 32 firms only.

4. Empirical Results
4.1 The net contribution of IFR

Figure 2 and Table 4 display a comparison between SIFR and STFR disclosure scores for each category of information and for the overall information. Results in Table 4 clearly demonstrate that, on average, Canadian firms do publish additional financial information on their websites. The mean difference between IFR scores and TFR scores is positive for the overall information and for each category of information, except for MD&A, indicating that corporate websites are used to provide incremental information to market participants. Observing non-significant differences between IFR scores and TFR scores for MD&A is not unexpected as the MD&A is regulated and often boilerplated. As explained earlier, compared with TFR the information published on corporate websites may be disaggregated or incremental (Ashbaugh et al., 1999).Figure 3 compares disaggregated to incremental disclosures. It highlights the fact that, except for the summary of historical results (SHR) and MD&A categories, financial disclosure on Canadian firms websites is mostly incremental. Table 5 presents descriptive statistics for the 57 firms of our sample using their websites to release additional information,i.e., those with AIFR strictly positive. Table 6 contains the proportion of firms attaining the additional IFRdisclosure item. The background incremental information score (BI) varies between 0 and 10,with an average of 4.16 (standard deviation = 2.53). For the 57 firms using IFR, this information may, for instance, take the form of a discussion of the impact of entry barriers on current profit (16 percent). It may also consist of identifying the main markets of the firm and their specific characteristics (77 percent). Moreover, a few firms disclose information on the impact of their competitive environment on future profits (5 percent). Regarding key non-financial statistics (KNFS), the average score for this category is 4.49 suggesting that Canadian firms usually provide

Samir Trubelsi, R h l Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4,No 2 (December 2008)

135 120-155

Figure 2 Comparison between Average IFR Scores (SIFR) and Average TFR Scores (STFR)
I

60 50

40 30 20 10
Internet Financial Reporting (IFR) Traditional Financial Reporting (TFR)

5J9

I
BI: background information; SHR: summary of historical results; KNFS: key non-financial statistics; PINF projected information; MD&A: management discussion and analysis; IA: information on intangible assets; SEI: social and environmental information; SIFR: extent of voluntary disclosure on the internet; STFR: extent of voluntary disclosure on traditional financial reporting.

Table 4 Tests for the Absence of Mean Difference between SIFR and STFR Categories BI SHR KNFS PINF MD&A IA SEI Mean difference
2.508 0.246 2.153 4.576 0.034 2.780 2.288 14.585

T statistics
8.671 3.455 8.333 8.245 1.420 8.253 7.377 9.438

Sig. (2-tailed)

Z statistics
6.35 1 3.656 6.286 6.281 1.414 6.228 5.920 6.569

Sig. (2-tailed)
0.000 0.000 0.000 0 .ooo 0.157
0.000 0.000

0.om
0.001 0 .Ooo 0 .Ooo

AIFR

0.158 0.000 0 .Ooo 0 .Ooo

0. m o

Parametric and non-parametric tests compare disclosure scores attributed to internetjnancial reporting (SIFR) with those attributed to traditionalfinancial reporting (STFR).As it refers to SIFR minus STFR, a positive mean difference indicates that IFR scores are higher than TFR ones. Test are given bothfor each category of information and for the overall information (AIFR).

BI: background information; SHR: summary of historical results; KNFS: key non-financialstatistics;PINF: projected information; MD&A: management discussion and analysis; I A information on intangible assets; SEI social and environmental information; SIFR: extent of voluntary disclosure on the internet; STFR: extent of voluntary disclosure on traditional financial reporting; AIIFR:extent of additional voluntary disclosure on IFR over TFR.

136

Samir Trabelsi, Rial Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-I55

such statistics on their websites. For instance, some companies disclose changes in the number of units sold (11 percent) or in their market share (37 percent). Other firms disclose a list of their key employees. Incremental disclosure in Table 5 also includes projected information (PINF) in addition to that presented in the MD&A, these being regulated.This category has the highest average score of 9.51. Some firms provide a comparison between current and forecasted earnings (67 percent). Other firms disclose a comparison of current and forecasted annual sales (5 1 percent). A few firms discuss the implication of their sales-related opportunities and risks on their profitability (25 percent), on their future capital or R&D expenditures (60 pecent). Finally, other firms disclose forecasts about profits or sales. The information on intangible assets (IA) and social and environmental information (SEI) varies from a minimum score of 0 to a maximum score of 12, with average scores of 5.18 and 4.28 respectively. The first category refers essentially to information on the patents and licenses used by the firm (54 percent). Some firms disclose a description of their leading clients and suppliers (53 percent). With respect to SEI, companies mainly release statements of their social objectives as well as descriptions of their donations, grants or financial contributions (58 percent). Other firms include a description of anti-pollution activities linked to their operations and the actions undertaken to treat, manage or recycle products and waste (49 percent).

Figure 3 Disaggregated and Incremental Financial Disclosure on the Internet

30
25 20 15

Disaggregated Financial Disclosure Incremental Financial Disclosure

10
5

BI: background information; SHR: summary of historical results; KNFS: key non-financial statistics: PINF projected information; MD&A: management discussion and analysis; IA: information on intangible assets; SEI: social and environmental information; SIFR: extent of voluntary disclosure on the internet; EADI : extent of additional voluntary disclosure on IFR over TFR (EADI = AIFR).

Samir Trabelsi, Rial Labelle and Pascal Dumontier 137 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-IS5
Table 5 Descriptive Statistics Pertaining to Scores Attributed to Incremental and DisaggregatedDisclosures
~

_____

N Panel A: Disaggregated disclosures

Minimum

Maximum

Mean

Std. Deviation

BI SHR KNFS PINF MD&A IA SEI

Am I

57 57 57 57 57 57 57 57

1.04 0.21 0.00 0.07 0 .00 0.42 0.46 2.16

1.117 0.773 0.000 0.417 0.000 1.149 1.240 2.153

Panel B: Incremental disclosures


BI SHR KNFS PINF MD&A IA

SEI AIFR

57 57 57 57 57 57 57 57

10
4

10 25 2 12 12 48

4.16 0.30 4.49 9.51 0.04 5.18 4.28 27.98

2.534 0.654 2.487 5.613 0.265 3.112 3.347 10.621

BI: background information; SHR: summary of historical results; KNFS: key non-financial statistics; PINF: projected information; MD&A: management discussion and analysis; IA: information on intangible assets; SEI: social and environmental information; SIFR: extent of voluntary disclosure on the internet; STFR: extent of voluntary disclosure on traditional financial reporting; AIFR: extent of additional voluntary disclosure on IFR over TFR.

4.2 Determinants of IFR


This section describes the results of univariate and multivariate statistical analyses. Table 7 displays results of univariate tests on the mean differences between the characteristicsof firms disclosing additional information through their website (IFR = 1) and those providing the same level of information (NON-IFR firms = 0). A firm is identified as publishing additional information on the internet if the (IFR - TFR) difference is strictly positive. As predicted IFR firms exhibit a higher share trading volume or a higher turnover in their shareholders, and they are more widely held than NON-IFR ones. These differences are statistically significant at the 1percent level. Their traditional financial information appears less informative (RELEV and ASYM), and they spend more on R&D investments. These characteristics, which tend to translate to greater information asymmetry, could explain why these firms use IFR to be more forthcoming. Furthermore, reflecting the greater facility provided by IFR, 40 percent of IFR firms are cross listed (LISTING), as opposed to only 13 percent for NON-IFR firms. Also as predicted 54 percent of IFR firms, versus

138

Samir Trabelsi, Rial Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

Table 6 Percentage of Firms Attaining the Additional CFI Disclosure Item Disclosure Items
% of Firms Attaining the Item 93

I . Background Information
a. A statement of corporate goals or objectives is provided. b. A general statement of corporate strategy is provided. C . Actions taken during the year to achieve the corporate goal are discussed. d. Planned actions to be taken in future years are discussed. e. A time frame for achieving corporate goals is provided. f. Barriers to entry are discussed. g. Impact of barriers to entry on current profits is discussed. h. Impact of barriers to entry on future profits is discussed. 1. The competitive environment is discussed. J. The impact of competition on current profits is discussed. k. The impact of competition on future profits is discussed. 1. A general description of the business is provided. m The principal products produced are identified. n. Specific characteristics of these products are described. 0 . The principal markets are identified. P. Specific characteristics of these markets are described.

12 9 19 54 4 21 16 5 18 35 30 5 12 49

60
77
28

II. Summary of Historical Results


a. Return-on-assets or sufficient information to compute return-on-assets (i.e. net income, tax rate, interest expense and total assets) is provided. b. Net profit margin or sufficient information to compute net profit margin ( i s . net income, tax rate, interest expense and sales) is provided. c. Asset turnover or sufficient information to compute asset turnover ( i t . sales and total assets) is provided. d. Return-on-equity or sufficient information to compute return-on-equity (i.e. net income and stockholders equity) is provided. e. A summary of sales and net income for at least the most recent eight quarters is provided.
111. Key Non-Financial Statistics

9
9 9

0
25
89

a. b. c. d. e. f. g. h. i. j. k.

Number of employees. Average compensation per employee. Order backlog. Percentage of order backlog to be shipped next year. Percentage of sales in products designed in the last five years. Market share. Dollar amount of new orders placed this year. Units sold. Unit selling price. Growth in units sold. Rejectionldefect rates. 1. Production lead time. m. Age of key employees.

39 22 32 15 12 37
14

11 22 47 25 16 33

Samir Trabelsi, Rkal Labelle and Pascal Dumontier 139 Journal of Contemporary Accounting & Economics Vo14,No 2 (December 2008) 120-155
Table 6 (Cont.) Percentageof Firms Attaining the Additional CFI Disclosure Item
n. Sales growth in key regions not reported as geographic segments. 0. Break-even sales $'s. p. Volume of materials consumed. q. Prices of materials consumed. r. Ratio of inputs to outputs. s. Average age of key employees. t. Growth in sales of key products not reported as product segments.
20 8 12 7

4
2 13

IV. Projected Information


a. b. c. d. e. f. g. h. i.
A comparison of previous earnings projections to actual earnings is provided. A comparison of previous sales projections to actual sales is provided. The impact of opportunities available to the firm on future sales or profits is discussed. The impact of risks facing the firm on future sales or profits is discussed. A forecast of market share is provided. A cash flow projection is provided. A projection of capital expenditures or R&D is provided. A projection of future profits is provided. A projection of future sales is provided.

91 67 51 42 25 9 5 60 42 39
4

V. Management Discussion and Analysis a. b. c. d. e. f. Change in sales. Change in operating income. Change in cost of goods sold. Change in cost of goods sold as a percentage of sales. Change in gross profit. Change in gross profit as a percentage of sales. Change in selling and administrative expenses. Change in interest expense or interest income. Change in net income. Change in inventoly. Change in accounts receivable. Change in capital expenditures or R&D. Change in market share.

0 0 0
0

g. h. i. j.

k.
1.

m.

0 0 0 0 0 2 0 0 2
89

VI. Intangibles Assets

a. b. c. d. e. f. g. h. i.

Description of major customers. Discussion of the investment in customer service. Description of the size and productivity of the sales force. Description of major suppliers. Information on the employees engaged in R&D. Information on the R&D program. Information on R&D alliances. Discussion of the importance of the trademark. Discussion of the licences employed.

53 40 44
56 9 23 33

35 54

140

Samir Trabelsi, Rkal Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4,No 2 {December 2008) 120-155

Table 6 (Cont.) Percentage of Firms Attaining the Additional CFI Disclosure Item
VII. Social and Environmental Infurmatiun
81

a. b. c, d. e. f.

Statement of corporate social goals is provided. Description of responsibility activities is provided. Description of donations, grants, financial assistance. Statement of corporate environmental goals is provided. Description of specific activities to reduce pollution generated by the firm is provided. Description of actions taken for the transportation, and treatment of the waste products. g. Description of the activities for preservation and restoration of green spaces and natural sites.

63 68 58 44 44 49 54

23 percent of NON-IFR ones, have issued securities in one of the three years following the quarter under study (FIN). Finally, in conformity with the hypothesis that the level of competition may discourage managers from using IFR to disclose additional information, IFR firms exhibit a higher Herfindahl index which is significant at the 1 percent level. However, the difference is not significant with regard the other measure of competition used, i.e., the average ROE over the last five years. With respect to the control variables, the differences in firm size, analyst following and audit quality are all in the expected direction and significant at the usual statistical levels. Table 8 presents the results of the OLS regression3on the determinants of the extent of the incremental IFR information (AIFR). AIFR is left censored at zero to account for the fact that 5 1 firms in our sample do not use their website to communicate additional information. As expected, the extent of additional IFR voluntary disclosure (AIFR) is positively related to stock liquidity (ST) suggesting that firms with high trading volumes provide incremental information on their website to respond to investors demands for greater disclosure to avoid trading against better informed investors. This result confirms the results obtained by Bushee et al. (2003) in their examination of the determinants of conference calls. We also find a positive association between AIFR and the firms performance (PERF), in line with the good news hypothesis, which states that firms expecting good future performance are more forthcoming. AIFR is also positively associated with the level of R&D expenditure and with the degree of information asymmetry (ASYM) suggesting that firms are more forthcoming as they perceive net benefits from supplementing TFR disclosure with additional IFR disclosure. This AIFR may also avoid any potential expost

A Logit regression was first used to examine the likelihood that a firm would use its website for financial reporting purpose. Its results are not presented as most of them turned out to be redundant with the OLS analysis. Cross listing (LISTING) and the degree of informativeness (RELEV) as measured by the book to market ratio are the only variables positively and significantly related to the decision to use IFR but not to the decision to extend disclosure. This result suggests that (cross listed) firms use IFR mainly to reduce information asymmetry with (foreign) investors at a relatively modest cost but not to increase disclosure. This Logit model has statistically significant power (xz= 44.37, p-value = 0.0013). Of the 108 observations, the model correctly predicts 78.9 percent of the IFR firms, which indicates significant differences between those firms deciding to use IFR for additional voluntary disclosure and NON-IFR firms.

Samir Trabelsi, Rial Labelle and Pascal Dumontier 141 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-IS5
Table 7 Descriptive Statistics and Univariate Tests on the Difference between IFR and NON-IFR Firms Characteristics Decision
ST (in millions)

N
57 51 57 51 57 51 57 51 57 51 57 51 57 51 57 51 57 51 57 51 57 51 57 51 57 51

Average
37.900 19.158 25.581 38.330 0.659 1.198 618.399 62.730 0.050 0.100 0.540 0.230 0.400 0.130 -0.023 -0.066 0.101 0.020 0.051 0.011 2265.143 235.232 8.950 5.830 0.960 0.870

Minimum
(17.200 5.900 0.100 0.300 0.102 0.04 1 0.000 0 .Ooo

Maximum
165.289 758.146 79 .OOO 85.900 1.923 6.250 20426.067 1868.658 1.000 1.000 1.Ooo 1.om 1.Ooo 1.om 0.345 0.343 0.143 0.065 0.657 0.457 29310.706 5886.147 3 1.000 26.000 1.000 1.000

T-value sign. level


1.270 0.207 2.990 0.003** 3.067 0.003 ** 1.861 0.071 0.93 1 0.354 3.569 0.001** 3.502 0.001** 0.932 0.353 2.509 0.014 3.417 0.001** 2.781 0.006 1.902 0.062 1.885 0.062*

2-value sign. level


5.517 O.OOO** 3.016 0.003 * * 2.442 0 .O 15** 3.084 0.020 0.931 0.352 3.401 0.001** 3.345 0.001** 1.013 0.311 3.016 0.003 3.762 O.OOO** 5.528 0.Ooo 2.390 0.017** 1.864 0.062*

IFR

NON-IFR oc IFR NON-IFR RELEV IFR NON-IFR R&D ($ millions) IFR NON-IFR PERF IFR NON-IFR FrN IFR NON-IFR LISTING IFR NON-IFR COMPET IFR NON-IFR HERF IFR NON-IFR ASYM IFR NON-IFR SIZE ($ millions) IFR NON-IFR NAF IFR NON-IFR BIG IFR NON-IFR

0 .om 0.Ooo 0 .Ooo


0.om

0 .Ooo 0 .OOO
-1.083 - 1.083 0.013 0.002 0.014 0.004 2.957 1.302 1.000 1.000 0.000 0.000

AIFR: extent of additional voluntary disclosure on IFR over IFR; S T monthly average number of shares traded average number of outstanding shares from January to December 2001; OC: percentage of shares held by executives and major shareholders; RELEV book value/market value ratio; R&D: natural logarithm of current spending on research and development; PERF: 1 if net income (NI) 2002 > NI 2001 and 0 otherwise, firm performance is measured by a dichotomous variable obtained by comparing the 2002 NI and the 2001 NI; FIN: 1 if the firm has issued shares or debt in 2003,2002, or 2001, and 0 otherwise; LISTING: 1 if the firm is cross listed on a foreign exchange, 0 otherwise; COMPET the firms average return on equity for the last five years; HERF Herfindahl index constructed using quarterly StockGuide data; ASYM: standard deviation of adjusted stock market returns over the previous ten years; SIZE: natural logarithm of the firms total assets; N A F the average number of analysts following the firm during 2000; BIG: 1 if the firm is audited by one of the Big 4 firms, otherwise 0; SIC: a dummy variable based on the firms one-digit SIC code.

or ex unfe misinterpretation (Gibbins et al., 1990). However, contrary to expectations, the relevance variable (RELEV) and the degree of ownership concentration (OC) are not significantly related to AWR. Verrecchia (1983) argues that firms in more competitive industries are less forthcoming because the potential adverse effect of disclosure is more severe in these industries. In line

142

Samir Trabelsi, Real Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

Table 8 Results from an OLS Analysis of the Determinants of the Extent of Additional IFR Voluntary Disclosure (MFR) Dependent variable: AIFR
Constant ST

Sign

Estimation
-26.895** 12.876** -0.203 0.370 0.202** 1.804* 1.234* 0.248 -4.527** 3.689* 3.476** 3.426**
1.157**

oc
RELEV R&D PERF FIN LISTING COMPET HERF ASYM SIZE NAF BIG SIC Number of observations = 108 Censored obs = 5 1 Uncensored obs = 57 Wald chi2(20)= 53.25 Prob > chi2= 0.001 McFadden R-Squared = 0.354

-0.932 NS

**, * indicate statistical significance at the p = 0.01 and 0.05 levels respectively (two-tailed test)
AIFR:extent of additional voluntary disclosure on IFR over TFR; ST: monthly average number of shares traded average number of outstanding shares from January to December 2001; OC: percentage of shares held by executives and major shareholders; RELEV book valuelmarket value ratio; R&D: natural logarithm of current spending on research and development; PERF 1 if net income (NI) 2002 > NI 2001 and 0 otherwise, firm performance is measured by a dichotomous variable obtained by comparing the 2002 NI and the 2001 NI; FIN: 1 if the firm has issued shares or debt in 2003,2002,or 2001, and 0 otherwise; LISTING: 1 if the firm is cross listed on a foreign exchange, 0 otherwise; COMPET the firms average return on equity for the last five years; HERF: Herfindahl index constructed using quarterly StockGuide data; ASYM: standard deviation of adjusted stock market returns over the previous ten years; SIZE: natural logarithm of the firms total assets; N A F the average number of analysts following the firm during 2000; BIG: 1 if the firm is audited by one of the Big 4 firms, otherwise 0; SIC: a dummy variable based on the firms one-digit SIC code.

with this hypothesis, we find that AIFX is positively and significantly associated with the Herfindahl index aimed at capturing the level of competition in the firms industry.We also find that a higher level of competition (COMPET), as measured by the average last five years ROE, reduces the extent of additional voluntary disclosure. This finding is likely due to the fact that firms exhibiting a high profitability strategically choose to hide their sustainable performance (Hayes and Lundholm, 1996;Clarkson et al., 1999). Regarding control variables, size and analyst following are both positively associated with the extent of additional voluntary disclosure. These findings are consistent with prior studies on voluntary disclosure (Healy and Palepu, 2001). Surprisingly, and contrary to expectations, we find that audit quality is not associated with AIFR.This result may be ascribable to the fact that IFR is not audited.

Samir Trabelsi, RLal Labelle and Pascal Dumontier 143 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

4.3 The impact of IFR on analysts forecasts


Table 9 reports results from the OLS regression of forecast accuracy on the extent of additional IFR financial disclosure. As expected, the coefficient on the &FR variable is significantly positive. This result confirms that additional disclosure on corporate websites increases the accuracy of analyst forecasts by improving the quality of common information shared by all analysts. The results for control variables are largely consistent with prior studies. Forecast accuracy increases with the quality of the firms information environment as proxied by firm size and analyst following, but it decreases with forecast dispersion. As expected, losses and earnings decreases are associated with less accurate forecasts. When the extent of additional voluntary IFR over TFR (AIFR)is included in the model, the explanatory power of the model increases substantially, the adjusted R2 increasing from 0.102 to 0.128. This sizable increase reflects the usefulness for financial analysts of additional voluntary information posted on corporate websites. Comparing the R2sof the two models to determine whether the R2 of Model 2 is significantlyhigher that the one of Model 1 provides an incremental F-statistic associated with testing Model 2 against Model 1 significant at the 5 percent level.

Table 9 Regression of Forecast Accuracy on Disclosure Score and Control Variables Model 1
Intercept AIFR MV NAF DISPERSION LOSS SURPRISE Model Fit (F) Adj. R2 Model F Tests: Model 2 vs. Model 1, F = 3.47*

Model 2
2.129** 0.110** 1.009** 0.073 -0.3 14** -0.190* -0.03 1* 3.994** 0.128

2.176**

I .089** 0.113 - 0.310**

- 0.186*
-0.028*

2.199* 0.102

**, * indicate statistical significance at the p = 0.01 and 0.05 levels respectively (two-tailed test)
ACCURACY opposite of the absolute value of analyst forecast errors deflated by stock price; DISPERSION: inter-analyst standard deviation of forecasts deflated by stock price; AFR: extent of additional voluntary disclosure on IFR over TFR; M V natural logarithm of the firms market value; NAF: number of analysts following the fr; i m LOSS: 1 if the fourth quarter earnings is a loss; SURPRISE 1 if the difference between the current years fourth quarter earnings and the preceding years fourth quarter earnings is negative.

Table 10 provides OLS evidence on a negative relation between forecast dispersion and the extent of additional IFR disclosure, indicating that additional disclosure on corporate websites increases convergence of beliefs among analysts. When the extent of additional

144

Samir Trabelsi, Rdal Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

IFR disclosure score is included in the model, the explanatory power of the model increases substantially, the adjusted R2 increasing from 0.080 to 0.107. This increase suggests that more forthcoming firms enjoy lower dispersion among analysts forecasts. The incremental F-statistic associated with testing Model 2 against Model 1 is significant at the 5 percent level. Taken together, the increase of forecast accuracy and the decrease of forecast dispersion resulting from an increase of the information disseminated to market participants through the internet suggests that analysts private information results more from private communications with the management of firms than from contradictory interpretations of common information.

Table 10 Regression of Dispersion on Disclosure Score and Control Variables Independent Variables
Intercept AIFR MV NAF LOSS SURPRISE Model Fit (F) Adj . R2 Model F Tests: Model 2 vs. Model 1, F = 2.47*

Model 1
3.956**
-

Model 2
3.419** -0.132** - 0.023*
0.108

0.073 0.060 0.463 ** 0.240* 3.555* 0.080

0.456** 0.243* 2.825* 0.107

**, * indicate statistical significance at the p = 0.01 and 0.05 levels respectively (two-tailed test).
DISPERSION: inter-analyst standard deviation of forecasts deflated by stock price; AIFR: extent of additional voluntary disclosure on IFR over TFR; M V natural logarithm of the firmsmarket value; N A F number of analysts i m LOSS: 1 if the fourth quarter earnings is a loss; SURPRISE: 1 if the difference between the following the f r ; current years fourth quarter earnings and the preceding years fourth quarter earnings is negative.

5. Conclusion
The steady increase in the number of firms using their websites for financial reporting purpose suggests that IFR is becoming a standard practice for most firms. However, while some firms do not use IFR, others provide various levels of voluntary information on their websites in addition to that available in TFR sources as posted on SEDAR. This study was aimed at better appraising the incremental contribution of the internet to the firms financial reporting policy and at analyzing the determinants and consequencesof that effect. Corporate web disclosures are truly voluntary. So, this setting allows more powerful tests of the predictions of voluntary disclosure theories compared to studies examining,for instance, the level of disclosure in financial statements including MD&A, which contains many mandatory and/or boilerplate disclosures. We use the corporate website setting to extend voluntary disclosure research by examining whether the IFR decision depends on whether

Samir Trabelsi, Rkal Labelle and Pascal Dumontier 145 Journal of Contemporary Accounting & Economics Vol4,No 2 (December 2008) 120-155

the firm benefits from the internet democratization effect on business reporting. Indeed, IFR allows managers to reach a larger population of information users and to comment on the firms current financial position and future prospects in continuous time. To study the impact of the internet on the firms financial reporting strategy, we first document that Canadian firms provide incremental information on the internet. An enhanced version of Botosans (1997) disclosure index scoring sheet is used to measure the information in IFR as well as in TFR in order to determine the incremental effect of IFR over TFR. This is an important contribution as prior studies examined IFR without isolating its effect on financial reporting. The extent of this difference was measured for a random sample of 108 Canadian public firms. Of those, only 57 provided incrementaldisclosure on their websites mainly in the form of background information, key non-financial statistics, projected information in addition to that presented in MD&A, information on intangible assets, and social and environmental information. In order to mitigate some of the shortcomings of prior studies, we use the two-stage estimation procedure developed by Heckman (1979) to take into consideration that IFR incremental disclosure is the product of a two-stage voluntary disclosure decision. First, managers decide to use the internet to broaden the access to the firms financial information. Second, they set the extent of this information. Our results indicate that the nature of the firms investor base and financial reporting, the good news hypothesis and the level of competition influence the decision to maintain a website and use it to disclose voluntarily additional information. In line with the findings obtained by Bushee et al. (2003) with regard to conference calls, by Clarkson et al. (1999) with regard to MD&A, and by Lang and Lundholm (1993) with regard to investors relations, our results lead us to conclude that IFR is a part of the firms overall disclosure package and is likely managed in the same manner as TFR. We also extend the work of Bushee et al. (2003), who studied the use of open (like IFR) versus closed conference calls reserved for analysts, while assuming the extent of disclosure to be constant in both. Finally, as opposed to prior IFR studies which mostly focused on large firms, our study is based on a random sample of Canadian firms which enhances the external validity of our results. To assess the consequences of website disclosure, we investigate the relation between the magnitude of the additional disclosures on corporate websites and the accuracy and standard deviation of earnings forecasts. We document that more forthcoming firms on the internet enjoy higher forecast accuracy and lower dispersion among individual analysts forecasts. These results suggest that website disclosures provide useful information to analysts. Our results should be of interest to standard-setters and regulators in allowing them to compare their normative positions with the reporting practices analyzed in this study. our Insofar as the IASC task force has recommended establishinga code of conduct for IFR, F results may help identify the components of I R that should be more closely monitored. There are still several unanswered questionsleft for future research. We took a voluntary disclosure approach focusing on the motives of the preparers reporting decision as well as a user perspective approach as suggested by Penman (2003) and Schipper and Vincent (2003). We found that technological enhancements provided by IFR are useful in better understanding the economic prospects of firms. One could further examine this issue by applying the methodologies used by Bryant (1997) for MD&A and by Gu and Lev (2004)to

146

Samir Trabelsi, Rkal Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4,No 2 (December 2008) 120-155

study the effects of disclosing the outcome of patent licensing with regard to royalty income. Furthermore, as existing software can track the time and the date of internet disclosures, it is henceforth possible to study the stock market reaction to continuous financial reporting. More specifically,the methodology of Asthana and Balsam (2001) could be used to study the price and trading volume reaction following the disclosure of various categories of additional information identified in our disclosure index sheet.

Appendix A Disclosure Index Scoring Sheet


IName of the firm: Medium Analysed : Site web 0 SEDAR 0 Disclosure practice or source: Annual Report (RA) and others [Press release (PR) / presentation (PPT) /Conference1

Date of content analysis :

End of the accounting period

I.

Background Information

Qualitative

Quantitative

Source

p. Specific characteristics of these markets are described


II. Summary of Historical Results

Ten years or Less than ten more years

1
Source

/a. Return-on-assets or sufficient information to compute return-on-l assets (is., net income, tax rate, interest expense and total assets) is provided
b.

Net profit margin or sufficient information to compute net profit margin (is., net income, tax rate, interest expense and sales) is provided

c. Asset turnover or sufficient information to compute asset turnover (ie sales and total assets) is provided

Samir Trabelsi, Rkal Labelle and Pascal Dumontier 147 Journal o Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155 f

Appendix A (Cont.)
d. Return on equity or sufficient information to compute return equity (i.e., net income and stockholders equity) is provided
01

Yes e. A summary of sales and net income for at least the most recen eight quarters is provided
11 Key Non-FinancialS?a?isfics 11

No

Source

Number or amount

Source

a. c.

Number of employees Order backlog

b. Average compensation per employee d. Percentage of order backlog to be shipped next year e. Percentage of sales in uroducts designed in the last five years

f.

Market share

g. Dollar amount of new orders placed this year h. Units sold


i.

Unit selling price Growth in units sold

i.

k. Rejection/defect rates I. Production lead time


m. Age of kev emulovees

n. Sales growth in key regions not reported as geographic segments

o. Break even sales $'s p. Volume of materials consumed a . Prices of materials consumed
r. s. t. Ratio of inputs to outputs Average age of key employees Growth in sales of key products not reported as product segments

I
Source

IV. Projected Informarion a. A comparison of previous earnings projections to actual earnings b. A comparison of previous sales projections to actual sales ir

or profits is discussed
5 d. The impact of risks facing the firm on future sales or profits i discussed
~ ~

e. f.

A forecast of market share is provided A cash flow projection is provided

g. A projection of capital expenditures or R&D is provided

148

Sumir Trabelsi, Rkal Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4,No 2 (December 2008)

120-155

Appendix A (Cont.)

VII. Social and Environmental Informarion


Statement of corporate social goals is provided Description of responsibility activities is provided

Qualitative

Quantitative

Source

a. b.

c. Description of donations, grants, financial assistance d. Statement of corporate environmental goals is provided
e.

Description of specific activities to reduce pollution generated by im the fr is provided Description of actions taken for the transportation and treatment of waste products

f.

g. Description of the activities for preservation and restoration of green spaces and natural sites

I I

I I

Categories I to V are those of Botosan (1997). Categories VI to VII have been added to the Botoson disclosure index.

Samir Trabelsi, Rial Labelle and Pascal Dumontier 149 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

Appendix B Scoring Procedures Background Information


Corporate goals:

One point is awarded if the annual report includes a statement of the corporate goal or mission. One additional point is awarded if the goal is stated in quantitative terms, e.g., ROE of X %. Maximum points = 2.0
Statement of business strategy:

One point is awarded if the annual report includes a general statement of corporate strategy. One point is awarded if specific actions taken during the current year are outlined and one additional point is given if this includes quantitative infomation. One point is given if the annual report outlines specific actions to be taken in future years and one additional point is given if this includes quantitative information. One point is awarded if the annual report gives a time frame for attaining the corporate goal. Maximum points = 6.0
Competition.

One point is awarded if the annual report discusses barriers to entry. One point is awarded if the report discusses the impact of these on current firm profits and an additional point is given if the impact is quantified. One point is awarded if the annual report discusses the impact of barriers to entry on future firm profits and an additional point is given if this impact is quantified. One point is awarded if the report discusses the intensity of competition. One point is awarded if the impact of this on current firm profits is discussed and an additional point is given if this impact is quantified. One point is awarded if the report discusses the impact of competition on future firm profits and one additional point is given if this impact is quantified. Maximum points = 10.0
Description o the business: f

One point is awarded if the annual report gives a general description of the companys business activities. Maximum points = 1.O
Principal products:

One point is awarded if the annual report lists the principal products produced. One point is awarded if the characteristics of these products are described and an additional point is given if the description includes specific,quantified information, e.g., our machine can produce widgets five times faster than competitors similar machines. Maximum points = 3 .O.

150

Samir Trabelsi, Rial Labelle and Pascal Dumontier Journal o Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155 f

Appendix B (Cont.)
Principal markets:

One point is awarded if the report lists the principal markets that buy the firms products. One additional point is awarded if this is quantified eg percentage of sales to each market. One point is awarded if the annual report describes these markets and one additional point is given if this description is quantified, e.g., the size of the heavy duty truck market in 1989 was X units. Maximum points = 4.0

Summary of Historical Results


Anplual summary:

One point is awarded if sufficient information is given to compute return on assets for nine or fewer years (typically five or ten years is given). One additional point is awarded if sufficient information is given to compute return on assets for ten or more years. One point is awarded if sufficient information is given to compute the net profit margin for nine or fewer years and one additional point is given if sufficient information is available to compute the net profit margin for ten or more years. One point is awarded if sufficient information is given to compute asset turnover for nine or fewer years. One additional point is awarded if sufficient information is available to compute asset turnover for ten or more years. Since any two of ROA, profit margin or asset turnover is sufficient to compute the third, companies may only earn points for two of the three if all three are presented. One point is awarded if sufficient information is available to compute return on equity for nine or fewer years and one additional point is awarded if sufficient information is available to compute return on equity for ten or more years. Maximum points = 6.0
Quarterly summary:

One point is awarded if quarterly sales and net income is available for at least the most recent eight quarters. Maximum points = 1.O

Key Non-Financial Statistics


Two points are awarded for each of the 20 items listed. In some instances firms may provide one item more than once. For example, several market share figures may be present if the firm operates in more than one line of business. Only the first instance of disclosure is counted. Maximum points = 40.0

Samir Trabelsi,Rial Labelle and Pascal Dumontier 151 Journal of ContemporaryAccounting & Economics Vol4, No 2 {December2008) 120-155

Appendix B (Cont.) Projected Information


Comparison of last year 2 forecast to actual:

One point is awarded if the annual report includes a comparison of prior earnings projections to this years actual earnings. One additional point is given if this comparison is quantified. One point is awarded if the annual report includes a comparison of prior sales projections to this years actual sales. One additional point is given if this comparison is quantified. Maximum points = 4.0
Anticipated impact of opportunities and /or risks:

Two points are awarded if the report discusses the profit or sales implications of opportunities available to the firm. One additional point is given if these implications are quantified. Two points are awarded if the report discusses the profit or sales implications of risks facing the firm. One additional point is given if these implications are quantified. Maximum points = 6.0
Projected market share:

Two points are awarded if the annual report includes a forecast of future market share and one additional point is given if this is quantified. Maximum points = 3.0
Projected cash flow:

Two points are given if the firm provides a discussion of future cash flow requirements and how they will be met. One additional point is given if these requirements are quantified. Maximum points = 3 .O
Projected capital andlor research and development expenditures:

Two points are given if the annual report includes a discussion of future capital expenditure other than those already committed to. One additional point is given if the amount of planned capital expenditure is given. Two points are given if the annual report includes a discussion of future expenditures on research and development. One additional point is give if the amount of planned research and development expenditure is given. Maximum points = 6.0
Earnings andlor sales projections:

For a single segment firm, two points are given if the firm provides a profit projection and one additional point is given if the projection is a point estimate. Two points are given

152

Samir Trabelsi, Rkal Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

Appendix B (Cont.)
~~ ~ ~ ~

if the firm provides a sales projection and one additional point is given if the projection is a point estimate. For multi segment firms, the points awarded are weighted by the fraction of consolidated earnings or sales being forecast. One point multiplied by the weight as determined above is awarded for each segment forecast. One additional point is awarded if a consolidated forecast is made or if forecasts are provided for all of the firm segments and it is possible to determine the overall forecast from this information. If the forecasts are point estimates, one additional point times the appropriate weight is given. Using this procedure the maximum number of points available is 6.0 for both multi segment and single segment firms.

Management Discussion and Analysis


For a single segment firm, one point is awarded for each item (18 items total) discussed provided reasons for the change is given. One additional point per item is given if the explanation includes quantitative data, e.g., market share increased 5 percent over last year. Maximum points = 36.0 For multi segment firms, the procedure is the same except for sales and operating income. For these two items a firm is given only half of the points otherwise available if the report discusses the change in consolidated sales (or operating profit) and does not discuss the change on a segment by segment basis in a manner consistent with the breakdown provided in their segmented disclosure.

Intangible Assets
One point is awarded for each item (nine items) discussed. One additional point per item is given if the explanation includes quantitative data.

Social and Environmental Information


One point is awarded for each item (seven items) discussed. One additional point per item is given if the explanation includes quantitative data.

References
Amir, E. and B. Lev., 1996,Value-relevance of non-financial information: The wireless communications industry, Journal of Accounting and Economics 22( 1-3), 3-30. Ashbaugh, H., K. M. Johnstone, and T. Warfield, 1999,Corporate reporting on the internet, Accounting Horizons 13(3), 241-257. Ashbaugh, H. and M. Pincus, 2001, Domestic accounting standards, international accounting standards, and predictability of earnings, Journal of Accounting Research 39(3), 417-435.

Samir Trabelsi, Rial Labelle and Pascal Dumontier 153 Journal of Contemporary Accounting & Economics Vol4,No 2 (December 2008) 120-155

Asthana, S. and S. Balsam, 2001, The effect of EDGAR on the market reaction to 10-K filings, Journal of Accounting and Public Policy 20(4), 349-372. Banon, O., 0. Kim, S. Lim, and D. Stevens, 1998, Using analysts forecasts to measure properties of analysts information environment. The Accounting Review 73(4), 421-433. Bryant, S. H., 1997, Incremental information content of required disclosures contained in management discussion and analysis, The Accounting Review 72(2), 285-301. Botosan, C. A., 1997, Disclosure level and the cost of equity capital, The Accounting Review 72(3), 323-350. Botosan, C. A. and M. S. Harris, 2000, Motivations for change in disclosure frequency and its consequences:An examination of voluntary quarterly segment disclosure, The Accounting Review 38(2), 329-353. Bushee, B. J., D. A. Matsumoto, and G. S. Miller, 2003, Open versus closed conference calls: The determinants and effects of broadening access to disclosure, Journal of Accounting and Economics 34( l), 149-180. Byard, D. and K. W. Shaw, 2003, Corporate disclosure quality and properties of analysts information environment, Journal of Accounting, Auditing and Finance 18(3), 355-378. Byard, D., Y. Li, and J. Weintrop, 2006, Corporate governance and the quality of financial analysts information, Journal of Accounting and Public Policy 25(5), 609-625. Clarkson, P. M., J. L. Kao, and G. D. Richardson, 1999, Evidence that management discussion and analysis (MD&A) is a part of a firms overall disclosure package, Contemporary Accounting Research 16(l ) , 111-134. Collins,D.,S.P.Kothari,S. P.Raybum,and J.Dawson, 1987,Firmsize andthe information content of prices with respect to earnings, Journal of Accounting and Economics 9(2), 111-138. Cooke, T. E., 1989, Disclosure in the corporate annual reports of Swedish companies, Accounting and Business Research 19(74), 113-124. Cronbach,L. J., 1951,Coefficient alpha and the internal structure of tests, Psychometrika 16(3), 297-334 Debreceny, R., G. L. Gray, and A. Rahman, 2002, The determinants of internet financial reporting, Journal of Accounting and Public Policy 21(2), 371-394. Ettredge, M., V. J. Richardson, and S . Scholz, 2002, Dissemination of information for investors at corporate websites, Journal of Accounting and Public Policy 21( l ) , 357-369. Financial Accounting Standards Board (FASB),2000,Business Reporting Research Project: Electronic Distribution of Business Reporting Information (FASB,Nonvalk, CT). Financial Accounting Standards Board (FASB), 200 1, Improving Business Reporting: Insights Into Enhancing Voluntary Disclosure (FASB, Nonvalk, CT). Gelb, D., 2000, Corporate signaling with dividends, stock repurchases, and accounting disclosures: An empirical study, Journal of Accounting, Auditing & Finance 15(2), 99-120. Gibbins, M., A. Richardson, and J. Waterhouse, 1990, The management of corporate financial disclosures: Opportunism, ritualism, policies and processes, Journal o f Accounting Research 28(1), 121-143.

154

Samir Trabelsi, Rial Labelle and Pascal Dumontier Journal of Contemporary Accounting & Economics Val 4 , No 2 {December2008) 120-155

GouriCroux, C., 1989, EconomLtrie des Variables Qualitative, 2nd Cdition (Economica, Paris). Greene, W. H., 2003, Econometric Analysis, 5th edition (MacMillan, New York). Gu, F. and B. Lev, 2004, The information content of royalties, Accounting Horizons 18(1),1-12. Gu, Z . and J. Wu, 2003, Earnings skewness and analyst forecast bias, Journal of Accounting and Economics 35( l ), 5-29. Hayes, R. and R. Lundholm, 1996,Segment reporting to the capital market in the presence of a competitor, Journal of Accounting Research 34(2), 26 1-279. Heckman, J., 1979, Sample selection bias as a specification error, Econometrica 47( l), 153-161. Healy, P. and K. Palepu, 1993, The effect of firms financial disclosure strategies on stock prices, Accounting Horizons 7( l), 1-1 1. Healy, P., and K. Palepu, 2001, Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature, Journal of Accounting and Economics 3 1(1-3), 405-440. Hope, 0-K., 2003, Disclosure practices, enforcement of accounting standards, and analysts forecast accuracy: An international study, Journal of Accounting Research 41(2), 235-272. Huddart, S. and B. Ke, 2007, Information asymmetry and cross-sectional variation in insider trading, Contemporary Accounting Research 24( 1), 195-232. Hwang, L. S., C. L. Jan, and S. Basu, 1996, Loss firms and analystsearnings forecast errors, Journal of Financial Statement Analysis 1(2), 18-30. Jenkins Committe Report, 1994, Improving Business Reporting - A Customer Focus: A Comprehensive Report of the Special Committe on Financial Reporting (American Institute of Certified Public Accountants, New York). Labelle, R., 2004, The statement of corporate governance practices (SCGP): A voluntary disclosure and corporate governance perspective, Working paper, HEC, Montreal. Lang, M. H. and R. J. Lundholm, 1993, Cross-sectional determinants of analyst ratings of corporate disclosures, Journal of Accounting Research 3 1(2), 246-27 1. Lang, M. H. and R. J. Lundholm, 1996,Corporate disclosure policy and analyst behavior, The Accounting Review 7 1(4), 467-492. Lev, B. and S. Penman, 1990, Voluntary forecast disclosure, non-disclosure, and stock prices, Journal of Accounting Research 28( 1), 49-76. Lys, X . and L. Soo., 1995 Analysts forecast precision as a response to competition, Journal of Accounting, Auditing and Finance 10(4), 751-765. Maddala, G. S., 1991, A perspective on the use of limited-dependent and qualitative variables models in accounting research, The Accounting Review 66(4), 788-806. Marston, C. and P. Shrives, 1991, The use of disclosure indices in accounting research: A review article, Brirish Accounting Review 23(3), 195-210. Merton, R. C., 1987, A simple model of capital market equilibrium with incomplete information, Journal of Finance 42(3), 483-5 10. Myers, S.C., 1977,Determinants of corporate borrowing,Journal of Financial Economics 5(2), 147-175. Ontario Securities Commission (OSC), 2004, Stafs Notice 51-715: Corporate Finance

Samir Trabelsi, Rial Labelle and Pascal Dumontier 155 Journal of Contemporary Accounting & Economics Vol4, No 2 (December 2008) 120-155

Review Program Report - October 2004, available at http://www.osc.gov.on.ca/ Regulation/ RulernakinglCurrentJPart5/sn-20041022-5 1-7 15-cop-fin-review-report . jsp. Penman, S. H., 2003, The quality of financial statements: Perspectives from the recent stock market bubble, Accounting Horizons 17(supplement),77-96. Piotroski, J. and D. Roulstone, 2004, The influence of analysts, institutional investors, and insiders on the incorporation of market, industry, and firm-specific information into stock prices, The Accounting Review 79(4), 1119- 1 151. Pirchegger, B. and A. Wagenhofer, 1999, Financial information on the internet: A survey of the homepages of Austrian companies, European Accounting Review 8(2), 383395. Richardson,A. and M. Welker, 2001, Social disclosure, financial disclosure and the cost of equity capital, Accounting, Organizations and Society 26(7-8), 597-61 6. Shehata, M., 1991, Self-selection bias and the economic consequences of accounting regulation: An application of two-stage switching regression to SFAS No. 2, The Accounting Review 66(4), 768-787. Schipper, K. and L. Vincent, 2003, Earnings quality, Accounting Horizons 17(supplement), 97-110. Singhvi, S . and H. Desai, 1971,An empirical analysis of the quality of corporate financial disclosure, The Accounting Review 46( l), 129-138. Toronto Stock Exchange (TSE), 1999, Electronic Communications Disclosure Guidelines (TSE, Toronto). Trabelsi, S ., R. Labelle, and C. Laurin, 2004, The management of financial disclosure on corporate websites: A conceptual model, Canadian Accounting Perspectives 3(2), 235-259. Verrecchia R., 1983, Discretionary disclosure, Journal of Accounting and Economics 5(1), 179-194. Verrecchia, R., 200 1, Essays on disclosure. Journal of Accounting and Economics 32(1-3), 97-180. Xiao, J., Z. Yang, H. Chow, and W. Chee, 2004, The determinants and characteristics of voluntary internet-based disclosures by listed Chinese companies, Journal of Accounting and Public Policy 23(3), 191-225. Wallace, R. S. O., and K. Naser, 1995, Firm-specificdeterminants of comprehensiveness of mandatory disclosure in the corporate annual reports of firms on the stock exchange of Hong Kong, Journal ofAccounting and Public Policy 14(1),311-368. between comprehensiveness Wallace,R. S. O.,K.Naser,andA.Mora, 1994,Therelationship of corporate annual reports and firm characteristicsin Spain,Accounting and Business Research 25(97), 41-53. White, H., 1980, A heteroskedasticity-consistentcovariance matrix estimator and a direct test for heteroskedasticity,Econometrica 48(4), 8 17-838. Wang, J., 1993, A model of intertemporal asset prices under asymmetric information, Review of Economic Studies 60(2), 249-282.

Das könnte Ihnen auch gefallen