Audited financial reporting and voluntary disclosure: International evidence on management earnings forecasts
Abstract
In this paper, we extend prior research on the link between audited financial reporting and voluntary disclosure by examining international differences in the relationship between commitment to higher levels of audit verification of actual financial outcomes and management earnings forecasts (our proxy for voluntary disclosure), using firm-level data from 30 non-US countries. Our evidence that commitment to higher levels of audit verification (proxied by the choice of a Big 4 auditor, the amount of audit fees, and excess audit fees) is positively associated with the incidence and frequency of management forecasts, and with stock market reactions to such forecasts, supports the notion that audited financial reporting and voluntary disclosure of managers' private information are complements in countries around the world. We further find that the relation between audited financial reporting and management earnings forecasts is weaker for firms in countries with relatively stronger capital market development or with higher levels of investor protection, suggesting that audited financial reporting plays a more important complementary role in voluntary disclosure in countries with less-developed institutions. Overall, our findings suggest that firm-level commitment to better audited financial reporting and the strength of country-level institutional characteristics play substitute roles in corporate voluntary disclosure decisions.
1 INTRODUCTION
The research question we investigate in this study is whether commitment to higher levels of audit verification can serve as an effective mechanism in enhancing the perceived credibility of voluntary disclosure, which in turn affects a firm's voluntary disclosure decisions (Ball, Jayaraman, & Shivakumar, 2012) in countries outside the USA. More importantly, we also investigate the extent to which the influence of audited financial reporting on voluntary disclosure differs across economic settings.
There are several competing views on whether and how audited financial reporting can affect firms' future-oriented voluntary disclosure in an international setting. We argue that audited financial reporting matters more to a firm's voluntary disclosure in countries with stronger institutional characteristics. That is, the strength of country-level institutions complements the role that firm-level commitment to higher levels of audit verification plays in voluntary disclosure by enhancing the credibility-enhancing/signaling role of such commitment on voluntary disclosure in countries with better developed capital markets and higher levels of investor protection.1 To the extent that commitment to higher levels of audit verification signals the credibility of a firm's voluntary disclosure, especially given the more important role of auditing in countries with better-developed institutions, we predict a strengthened relationship between audited financial reporting and the perceived informativeness and decisions of a firm's voluntary disclosure.
However, one can also argue that the strength of country-level institutions can play a substitutive role in the relationship between audited financial reporting and voluntary disclosure, and thereby weakens the role that commitment to audited financial reporting plays in a firm's voluntary disclosure. Given the presence of a strong institutional infrastructure and legal institution that plays a positive role in enhancing/signaling the credibility of voluntary disclosure, any firm-level mechanisms, such as commitment to higher levels of audit verification, will play a relatively limited role and ultimately have little effect on a firm's voluntary disclosure decisions (Choi & Wong, 2007; Durnev & Kim, 2005).2 This alternative view thus predicts a weakened link between a firm-specific mechanism, in our case audited financial reporting, and the perceived informativeness and decisions of a firm's voluntary disclosure.
We examine these competing hypotheses and test the possible variations on the relationship between audited financial reporting (proxied by the choice of a Big 4 auditor, the amount of audit fees, and excess audit fees; i.e., the residual of the regression of total actual audit fees on major firm-level determinants identified by prior studies) and firms' voluntary disclosures across countries by focusing on a sample of management earnings forecasts (our proxy for voluntary disclosure) from 30 non-US countries covered by CapitalIQ COMPUSTAT during the period 2004–2010. Our evidence confirms the complementary role between audited financial reporting and voluntary disclosure in countries around the world. Specifically, we find a consistently positive relationship in our sample between all three proxies of commitment to higher levels of audit verification and the likelihood and frequency of management earnings forecasts and stock market reaction to those forecasts. More importantly, we find this positive relation decreases as the capital market development or strength of investor protection in a country increases. This finding suggests that firm-level commitment to higher levels of audit verification will have a stronger (weaker) effect on firms' voluntary disclosure practices when country-level institutions are weak (strong). These results are robust to various country-level measures of capital market development and level of investor protection, to either standalone or bundled earnings forecasts, to inclusion and exclusion of various countries in the sample (such as removing the three countries with the largest number of observations), to conducting a country-by-country regression, and to one-way firm-level or two-way firm-year-level clustering of standard errors.
This study provides new extensions to a few lines of research in the literature. First, by examining the heterogeneity in the relation between audited financial reporting and management earnings forecast practices in countries around the world, our study contributes to the literature on management earnings forecasts, which has primarily focused on US firms (Hirst, Koonce, & Venkataraman, 2008), and adds to the broad literature on cross-country determinants of corporate transparency (e.g., Bhattacharya, Daouk, & Welker, 2003; Chaney, Faccio, & Parsley, 2010; DeFond, Hung, & Trezevant, 2007; Lang et al., 2012). For example, one interesting result from our study is the positive relation between the number of earnings announcements in a year and the likelihood and frequency of management forecasts. Given the variations in the number of financial statements mandated by the disclosure requirements of each country and the subsequent earnings announcements issued by firms from different countries, we believe that our study has implications for global regulators and policy-makers when setting disclosure requirements.3
Second, studies have shown that commitment to higher levels of audit verification is positively related to firms' voluntary disclosure practices (e.g., Ball et al., 2012; Chen, Srinidhi, Tsang, & Yu, 2016). Although studies have shown the importance of firm-level credibility-enhancing mechanisms in affecting the perceived informativeness of voluntary disclosure, which in turn has an effect on managers' voluntary disclosure incentives, evidence from a single country such as the USA provides limited insight into whether such a finding can be generalized to other countries. By extending studies that analyze the relation between audited financial reporting and firms' voluntary disclosure (e.g., management earnings forecasts or nonfinancial corporate social responsibility reporting) from the USA to an international setting, our research confirms the important role that audited financial reporting plays in firms' voluntary disclosure in countries around the world.
In addition, while most research to date on corporate transparency in the international setting has either focused on mandatorily reported information or has studied general information disclosure without distinguishing between voluntary and mandatory disclosures, our study attempts to examine and emphasize the intertwined relation between mandatory and voluntary disclosures by investigating the relationship between audit verification of actual financial outcomes and a firm's management earnings forecasts (Ball et al., 2012; Beyer, Cohen, Lys, & Walther, 2010; Dutta & Gigler, 2002; Lennox & Park, 2006). Our findings suggest that the perceived credibility of voluntary disclosure and its subsequent effect on firms' voluntary disclosure decisions is jointly determined by both firm-level commitment to better audit verification and the level of country-level capital market development and/or investor protection in the country where a firm is located. As such, our study not only responds to calls for research on the effects of management earnings forecasts in international contexts (Bushman, Piotroski, & Smith, 2004; Francis & Wang, 2008; Hirst et al., 2008), it also tests the possible interaction between voluntary and mandatory disclosures, which is important for a better understanding of firms' disclosure decisions (Beyer et al., 2010).
Finally, our research also adds to the number of accounting and finance studies that have found that firm-level governance mechanisms and country-level institutional environments are substitutes (e.g., Choi & Wong, 2007; Lang et al., 2012; Lang, Lins, & Miller, 2003; Lang, Lins, & Miller, 2004). Important practical implications can be drawn from this finding. For example, in light of the importance of trust-building between firms and their stakeholders in general, and shareholders in particular, our finding suggests that when making decisions on the level of commitment to audit verification of actual financial outcomes for signaling purpose, managers should take into consideration the strength of institutional infrastructure in the country where the firm is located.
The remainder of this paper is organized as follows. In Section 2 we provide further discussion of the two competing views on the relation between commitment to higher levels of audit verification and firms' voluntary disclosures, and develop our hypotheses. Section 3 describes the sample, data, and empirical models. Our findings are presented in Section 4. Section 5 concludes the paper.
2 DEVELOPMENT OF HYPOTHESES
In most countries around the world, management earnings forecasts represent one of the key forms of voluntary disclosure through which firms communicate private information to market participants (Healy & Palepu, 2001), thereby affecting shareholders' decision-making (see Hirst et al. (2008) for a review of the management earnings forecasts literature). However, management earnings forecasts tend to be future orientated and subject to few reporting guidelines. Therefore, the information provided by management earnings forecasts is difficult for market participants to verify ex ante, which in turn increases market participants' concerns about the credibility of firms' voluntary disclosures. As a result, mechanisms by which managers can signal the truthful disclosure of private information are important to both managers and market participants (Hirst, Koonce, & Venkataraman, 2007).
Ball et al. (2012) argue that an important firm-level mechanism by which managers can signal truthful disclosure of private information, and thereby increase the perceived credibility of management earnings forecasts, is through commitment to higher levels of audited financial reporting. The rationale underlying this argument is that audited financial reporting provides confirmation of the voluntary earnings forecasts issued by firms ex post. Given that managers will face various costs (e.g., lowered reputation and credibility, and even litigation costs) if the earnings forecasts they issue voluntarily are not consistent with or confirmed by actual future performance outcomes, the resources firms commit to audited financial reporting, as measured by a higher level of audit fees, are predicted to be an increasing function of the extent of their management earnings forecasts (and vice versa). In other words, the decisions concerning voluntary disclosure and commitment to audited financial reporting are jointly determined. Consistent with their arguments, using management earnings forecasts issued by firms in the USA, Ball et al. (2012) find that the issuance of management earnings forecasts and stock market reaction associated with the forecasts increase with a firm's audit fees. In line with Ball et al. (2012), Chen et al. (2016) also find a positive association among US firms between audit fees and one type of voluntary nonfinancial disclosure; namely, corporate social responsibility disclosure.
Together, studies suggest that higher levels of audited financial reporting enhance the credibility of private information that is not directly verifiable by investors ex ante, and thus that audited financial reporting and corporate voluntary disclosure are complementary mechanisms in improving the overall information environment (Ball et al., 2012; McNichols & Trueman, 1994; Verrecchia, 1982). Complementing this line of research on the link between audited financial reporting and voluntary disclosure, the objective of this paper is to expand this link to an international setting and examine possible variation in the relationship between auditing of actual financial outcomes and corporate voluntary disclosure in countries around the world. As such, the first objective of this paper is to extend this finding from previous studies with a US focus to an international setting, and to reexamine the association between commitment to higher levels of audit verification and firms' voluntary financial disclosures, as measured by management earnings forecasts, for firms outside the USA. Our first hypothesis (H1) is thus formally stated as follows.
H1.Ceteris paribus, commitment to higher levels of audited financial reporting is positively associated with firms' voluntary disclosures.
An underlying assumption of the auditing literature is that independent audits of firms' financial reporting facilitate contracting by reducing information asymmetry and by providing governance and monitoring for the performance of the contracting parties (Watts & Zimmerman, 1986). However, a number of auditing studies investigating auditors' governance roles in cross-country settings (e.g., Choi & Wong, 2007; Francis & Wang, 2008) have suggested that the role of auditors and audited financial reporting in capital markets will vary with country-level institutional characteristics. In addition, other studies (e.g., Cao et al., 2017; Francis, Khurana, & Pereira, 2005) argue that the capital market consequence of voluntary disclosure, which potentially serves as an important factor affecting firms' voluntary disclosure decisions (Hirst et al., 2008), can also be affected by cross-country variations in institutional environments. Thus, by taking advantage of international differences in the strength of country-level institutions, our study examines the possible variation in the link between auditing fees and voluntary disclosure across countries.
Several mechanisms for this variation have been proposed in the literature. First, a number of studies have suggested that, in countries with stronger institutions, the role that auditing plays in reducing information asymmetry and/or signaling the credibility of firms' voluntary disclosure can be strengthened (Ball, 2001; Francis & Wang, 2008). In countries with stronger (weaker) country-level institutions, auditors are more (less) concerned about reputation and litigation costs, and thus they are perceived to play a relatively stronger (weaker) role in the quality of firms' financial reporting, and similarly in signaling the credibility of firms' voluntary disclosures. Following this discussion, we predict a strengthened role of audited financial reporting in the perceived informativeness of a firm's voluntary disclosure, and ultimately in a firm's voluntary disclosure decisions, in countries with better developed institutions.4
Second, findings from Durnev and Kim (2005) and Choi and Wong (2007) suggest that commitment to firm-level governance mechanisms, such as the engagement of a Big 4 auditor, can substitute for weak country-level institutions that negatively affect the perceived credibility of voluntary corporate disclosure. For example, research has documented greater financial transparency, higher value relevance of earnings, more informative earnings announcements and management earnings forecasts, and lower earnings management in countries with stronger institutions (Bhattacharya et al., 2003; Bushman et al., 2004; Cao et al., 2017; DeFond et al., 2007; Huang, Kong, & Tsang, 2013; Hung, 2001; Leuz, Nanda, & Wysocki, 2003). Following the findings of these studies, to the extent that better institutional development of a country facilitates communication between firms and market participants, in countries with better developed institutions one can predict a weakened role of audited financial reporting in a firm's management earnings forecasts.
Lastly, audit fees have been shown to be a function of the risk associated with an auditing engagement (Johnstone & Bedard, 2001; Venkataraman, Weber, & Willenborg, 2008). Consistent with this view, a recent study by Albring, Elder, and Xu (2016) argues that higher audit fees contain information on unobserved audit costs and client control risks, and thus finds that excess audit fees help predict a firm's future internal control weakness. More directly related to voluntary disclosure, Krishnan, Pevzner, and Sengupta (2012) explain their finding of a positive association between audit fees and voluntary disclosure by arguing that auditors consider management earnings forecast behavior to be associated with higher business risks, due to a greater concern for firms' earnings management behavior following firms' earnings forecasts (Kasznik, 1999). Following this discussion, to the extent that the strength of country-level institutions attenuates client firms' internal control weakness and/or earnings management risk, we again expect to find a less positive association between audit fees and voluntary disclosure in countries with better developed institutions.
In response to these competing views concerning the link between audited financial reporting and voluntary disclosure in countries with better developed institutions, we present our second hypothesis (H2) formally as follows:
H2.The strength of institutional development in a country has no effect on the relation between firms' commitment to higher levels of audited financial reporting and their voluntary disclosure.
3 DATA, SAMPLE, AND EMPIRICAL MODELS
3.1 Data and sample
We obtain global data for management earnings forecasts from CapitalIQ COMPUSTAT (CapitalIQ hereafter), which provides the original texts of management forecasts for firms across approximately 90 countries/regions beginning in 2004.5 According to CapitalIQ, the raw text forecasts are extracted from various sources, such as newspapers, filings, subscriptions, and announcements of transactions. CapitalIQ translates all forecasts published in languages other than English into English.6
We start with countries with at least one management earnings forecast, which would indicate that the country is followed by CapitalIQ. We define a firm's home country based on the country where a firm's stock is primarily listed, as presumably a firm's voluntary disclosure is more likely to be issued for investors in the primary listed stock market.7 We exclude countries/regions that lack the country-year-level institutional variables used in our study. We further exclude the USA from our analysis, because the purpose of our study is to extend prior studies showing a link between audited financial reporting and voluntary disclosure within a US setting to a global setting (e.g., Ball et al., 2012; Krishnan et al., 2012).8 We also exclude Japan, because management earnings forecasts in Japan are de facto mandatory (Kato, Skinner, & Kunimura, 2009). Our final sample consists of 30 countries during the sample period of 2004–2010, representing 16,319 unique firms issuing a total of 32,063 management earnings forecasts and 19,049 firm-years observations with earnings forecasts.
In addition, recent research shows that an increasing number of firms tend to issue management earnings forecasts together with their earnings announcements (e.g., Lansford, Lev, & Tucker, 2013; Merkley, Bamber, & Christensen, 2013; Rogers & Van Buskrik, 2013). As a result, to control for the possibility that management's earnings forecast decisions are affected by the issuance decisions of firms' earnings announcements, we create a separate sample of management earnings forecasts that includes only standalone, nonbundled earnings forecasts (i.e., a sample excluding all management earnings forecasts that were issued together with firms' earnings announcements) in examining the relation between commitment to higher levels of audited financial reporting and firms' management earnings forecast practices.
3.2 Empirical model

Following prior studies, we use three variables to measure a firm's commitment to higher levels of audit verification. Given research indicating that Big 4 auditors provide higher quality audits, and that the decision to hire a Big 4 auditor represents a firm's commitment to better financial reporting quality, both in the USA and globally (Choi & Wang, 2007; Francis, Maydew, & Sparks, 1999; Francis & Wang, 2008; Guedhami, Pittman, & Saffar, 2014), our first measure of audit commitment is the choice of a Big 4 auditor (Big4Auditor) made by a firm in a given year.9 In addition, following previous studies (e.g., Ball et al., 2012; Chen et al., 2016), we use the amount of total actual audit fees (AuditFee) and excess audit fees (ExcessFee) to measure a firm's commitment to higher levels of audit verification.
Excess audit fees represent audit fees that are incremental to those associated with previously identified determinants of audit fees. In particular, it constitutes the residual from the regression of audit fees (AuditFee) on firm-level audit fees' determinants, as summarized by Ball et al. (2012). These determinants are (1) a variable controlling for firm size, the log of total assets (LnAssets); (2) variables controlling for audit complexity, including total accruals scaled by total assets (Accruals), the ratio of current assets to total assets (Current), the ratio of foreign sales to total sales (Foreign), and the number of business segments (Segment); (3) variables controlling for audit risk, including return on assets (ROA), the ratio of total liabilities to total assets (Leverage), an indicator variable for firm-years with negative income (Loss), and the lag between the fiscal period end and the earnings announcement date (Lag); and (4) a variable controlling for seasonal peaks in audit fees, which is an indicator variable for firms with December fiscal year-ends (DEC).10
Similar to all the control variables, the three proxies to measure a firm's commitment to higher levels of audit quality (i.e., Big4Auditor, AuditFee and ExcessFee) are all lagged 1 year to capture the ex-ante level of audit commitment.11 As such, in Equation 1, β1 is our coefficient of interest. Specifically, it estimates the relation between a commitment to higher levels of audit verification and management's earnings forecast decisions.
We use two measures for voluntary disclosure. Specifically, voluntary disclosure is measured as either the incidence of management forecasts (MF Occu), an indicator variable that takes the value of one if a firm issues an earnings forecast in a given year, and zero otherwise, or the frequency of management forecasts (MF Freq), a count variable indicating the total number of earnings forecasts issued by a firm in a given year. We set earnings forecast frequency to zero if a firm issues no forecasts in a given year. Owing to the difference in the definition of each forecast variable, models with MF Occu (MF Freq) as the dependent variable are estimated using logistic (Poisson) regressions.12
In addition to examining the link between a firm's commitment to higher levels of audit verification and its voluntary disclosure decisions, as measured by the likelihood and frequency of management earnings forecasts, we directly examine whether stock market reactions to management earnings forecasts vary with a firm's audit commitment. Stock market reaction associated with management earnings forecasts (AbsCAR) is defined as the absolute value of the 2-day cumulative market-adjusted return during the [0,1] earnings forecast window, with day 0 equal to the forecast date.13 Examining the stock market reaction to earnings forecasts issued by firms with higher audit fees is important, because one can argue that higher audit fees could result from a firm's incentive to become more financially important to its auditors (e.g., intentional overpayment), and thereby a higher level of audit fees might actually reduce the perceived audit quality due to impaired auditor independence. Similarly, if audit fees are considered as a function of the risk associated with an auditing engagement as discussed in hypothesis development, we also predict a lowered stock market reaction to earnings forecasts issued by firms with higher audit fees. Alternatively, if investors regard higher audit fees or the appointment of a Big 4 auditor as a signal of commitment to better information quality, then we predict that investors will react more strongly to earnings forecasts issued by these firms.
In examining stock market reactions to earnings forecasts, following prior work (Hirst et al., 2008), we further obtain some other variables measuring the properties of management earnings forecasts: (1) forecast precision (MF Prec), which measures how precise or quantitatively specific the forecast is;14 (2) forecast timeliness (MF Time), defined as the number of days between a forecast's release and the earnings realization date (i.e., annual report filing date); and (3) forecast error (MF Error), defined as the absolute difference between the forecast and the actual performance of the item forecasted, divided by the actual performance (in percentage form).15
Furthermore, we draw on the literature to identify and control for a wide range of firm- and industry-level control variables that may influence firms' earnings forecast decisions: the logarithm of total assets (LnAssets), to control for firm size; return on assets (ROA), to control for financial performance; the level of a firm's asset-scaled total accruals (Accruals), to control for financial opacity; the number of business segments (Segment), to control for forecast complexity; whether a firm reports a loss (Loss); the number of analysts following a firm (Analysts), to control for the overall information environment; the percentage of shareholding of institutional investors (Institution) and insiders (Insider); the percentage of a firm's sales growth from year t to year t+1 (SalesGrowth); the number of exchanges on which a stock is listed (StkExch); whether a firm cross-lists its stock on stock exchanges in the USA (ADR); the sales-based Herfindahl index multiplied by −1 to control for industry-level competition (Competition); and whether a firm is in the high-tech industry (HiTech). In addition to all these control variables, we add the total number of earnings announcements issued by a firm in a given year (NumEA), to control for the possibility that a higher incidence (or frequency) of earnings forecasts is more likely to be observed when a firm issues multiple earnings announcements in a specific year, given the high tendency of firms to bundle the issuance of earnings forecasts and earnings announcements. We provide detailed definitions of these variables in Appendix A. We estimate Equation 1 at the firm-year level, and in all regressions we include country, industry, and year indicators and cluster all standard errors by both firm and year.16


In Equation 2a, CAPMKT is a time-variant measure for the level of capital market development in a country. It is defined as the total stock market capitalization of listed companies as a percentage of GDP for each country-year, as obtained from the World Bank (Beck, Demirgüç-Kunt, & Levi, 2003; Doidge, Karolyi, & Stulz, 2007). INVPRO in Equation 2b is a time-variant measure for the strength of investor protection in a country. It is obtained from ‘Doing Business Indicators’ by the International Finance Corporation and the World Bank (Doidge, 2004; Francis & Wang, 2008).17 In both equations, the coefficient estimate of β2 is our variable of interest. A significant and positive (negative) estimated coefficient will thus support a strengthened (weakened) role of commitment to audited financial reporting in firms' voluntary disclosure decisions in countries with stronger institutions.
4 EMPIRICAL RESULTS
4.1 Univariate results
Table 1, panel A, presents the distribution of the sample, the mean of all management earnings forecast variables, and variables proxying commitment to higher levels of audit verification (Big4Auditor, AuditFee, and ExcessFee) by country. According to the table, Finland, Denmark, Ireland, Austria, and the Netherlands have the highest percentage of firm-years issuing management earnings forecasts. The variations in the percentage of firm-year observations with Big 4 auditors across countries are also notable. The substantial variations observed in both audit fees and excess audit fees support the view that audit fees can be partially affected by reasons other than audit effort, such as the desire to signal the degree of management's commitment to higher financial reporting quality in different countries. Some other observations are also worth noting. For example, 14 out of the 30 sample countries present negative excess fees on average during our sample period. However, countries with negative excess fees tend to be countries with relatively well-developed country-level institutions. Thus, this finding provides preliminary support to the prediction that audit fees can play a relatively limited role in firms' voluntary disclosures when a country's institutions are strong.
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | ||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
N | N (MF Occu = 1) | MF Occu | MF Occu SA | MF Freq | MF Freq SA | AbsCAR | AbsCAR SA | Big4Auditor | AuditFee | ExcessFee | ||
Panel A: By country | ||||||||||||
1 | Australia | 10,442 | 2,283 | 0.22 | 0.18 | 0.44 | 0.30 | 4.73 | 4.78 | 0.44 | 3.65 | −0.30 |
2 | Austria | 523 | 248 | 0.47 | 0.19 | 1.02 | 0.24 | 2.38 | 2.80 | 0.77 | 5.29 | −0.45 |
3 | Belgium | 925 | 246 | 0.27 | 0.11 | 0.54 | 0.16 | 3.44 | 3.61 | 0.56 | 5.13 | −0.32 |
4 | Brazil | 2,168 | 130 | 0.06 | 0.04 | 0.08 | 0.04 | 3.12 | 3.12 | 0.63 | 5.11 | −0.07 |
5 | Canada | 16,977 | 1,565 | 0.09 | 0.05 | 0.20 | 0.06 | 4.20 | 4.55 | 0.51 | 4.55 | 0.09 |
6 | Denmark | 1,147 | 628 | 0.55 | 0.24 | 1.54 | 0.33 | 3.19 | 3.67 | 0.80 | 5.84 | 0.61 |
7 | Finland | 936 | 639 | 0.68 | 0.34 | 1.73 | 0.47 | 3.42 | 3.82 | 0.89 | 5.40 | 0.05 |
8 | France | 4,811 | 1,292 | 0.27 | 0.16 | 0.59 | 0.26 | 3.01 | 2.48 | 0.55 | 5.58 | 0.07 |
9 | Germany | 6,193 | 2,203 | 0.36 | 0.17 | 0.91 | 0.27 | 3.03 | 2.69 | 0.49 | 5.12 | −0.16 |
10 | Greece | 1,276 | 125 | 0.10 | 0.06 | 0.13 | 0.07 | 2.93 | 2.96 | 0.29 | 4.54 | −0.60 |
11 | Hong Kong | 9,134 | 1,092 | 0.12 | 0.11 | 0.16 | 0.13 | 4.70 | 4.77 | 0.63 | 5.30 | 0.40 |
12 | Indonesia | 1,875 | 314 | 0.17 | 0.13 | 0.26 | 0.19 | 2.97 | 3.03 | 0.19 | 6.09 | 1.43 |
13 | Ireland | 335 | 160 | 0.48 | 0.42 | 1.09 | 0.73 | 4.15 | 4.00 | 0.86 | 5.70 | −0.06 |
14 | Israel | 1,686 | 64 | 0.04 | 0.02 | 0.07 | 0.03 | 2.97 | 3.33 | 0.19 | 5.20 | 0.26 |
15 | Italy | 1,719 | 486 | 0.28 | 0.15 | 0.49 | 0.20 | 2.37 | 2.04 | 0.44 | 5.48 | −0.86 |
16 | Malaysia | 6,438 | 820 | 0.13 | 0.06 | 0.17 | 0.07 | 2.58 | 2.27 | 0.57 | 4.03 | −0.54 |
17 | Mexico | 695 | 65 | 0.09 | 0.07 | 0.16 | 0.09 | 2.88 | 2.55 | 0.61 | 6.16 | 0.16 |
18 | Netherlands | 922 | 371 | 0.40 | 0.23 | 0.86 | 0.37 | 3.76 | 3.15 | 0.86 | 6.01 | 0.12 |
19 | New Zealand | 787 | 289 | 0.37 | 0.28 | 0.71 | 0.40 | 2.71 | 2.55 | 0.79 | 4.76 | −0.20 |
20 | Norway | 1,601 | 212 | 0.13 | 0.07 | 0.21 | 0.08 | 5.50 | 6.45 | 0.84 | 5.98 | 0.25 |
21 | Philippines | 1,258 | 207 | 0.16 | 0.12 | 0.30 | 0.17 | 2.57 | 2.44 | 0.29 | 5.31 | 1.22 |
22 | Poland | 2,583 | 297 | 0.11 | 0.07 | 0.21 | 0.11 | 3.23 | 3.41 | 0.22 | 4.54 | −0.08 |
23 | Singapore | 4,716 | 487 | 0.10 | 0.06 | 0.14 | 0.07 | 4.03 | 4.02 | 0.68 | 4.11 | −0.73 |
24 | South Africa | 1,790 | 343 | 0.19 | 0.15 | 0.27 | 0.20 | 3.23 | 3.34 | 0.63 | 6.14 | 0.74 |
25 | South Korea | 9,352 | 305 | 0.03 | 0.02 | 0.05 | 0.03 | 2.89 | 2.74 | 0.35 | 5.38 | 0.13 |
26 | Sweden | 2,941 | 336 | 0.11 | 0.05 | 0.20 | 0.07 | 4.02 | 4.15 | 0.78 | 5.84 | 0.43 |
27 | Switzerland | 1,926 | 606 | 0.31 | 0.13 | 0.57 | 0.18 | 3.41 | 3.57 | 0.91 | 5.63 | −0.06 |
28 | Taiwan | 10,058 | 284 | 0.03 | 0.02 | 0.06 | 0.04 | 2.36 | 2.34 | 0.65 | 5.09 | 0.30 |
29 | Thailand | 3,126 | 743 | 0.24 | 0.17 | 0.44 | 0.26 | 2.05 | 1.87 | 0.55 | 5.28 | 0.91 |
30 | UK | 11,574 | 2,209 | 0.19 | 0.15 | 0.30 | 0.21 | 5.76 | 5.69 | 0.57 | 4.77 | −0.29 |
Overall | 119,914 | 19,049 | 0.16 | 0.10 | 0.31 | 0.15 | 3.77 | 3.88 | 0.54 | 4.79 | −0.06 | |
Panel B: By year | ||||||||||||
1 | 2004 | 13,281 | 1,711 | 0.13 | 0.10 | 0.25 | 0.15 | 3.06 | 3.13 | 0.50 | 2.41 | −2.23 |
2 | 2005 | 15,182 | 1,838 | 0.12 | 0.09 | 0.22 | 0.13 | 3.40 | 3.45 | 0.55 | 4.02 | −0.65 |
3 | 2006 | 16,673 | 2,289 | 0.14 | 0.08 | 0.26 | 0.12 | 3.40 | 3.41 | 0.56 | 4.09 | −0.72 |
4 | 2007 | 17,781 | 2,519 | 0.14 | 0.08 | 0.27 | 0.11 | 3.59 | 3.67 | 0.55 | 5.39 | 0.48 |
5 | 2008 | 18,830 | 2,935 | 0.16 | 0.10 | 0.31 | 0.14 | 4.51 | 4.89 | 0.56 | 5.44 | 0.52 |
6 | 2009 | 19,459 | 3,290 | 0.17 | 0.11 | 0.33 | 0.16 | 4.51 | 4.59 | 0.55 | 5.42 | 0.53 |
7 | 2010 | 18,708 | 4,467 | 0.24 | 0.14 | 0.50 | 0.21 | 3.45 | 3.57 | 0.53 | 5.49 | 0.51 |
Overall | 119,914 | 19,049 | 0.16 | 0.10 | 0.31 | 0.15 | 3.77 | 3.88 | 0.54 | 4.79 | −0.06 |
Table 1 also presents the distribution of sample and summary statistics by year in panel B, demonstrating a positive trend in management earnings forecasts (i.e., incidence and frequency of management earnings forecasts and market reactions to those forecasts) and audit fees/excess audit fees over time. This evidence provides preliminary support for a positive correlation between commitment to higher levels of audit verification and firms' voluntary disclosure decisions.
We report the descriptive statistics for our test variables, variables related to management earnings forecast properties, and other control variables in Table 2. In our sample, approximately 16% (10%) of firm-year observations included issued (standalone) earnings forecasts during our sample period. The average earnings forecast frequency is 0.31, suggesting that firms that issue earnings forecasts on average do so about twice per year. In our sample, more than half of the observations were audited by Big 4 auditing firms during the sample period. Our sample firms had average total assets of USD 112 million, were followed by 2.4 analysts, and had an average institutional ownership of 14.5%, suggesting that our sample firms included both large and small firms from around the world.
N | Mean | SD | 1% | Median | 99% | |
---|---|---|---|---|---|---|
MF Occu | 119,914 | 0.16 | 0.37 | 0.00 | 0.00 | 1.00 |
MF Occu SA | 119,914 | 0.10 | 0.30 | 0.00 | 0.00 | 1.00 |
MF Freq | 119,914 | 0.31 | 0.86 | 0.00 | 0.00 | 4.00 |
MF Freq SA | 119,914 | 0.15 | 0.50 | 0.00 | 0.00 | 3.00 |
AbsCAR | 19,049 | 3.77 | 4.60 | 0.00 | 2.26 | 25.65 |
AbsCAR SA | 12,081 | 3.87 | 4.75 | 0.00 | 2.28 | 26.00 |
Big4Auditor | 119,914 | 0.54 | 0.50 | 0.00 | 1.00 | 1.00 |
AuditFee | 42,886 | 4.79 | 2.34 | 0.01 | 5.21 | 8.48 |
ExcessFee | 42,886 | −0.06 | 1.88 | −4.73 | 0.34 | 3.37 |
NumEA | 119,914 | 1.86 | 1.75 | 0.00 | 2.00 | 5.00 |
LnAssets | 119,914 | 4.72 | 2.31 | 0.19 | 4.55 | 10.98 |
ROA | 119,914 | −1.86 | 19.32 | −30.60 | 1.96 | 20.00 |
Accruals | 119,914 | 0.03 | 0.76 | −0.21 | 0.02 | 0.42 |
Segment | 119,914 | 1.23 | 0.59 | 1.00 | 1.00 | 3.00 |
Loss | 119,914 | 0.36 | 0.48 | 0.00 | 0.00 | 1.00 |
Analyst | 119,914 | 2.40 | 5.86 | 0.00 | 0.00 | 29.00 |
Institution | 119,914 | 14.51 | 20.85 | 0.00 | 4.99 | 89.00 |
Insider | 119,914 | 3.60 | 10.11 | 0.00 | 0.49 | 57.00 |
SalesGrowth | 119,914 | 16.50 | 40.48 | −26.46 | 5.88 | 69.60 |
StockExch | 119,914 | 1.44 | 0.90 | 1.00 | 1.00 | 5.00 |
ADR | 119,914 | 0.09 | 0.28 | 0.00 | 0.00 | 1.00 |
Competition | 119,914 | −0.29 | 0.23 | −1.00 | −0.23 | −0.04 |
HiTech | 119,914 | 0.15 | 0.36 | 0.00 | 0.00 | 1.00 |
Table 3 compares the difference in both incidence and frequency of management earnings forecasts by country across firms with a relatively high and low commitment to higher levels of audit verification. Specifically, it tabulates the mean incidence of earnings forecasts (columns 1–3), mean incidence of standalone earnings forecasts (columns 4–6), mean frequency of earnings forecasts (columns 7–9) and mean frequency of standalone earnings forecasts (columns 10–12) across firms with different levels of audit commitment for each of our sample countries. The results from these comparisons strongly support the conclusion that a higher level of audit commitment is associated with a higher likelihood and frequency of earnings forecasts, as indicated by the significantly positive difference in forecast likelihood and frequency across the two samples for most of the sample countries. The results are robust using either choice of auditor or amount of audit fees to measure a firm's audit commitment. For example, the results from Table 3, columns 1–3 (4–6), show that in 28 (24) out of the 30 sample countries, there is a significantly greater likelihood of issuing earnings forecasts (standalone earnings forecasts) for firms with higher audit commitment than for those with relatively low audit commitment. Similar results can also be observed from the frequency of earnings forecasts, and our result indicates that commitment to higher levels of audit verification is associated with greater forecast frequency in 28 (25) out of the 30 sample countries. Thus, the univariate comparison results provide preliminary evidence that the confirmation role of commitment to audited financial reporting on firms' voluntary disclosures that was observed in the US setting can be indeed generalized to firms outside of the USA.
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Country | MF Occu SA | MF Occu SA | MF Freq SA | MF Freq SA | |||||||||
Big4Auditor = 1 | Big4Auditor = 0 | Diff | AuditFee = High | AuditFee = Low | Diff | Big4Auditor = 1 | Big4Auditor = 0 | Diff | AuditFee = High | AuditFee = Low | Diff | ||
1 | Australia | 0.28 | 0.10 | 0.18*** | 0.45 | 0.14 | 0.31*** | 0.49 | 0.15 | 0.34*** | 0.82 | 0.21 | 0.61*** |
2 | Austria | 0.22 | 0.11 | 0.11*** | 0.21 | 0.10 | 0.11* | 0.26 | 0.16 | 0.10* | 0.23 | 0.10 | 0.13** |
3 | Belgium | 0.15 | 0.06 | 0.09*** | 0.21 | 0.07 | 0.14*** | 0.23 | 0.06 | 0.17*** | 0.31 | 0.09 | 0.23*** |
4 | Brazil | 0.05 | 0.01 | 0.04*** | 0.03 | 0.01 | 0.02 | 0.06 | 0.01 | 0.05*** | 0.04 | 0.02 | 0.02 |
5 | Canada | 0.08 | 0.01 | 0.07*** | 0.14 | 0.03 | 0.11*** | 0.11 | 0.02 | 0.09*** | 0.19 | 0.04 | 0.15*** |
6 | Denmark | 0.28 | 0.06 | 0.22*** | 0.37 | 0.16 | 0.20*** | 0.40 | 0.07 | 0.33*** | 0.52 | 0.23 | 0.29*** |
7 | Finland | 0.36 | 0.21 | 0.15*** | 0.42 | 0.33 | 0.09** | 0.48 | 0.31 | 0.17** | 0.58 | 0.42 | 0.16** |
8 | France | 0.22 | 0.08 | 0.14*** | 0.28 | 0.10 | 0.18*** | 0.37 | 0.12 | 0.26*** | 0.50 | 0.16 | 0.34*** |
9 | Germany | 0.26 | 0.08 | 0.18*** | 0.34 | 0.10 | 0.24*** | 0.44 | 0.11 | 0.33*** | 0.59 | 0.14 | 0.45*** |
10 | Greece | 0.10 | 0.04 | 0.06*** | 0.12 | 0.03 | 0.09*** | 0.12 | 0.05 | 0.07*** | 0.12 | 0.04 | 0.08** |
11 | Hong Kong | 0.11 | 0.10 | 0.01** | 0.12 | 0.08 | 0.04*** | 0.14 | 0.12 | 0.02** | 0.14 | 0.10 | 0.04*** |
12 | Indonesia | 0.26 | 0.10 | 0.16*** | 0.12 | 0.12 | 0.00 | 0.41 | 0.14 | 0.27*** | 0.19 | 0.17 | 0.02 |
13 | Ireland | 0.44 | 0.30 | 0.14* | 0.58 | 0.15 | 0.42*** | 0.78 | 0.43 | 0.35*** | 1.03 | 0.19 | 0.84*** |
14 | Israel | 0.07 | 0.01 | 0.06*** | 0.08 | 0.01 | 0.07*** | 0.10 | 0.02 | 0.08*** | 0.13 | 0.01 | 0.12*** |
15 | Italy | 0.19 | 0.11 | 0.08*** | 0.24 | 0.14 | 0.10* | 0.28 | 0.14 | 0.14*** | 0.38 | 0.15 | 0.23** |
16 | Malaysia | 0.07 | 0.05 | 0.02*** | 0.09 | 0.05 | 0.04*** | 0.08 | 0.06 | 0.02*** | 0.11 | 0.05 | 0.06*** |
17 | Mexico | 0.06 | 0.07 | −0.01 | 0.16 | 0.00 | 0.16*** | 0.09 | 0.10 | −0.01 | 0.23 | 0.00 | 0.23** |
18 | Netherlands | 0.26 | 0.05 | 0.21*** | 0.29 | 0.22 | 0.06 | 0.42 | 0.08 | 0.345*** | 0.49 | 0.31 | 0.18** |
19 | New Zealand | 0.31 | 0.17 | 0.14*** | 0.41 | 0.24 | 0.16*** | 0.45 | 0.22 | 0.23*** | 0.63 | 0.33 | 0.30*** |
20 | Norway | 0.08 | 0.04 | 0.04** | 0.13 | 0.06 | 0.07** | 0.09 | 0.04 | 0.05** | 0.15 | 0.07 | 0.08** |
21 | Philippines | 0.16 | 0.10 | 0.06*** | 0.14 | 0.06 | 0.08** | 0.23 | 0.14 | 0.09** | 0.24 | 0.07 | 0.17*** |
22 | Poland | 0.15 | 0.05 | 0.10*** | 0.06 | 0.08 | −0.02 | 0.27 | 0.07 | 0.20*** | 0.09 | 0.13 | −0.04 |
23 | Singapore | 0.06 | 0.05 | 0.01 | 0.07 | 0.04 | 0.03*** | 0.07 | 0.06 | 0.01 | 0.09 | 0.04 | 0.05*** |
24 | South Africa | 0.19 | 0.08 | 0.11*** | 0.23 | 0.13 | 0.09*** | 0.27 | 0.08 | 0.19*** | 0.33 | 0.16 | 0.17*** |
25 | South Korea | 0.05 | 0.01 | 0.04*** | 0.03 | 0.05 | −0.02 | 0.07 | 0.01 | 0.06*** | 0.05 | 0.05 | 0.00 |
26 | Sweden | 0.06 | 0.01 | 0.05*** | 0.12 | 0.03 | 0.09*** | 0.09 | 0.01 | 0.08*** | 0.19 | 0.04 | 0.15*** |
27 | Switzerland | 0.14 | 0.05 | 0.09*** | 0.18 | 0.10 | 0.07*** | 0.19 | 0.10 | 0.09** | 0.24 | 0.14 | 0.10*** |
28 | Taiwan | 0.03 | 0.01 | 0.02*** | 0.05 | 0.01 | 0.04*** | 0.05 | 0.02 | 0.03*** | 0.08 | 0.01 | 0.07*** |
29 | Thailand | 0.21 | 0.12 | 0.09*** | 0.14 | 0.14 | 0.00 | 0.33 | 0.18 | 0.15*** | 0.22 | 0.21 | 0.01 |
30 | UK | 0.20 | 0.08 | 0.12*** | 0.28 | 0.10 | 0.18*** | 0.28 | 0.10 | 0.18*** | 0.40 | 0.12 | 0.28*** |
Overall | 0.14 | 0.06 | 0.08*** | 0.22 | 0.09 | 0.13*** | 0.21 | 0.08 | 0.13*** | 0.34 | 0.12 | 0.23*** |
Table 4 presents the Pearson correlation matrix for the main variables used in our models. Results show that all three variables proxying for a firm's commitment to better audit verification used in our study are positively and highly correlated with each other. More importantly, all these variables have a significant and positive association with both the likelihood and frequency of firms' earnings forecasts. In an untabulated test, we also find significant and positive association between all audit commitment variables and stock market reactions to management earnings forecasts.
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | MF Occu | ||||||||||||||||||||
2 | MF Occu SA | .77 | |||||||||||||||||||
3 | MF Freq | .83 | .72 | ||||||||||||||||||
4 | MF Freq SA | .67 | .87 | .77 | |||||||||||||||||
5 | Big4Auditor | .17 | .14 | .17 | .13 | ||||||||||||||||
6 | AuditFee | .24 | .20 | .25 | .19 | .29 | |||||||||||||||
7 | ExcessFee | .06 | .03 | .05 | .02 | .03 | .81 | ||||||||||||||
8 | NumEA | .38 | .28 | .38 | .27 | .23 | .24 | .03 | |||||||||||||
9 | LnAssets | .32 | .28 | .33 | .28 | .40 | .60 | .01 | .34 | ||||||||||||
10 | ROA | .13 | .10 | .11 | .09 | .16 | .24 | .01 | .04 | .42 | |||||||||||
11 | Accruals | −.01 | −.01 | −.01 | −.01 | −.02 | −.01 | .01 | .02 | −.04 | −.02 | ||||||||||
12 | Segment | .27 | .25 | .32 | .27 | .16 | .26 | .00 | .32 | .36 | .02 | .02 | |||||||||
13 | Loss | −.14 | −.10 | −.13 | −.10 | −.18 | −.26 | −.02 | −.06 | −.44 | −.49 | .04 | −.03 | ||||||||
14 | Analyst | .40 | .35 | .46 | .38 | .24 | .35 | .01 | .39 | .54 | .15 | −.01 | .52 | −.19 | |||||||
15 | Institution | .23 | .21 | .22 | .19 | .24 | .28 | .05 | .26 | .37 | .12 | −.01 | .26 | −.13 | .30 | ||||||
16 | Insider | −.02 | −.02 | −.03 | −.03 | −.07 | −.06 | .00 | −.01 | −.08 | .05 | −.01 | −.06 | −.05 | −.06 | −.08 | |||||
17 | SalesGrowth | .01 | .00 | .00 | .00 | .01 | −.02 | −.04 | .02 | .05 | .10 | −.01 | −.01 | −.14 | .00 | .02 | .01 | ||||
18 | StockExch | .34 | .30 | .39 | .33 | .21 | .29 | .00 | .36 | .42 | .05 | .01 | .80 | −.07 | .59 | .32 | −.06 | −.02 | |||
19 | ADR | .26 | .27 | .30 | .29 | .20 | .30 | .01 | .26 | .43 | .10 | −.01 | .42 | −.14 | .52 | .29 | −.06 | .01 | .57 | ||
20 | Competition | −.11 | −.08 | −.11 | −.08 | −.06 | −.09 | .00 | −.06 | −.13 | −.05 | .03 | −.08 | .03 | −.13 | −.08 | .05 | .00 | −.10 | −.12 | |
21 | HiTech | .00 | .00 | .02 | .00 | .02 | −.08 | .00 | −.06 | −.12 | −.04 | −.02 | −.03 | .03 | −.02 | −.06 | .02 | .01 | −.03 | −.03 | .18 |
- This table describes the Pearson correlation coefficients for the main variables used in our model. Significant correlations are indicated in bold (p < .10, two-tailed test). All variables are defined in Appendix A.
4.2 Multivariate regression results: H1
Before formally testing our hypotheses, we start by examining the link between commitment to audited financial reporting and management earnings forecasts for an exclusively US sample. Although this test does not directly answer the question of whether and to what extent the findings from the USA on the importance of audited financial reporting to voluntary disclosure can be generalized to non-US settings, this validation process provides an assurance to subsequent tests in our attempts to extend the US setting to an international setting. The results tabulated in Appendix B suggest that our model can successfully replicate the results documented by Ball et al. (2012). Thus, our results using three different measures of audit commitment confirm the existence of a positive link between commitment to audited financial reporting and voluntary disclosure in the USA.
Table 5 reports the regression estimates of Equation 1 to formally test H1 using an international setting. Our results consistently show that the likelihood of firms' issuing earnings forecasts (MF Occu) or standalone earnings forecasts (MF Occu SA) is positively and significantly related to all our proxies of commitment to higher levels of audit verification for firms from around the world. In terms of economic significance, the results shown in column 1 of Table 5 suggest that the appointment of a Big 4 auditor increases the likelihood of firms' issuing an earnings forecast by 12.5%. Similarly, column 2 of Table 5 suggests that a commitment to audit fees that is one standard deviation higher increases the likelihood of firms issuing an earnings forecast by more than 8%.18
1 | 2 | 3 | 4 | 5 | 6 | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
All forecasts | Standalone forecasts | |||||||||||
Dep. variable | MF Occu | MF Occu | MF Occu | MF Occu SA | MF Occu SA | MF Occu SA | ||||||
Model | Logistic | Logistic | Logistic | Logistic | Logistic | Logistic | ||||||
Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | |
Big4Auditor | 0.118*** | 0.02 | 0.078*** | 0.03 | ||||||||
AuditFee | 0.033*** | 0.01 | 0.035*** | 0.01 | ||||||||
ExcessFee | 0.032*** | 0.01 | 0.034*** | 0.01 | ||||||||
NumEA | 0.538*** | 0.01 | 0.475*** | 0.01 | 0.475*** | 0.01 | 0.381*** | 0.01 | 0.282*** | 0.01 | 0.282*** | 0.01 |
LnAssets | 0.194*** | 0.01 | 0.172*** | 0.01 | 0.192*** | 0.01 | 0.234*** | 0.01 | 0.207*** | 0.01 | 0.228*** | 0.01 |
ROA | 0.021*** | 0.00 | 0.025*** | 0.00 | 0.025*** | 0.00 | 0.017*** | 0.00 | 0.020*** | 0.00 | 0.020*** | 0.00 |
Accruals | −0.004 | 0.02 | 0.005 | 0.03 | 0.005 | 0.03 | −0.020 | 0.03 | −0.025 | 0.04 | −0.025 | 0.04 |
Segment | 0.011 | 0.02 | 0.006 | 0.03 | 0.009 | 0.03 | 0.083*** | 0.02 | 0.059** | 0.03 | 0.062** | 0.03 |
Loss | −0.108*** | 0.03 | −0.087** | 0.04 | −0.085** | 0.04 | 0.037 | 0.03 | 0.099** | 0.05 | 0.101** | 0.05 |
Analyst | 0.036*** | 0.00 | 0.033*** | 0.00 | 0.033*** | 0.00 | 0.034*** | 0.00 | 0.033*** | 0.00 | 0.033*** | 0.00 |
Institution | 0.006*** | 0.00 | 0.005*** | 0.00 | 0.005*** | 0.00 | 0.006*** | 0.00 | 0.005*** | 0.00 | 0.005*** | 0.00 |
Insider | 0.008*** | 0.00 | 0.006*** | 0.00 | 0.006*** | 0.00 | 0.006*** | 0.00 | 0.004** | 0.00 | 0.004** | 0.00 |
SalesGrowth | 0.001*** | 0.00 | 0.001** | 0.00 | 0.001** | 0.00 | 0.001 | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 |
StockExch | 0.061*** | 0.02 | 0.051** | 0.02 | 0.050** | 0.02 | 0.006 | 0.02 | 0.017 | 0.02 | 0.017 | 0.02 |
ADR | −0.175*** | 0.04 | 0.021 | 0.05 | 0.021 | 0.05 | −0.087** | 0.04 | 0.104** | 0.05 | 0.104** | 0.05 |
Competition | 0.047*** | 0.05 | 0.030 | 0.07 | 0.030 | 0.07 | 0.107* | 0.06 | 0.106 | 0.08 | 0.106 | 0.08 |
HiTech | 0.266** | 0.06 | 0.080 | 0.08 | 0.080 | 0.08 | 0.242*** | 0.07 | 0.017 | 0.09 | 0.017 | 0.09 |
Country indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
Industry indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
Year indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
N | 119,914 | 42,886 | 42,886 | 119,914 | 42,886 | 42,886 | ||||||
N (dep. var. = 1) | 19,049 | 10,290 | 10,290 | 12,081 | 6,569 | 6,569 | ||||||
Pseudo R2 (%) | 44.50 | 43.15 | 43.13 | 34.39 | 32.26 | 32.25 |
- *** ,
- ** , and
- * indicate that the estimated coefficients are statistically significant at the 1%, 5%, and 10% levels respectively in two-tailed t-tests based on robust standard errors.
- This table reports the regression estimates of the relation between a firm's commitments to audited financial reporting (as measured by Big4Auditor, AuditFee, and ExcessFee) and management forecast likelihood (MF Occu). All firm-level continuous variables are winsorized at the 1st and the 99th percentiles. All regressions include country, industry, and year fixed effects. Refer to Appendix A for more detailed variable definitions.
The estimated effect of several other control variables on MF Occu is also worth noting. Issuing more earnings announcements in a year is positively associated with the likelihood of issuing forecasts, suggesting the importance of controlling for this variable in studies examining management earnings forecasts, especially in an international setting, given the variations in the number of financial statements and earnings announcements provided by firms from different countries. This finding is also consistent with studies that have found that earnings forecasts voluntarily issued by firms increasingly tend to be bundled with earnings announcements. Firms with higher assets (LnAssets), better financial performance (ROA), greater following by financial analysts (Analyst), and higher intuitional (Institution) and insider (Insider) ownership appear to have a higher tendency to issue earnings forecasts voluntarily. The finding that analyst following (Analyst) and institutional ownership (Institution) are positive and significant suggests that firms with better management disclosure policies attract greater analyst following and institutional investment. Alternatively, one can argue that a higher level of voluntary disclosure is driven by strong information demands from both financial analysts and institutional investors.19
Table 6 reports the regression estimates of Equation 1 using management earnings forecast frequency as the dependent variable instead of forecast likelihood. Following prior studies, the model is estimated using a Poisson regression model, given the count nature of the dependent variable (Balakrishnan et al., 2014; Bamber et al., 2010; Lennox & Park, 2006). Generally, results are consistent with the findings from Table 5 and suggest that a firm's commitment to better audit verification is positively associated with a firm's voluntary disclosure, as measured by the frequency of issuing earnings forecasts.
1 | 2 | 3 | 4 | 5 | 6 | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
All forecasts | Standalone forecasts | |||||||||||
Dep. variable | MF Freq | MF Freq | MF Freq | MF Freq SA | MF Freq SA | MF Freq SA | ||||||
Model | Poisson | Poisson | Poisson | Poisson | Poisson | Poisson | ||||||
Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | |
Big4Auditor | 0.022** | 0.01 | 0.001 | 0.00 | ||||||||
AuditFee | 0.009** | 0.00 | 0.007*** | 0.00 | ||||||||
ExcessFee | 0.009** | 0.00 | 0.008*** | 0.00 | ||||||||
NumEA | 0.109*** | 0.01 | 0.149*** | 0.02 | 0.149*** | 0.02 | 0.033*** | 0.00 | 0.034*** | 0.00 | 0.034*** | 0.00 |
LnAssets | 0.023*** | 0.00 | 0.027*** | 0.01 | 0.032*** | 0.01 | 0.020*** | 0.00 | 0.025*** | 0.00 | 0.030*** | 0.00 |
ROA | 0.001*** | 0.00 | 0.002*** | 0.00 | 0.002*** | 0.00 | 0.001* | 0.00 | 0.001** | 0.00 | 0.001** | 0.00 |
Accruals | −0.005*** | 0.00 | −0.006** | 0.00 | −0.006** | 0.00 | −0.001 | 0.00 | 0.001 | 0.00 | 0.001 | 0.00 |
Segment | 0.003 | 0.02 | 0.039** | 0.01 | 0.040** | 0.01 | 0.027* | 0.01 | 0.044*** | 0.01 | 0.045*** | 0.01 |
Loss | −0.043*** | 0.01 | −0.045* | 0.02 | −0.045* | 0.02 | −0.010* | 0.00 | −0.011 | 0.01 | −0.011 | 0.01 |
Analyst | 0.033*** | 0.00 | 0.030*** | 0.00 | 0.030*** | 0.00 | 0.018*** | 0.00 | 0.016*** | 0.00 | 0.016*** | 0.00 |
Institution | 0.001 | 0.00 | 0.001* | 0.00 | 0.001* | 0.00 | 0.001 | 0.00 | 0.001 | 0.00 | 0.001 | 0.00 |
Insider | 0.001*** | 0.00 | 0.001*** | 0.00 | 0.001*** | 0.00 | 0.001*** | 0.00 | 0.001* | 0.00 | 0.001* | 0.00 |
SalesGrowth | −0.001 | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 | −0.001** | 0.00 | −0.001*** | 0.00 | −0.001*** | 0.00 |
StockExch | 0.089*** | 0.01 | 0.057*** | 0.00 | 0.057*** | 0.00 | 0.033*** | 0.01 | 0.023** | 0.01 | 0.023** | 0.01 |
ADR | 0.042 | 0.02 | 0.088*** | 0.02 | 0.088*** | 0.02 | 0.088*** | 0.02 | 0.116*** | 0.02 | 0.116*** | 0.02 |
Competition | 0.013 | 0.02 | 0.046 | 0.03 | 0.046 | 0.03 | 0.001 | 0.02 | 0.024 | 0.02 | 0.025 | 0.02 |
HiTech | 0.090*** | 0.00 | 0.110*** | 0.02 | 0.110*** | 0.02 | 0.034*** | 0.01 | 0.023 | 0.02 | 0.023 | 0.02 |
Country indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
Industry indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
Year indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
N | 119,914 | 42,886 | 42,886 | 119,914 | 42,886 | 42,886 | ||||||
Adj. R2 (%) | 35.66 | 39.40 | 39.40 | 21.67 | 23.05 | 23.05 |
- *** ,
- ** , and
- * indicate that the estimated coefficients are statistically significant at the 1%, 5%, and 10% levels respectively in two-tailed t-tests based on robust standard errors.
- This table reports the regression estimates of the relation between a firm's commitments to audited financial reporting (as measured by Big4Auditor, AuditFee, and ExcessFee) and management forecast frequency (MF Freq). All firm-level continuous variables are winsorized at the 1st and the 99th percentiles. All regressions include country, industry, and year fixed effects. All regressions include robust standard errors clustered by both industry and year. Refer to Appendix A for more detailed variable definitions.
Table 7 tests whether the stock market reaction to a firm's earnings forecasts increases with its commitment to better audit verification. We use a model similar to Equation 1 but use the absolute value of the 2-day cumulative market-adjusted return during the earnings forecast window as the dependent variable. Given the effect of other forecast properties on the perceived informativeness of earnings forecasts issued by a firm, we further augment Equation 1 by including three major forecast properties in the model: forecast frequency, forecast precision, and forecast timeliness. More frequent forecasts, forecasts that are more precise, and forecasts that are timelier can presumably lead to a more positive stock market reaction to the forecasts.20 Because a firm can issue multiple forecasts in a single year, this analysis is conducted at the forecast level instead of the firm-year level. Each forecast is therefore counted as one individual observation. Across all three measures of audit commitment, and for both the full earnings forecast sample and standalone earnings forecasts sample, we continue to find a significant and positive association between commitment to audited financial reporting and stock market reaction (measured by the 2-day cumulative market-adjusted return during the 2-day event window around earnings forecasts) to a firm's earnings forecasts. This finding lends strong support to the notion that auditing plays an important role in enhancing or signaling the perceived credibility of a firm's voluntary disclosure.
1 | 2 | 3 | 4 | 5 | 6 | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
All forecasts | Standalone forecasts | |||||||||||
Dep. variable | AbsCAR | AbsCAR | AbsCAR | AbsCAR SA | AbsCAR SA | AbsCAR SA | ||||||
Model | OLS | OLS | OLS | OLS | OLS | OLS | ||||||
Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | |
Big4Auditor | 0.120** | 0.05 | 0.203*** | 0.06 | ||||||||
AuditFee | 0.047*** | 0.02 | 0.044** | 0.02 | ||||||||
ExcessFee | 0.046*** | 0.02 | 0.044** | 0.02 | ||||||||
MF Freq | 0.114*** | 0.02 | 0.127*** | 0.02 | 0.127*** | 0.02 | 0.080*** | 0.02 | 0.120*** | 0.03 | 0.120*** | 0.03 |
MF Prec | 0.074*** | 0.02 | 0.034 | 0.02 | 0.034 | 0.02 | 0.067*** | 0.02 | 0.004 | 0.03 | 0.004 | 0.03 |
MF Time | 0.001 | 0.00 | 0.001 | 0.00 | 0.001 | 0.00 | 0.001 | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 |
NumEA | −0.105*** | 0.02 | −0.060*** | 0.02 | −0.060*** | 0.02 | −0.094*** | 0.02 | −0.053* | 0.03 | −0.053* | 0.03 |
LnAssets | −0.299*** | 0.01 | −0.325*** | 0.02 | −0.297*** | 0.02 | −0.304*** | 0.02 | −0.324*** | 0.03 | −0.298*** | 0.03 |
ROA | −0.004 | 0.00 | 0.007 | 0.01 | 0.006 | 0.01 | −0.007 | 0.00 | 0.005 | 0.01 | 0.005 | 0.01 |
Accruals | 0.045 | 0.06 | 0.077 | 0.06 | 0.078 | 0.06 | −0.205*** | 0.08 | −0.286** | 0.14 | −0.285** | 0.14 |
Segment | 0.067** | 0.03 | 0.085** | 0.04 | 0.088*** | 0.04 | 0.011 | 0.03 | 0.042 | 0.06 | 0.046 | 0.06 |
Loss | 0.808*** | 0.07 | 0.742*** | 0.09 | 0.740*** | 0.09 | 0.852*** | 0.09 | 0.765*** | 0.13 | 0.767*** | 0.13 |
Analyst | 0.016*** | 0.00 | 0.013*** | 0.00 | 0.013*** | 0.00 | 0.011*** | 0.00 | 0.007* | 0.00 | 0.007* | 0.00 |
Institution | 0.009*** | 0.00 | 0.008*** | 0.00 | 0.008*** | 0.00 | 0.007*** | 0.00 | 0.008*** | 0.00 | 0.008*** | 0.00 |
Insider | 0.011*** | 0.00 | 0.010*** | 0.00 | 0.010*** | 0.00 | 0.011*** | 0.00 | 0.006 | 0.01 | 0.006 | 0.01 |
SalesGrowth | −0.001 | 0.00 | 0.001 | 0.00 | 0.001 | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 |
StockExch | 0.025 | 0.02 | 0.019 | 0.03 | 0.019 | 0.03 | 0.077*** | 0.03 | 0.045 | 0.04 | 0.048 | 0.04 |
ADR | −0.024 | 0.06 | 0.021 | 0.07 | 0.022 | 0.07 | −0.097 | 0.06 | −0.080 | 0.11 | −0.080 | 0.11 |
Competition | 0.197** | 0.08 | 0.116 | 0.11 | 0.114 | 0.11 | 0.118 | 0.10 | 0.043 | 0.16 | 0.043 | 0.16 |
HiTech | 0.371*** | 0.06 | 0.488*** | 0.08 | 0.484*** | 0.08 | 0.362*** | 0.07 | 0.453*** | 0.12 | 0.453 | 0.12 |
Country indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
Industry indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
Year indicators | Yes | Yes | Yes | Yes | Yes | Yes | ||||||
N | 32,063 | 16,776 | 16,776 | 22,292 | 11,563 | 6,569 | ||||||
Adj. R2 (%) | 3.98 | 4.94 | 4.94 | 5.62 | 5.57 | 5.57 |
- *** ,
- ** , and
- * indicate that the estimated coefficients are statistically significant at the 1%, 5%, and 10% level respectively in two-tailed t-tests based on robust standard errors.
- This table reports the regression estimates of the relation between committing to audited financial reporting (as measured by Big4Auditor, AuditFee, and ExcessFee) and stock market reaction to management forecasts defined as the absolute value of the 2-day cumulative market-adjusted return during the [0, 1] forecast window with day 0 equal to the management forecast date. All tests are conducted in forecast level. All firm-level continuous variables are winsorized at the 1st and the 99th percentiles. All regressions include country, industry, and year fixed effects. All regressions include robust standard errors clustered by both firm and year. Refer to Appendix A for more detailed variable definitions.
4.3 Multivariate regression results: H2
We next formally test H2 by estimating Equations 2a and 2b. The results for Equation 2a are presented in Table 8. For simplicity, we report the results of the full earnings forecasts sample only.21 Across all proxies of audit commitment, we consistently find the interaction term between firm-level audit commitment and the level of capital market development is significantly and negatively associated with firms' voluntary disclosures. This finding is robust using an alternative proxy of capital market development, such as a capital market development measure based on the total number of listed domestic firms divided by the total population of the country. In other words, the results from Table 8 suggest that while commitment to better audit verification is positively associated with the likelihood, frequency, and informativeness of management earnings forecasts, the effect of audit commitment on voluntary disclosure decreases with country-level capital market development.
Panel A: Management forecast likelihood and frequency | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 4 | 5 | 6 | ||||||
Dep. variable | MF Occu | MF Occu | MF Freq | MF Freq | MF Freq | |||||
Model | Logistic | Logistic | Poisson | Poisson | Poisson | |||||
Coef | SE | Coef | SE | Coef | SE | Coef | SE | Coef | SE | |
Big4Auditor | 0.218*** | 0.03 | 0.104** | 0.03 | ||||||
Big4Auditor × CAPMKT | −0.0009*** | 0.00 | −0.0006** | 0.00 | ||||||
AuditFee | 0.080*** | 0.01 | 0.060*** | 0.01 | ||||||
AuditFee × CAPMKT | −0.0004*** | 0.00 | −0.0004*** | 0.00 | ||||||
ExcessFee | 0.058*** | 0.01 | ||||||||
ExcessFee × CAPMKT | −0.0004*** | 0.00 | ||||||||
CAPMKT | −0.002*** | 0.00 | 0.001** | 0.00 | −0.001 | 0.00 | 0.002** | 0.00 | 0.003*** | 0.00 |
NumEA | 0.534*** | 0.01 | 0.474*** | 0.01 | 0.107*** | 0.01 | 0.148*** | 0.02 | 0.139*** | 0.02 |
LnAssets | 0.194*** | 0.01 | 0.172*** | 0.01 | 0.023*** | 0.00 | 0.028*** | 0.01 | 0.032*** | 0.01 |
ROA | 0.021*** | 0.00 | 0.025*** | 0.00 | 0.001*** | 0.00 | 0.002*** | 0.00 | 0.002*** | 0.00 |
Accruals | −0.003 | 0.02 | 0.005 | 0.03 | −0.005*** | 0.00 | −0.006** | 0.00 | −0.006** | 0.00 |
Segment | 0.005 | 0.02 | −0.002 | 0.03 | 0.003 | 0.02 | 0.029** | 0.01 | 0.024* | 0.01 |
Loss | −0.109*** | 0.03 | −0.090** | 0.04 | −0.045*** | 0.01 | −0.053** | 0.02 | −0.049* | 0.02 |
Analyst | 0.035*** | 0.00 | 0.033*** | 0.00 | 0.033*** | 0.00 | 0.030*** | 0.00 | 0.029*** | 0.00 |
Institution | 0.006*** | 0.00 | 0.005*** | 0.00 | 0.001* | 0.00 | 0.001 | 0.00 | 0.001* | 0.00 |
Insider | 0.009*** | 0.00 | 0.006*** | 0.00 | 0.001*** | 0.00 | 0.001*** | 0.00 | 0.001*** | 0.00 |
SalesGrowth | 0.001*** | 0.00 | 0.001** | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 | −0.001 | 0.00 |
StockExch | 0.064*** | 0.02 | 0.056*** | 0.02 | 0.090*** | 0.01 | 0.063*** | 0.00 | 0.065*** | 0.00 |
ADR | −0.160*** | 0.04 | 0.033 | 0.05 | 0.046* | 0.02 | 0.097*** | 0.02 | 0.104*** | 0.02 |
Competition | 0.043 | 0.05 | 0.026 | 0.07 | 0.014 | 0.02 | 0.043 | 0.03 | 0.060* | 0.03 |
HiTech | 0.264*** | 0.06 | 0.074 | 0.08 | 0.088*** | 0.00 | 0.104*** | 0.02 | 0.103*** | 0.02 |
Country indicators | Yes | Yes | Yes | Yes | Yes | |||||
Industry indicators | Yes | Yes | Yes | Yes | Yes | |||||
Year indicators | Yes | Yes | Yes | Yes | Yes | |||||
N | 119,914 | 42,886 | 119,914 | 42,886 | 42,886 | |||||
N (dep. Var. = 1) | 19,049 | 10,290 | ||||||||
Pseudo R2 (%) | 44.60 | 43.20 | 35.82 | 39.84 | 39.39 |
Panel B: Stock market reaction to management forecasts | ||||
---|---|---|---|---|
Dep Var | AbsCAR | AbsCAR | ||
Model | OLS | OLS | ||
Coef | SE | Coef | SE | |
Big4Auditor | 0.270*** | 0.08 | ||
Big4Auditor × CAPMKT | −0.002*** | 0.00 | ||
AuditFee | 0.067*** | 0.03 | ||
AuditFee × CAPMKT | −0.0002 | 0.00 | ||
ExcessFee | ||||
ExcessFee × CAPMKT | ||||
CAPMKT | 0.003*** | 0.00 | 0.002 | 0.00 |
MF Freq | 0.135*** | 0.02 | 0.134*** | 0.02 |
MF Prec | 0.081*** | 0.02 | 0.037* | 0.02 |
MF Time | 0.001 | 0.00 | 0.001 | 0.00 |
NumEA | −0.091*** | 0.02 | −0.058*** | 0.02 |
LnAssets | −0.297*** | 0.01 | −0.325*** | 0.02 |
ROA | −0.004 | 0.00 | 0.007 | 0.01 |
Accruals | 0.042 | 0.06 | 0.076 | 0.06 |
Segment | 0.040 | 0.03 | 0.076** | 0.04 |
Loss | 0.775*** | 0.07 | 0.726*** | 0.09 |
Analyst | 0.016*** | 0.00 | 0.016*** | 0.00 |
Institution | 0.009*** | 0.00 | 0.008*** | 0.00 |
Insider | 0.011*** | 0.00 | 0.011*** | 0.00 |
SalesGrowth | −0.001 | 0.00 | 0.001 | 0.00 |
StockExch | 0.039 | 0.02 | 0.024 | 0.03 |
ADR | −0.059 | 0.06 | 0.009 | 0.07 |
Competition | 0.111 | 0.08 | 0.081 | 0.11 |
HiTech | 0.397*** | 0.06 | 0.488*** | 0.08 |
Country indicators | Yes | Yes | ||
Industry indicators | Yes | Yes | ||
Year indicators | Yes | Yes | ||
N | 32,063 | 16,776 | ||
Pseudo R2 (%) | 4.19 | 4.99 |
- *** ,
- ** , and
- * indicate that the estimated coefficients are statistically significant at the 1%, 5%, and 10% levels respectively in two-tailed t-tests based on robust standard errors.
- This table examines whether and how the relation between a firm's commitments to audited financial reporting (as measured by Big4Auditor, AuditFee, and ExcessFee) and the likelihood, frequency, and informativeness of management earnings forecasts varies with the strength of country-level institutions as measured by country-level stock market development (CAPMKT). All firm-level continuous variables are winsorized at the 1st and the 99th percentiles. All regressions include country, industry, and year fixed effects. Refer to Appendix A for more detailed variable definitions.
Untabulated results confirm that the effect of audit commitment on voluntary disclosure also decreases with the strength of country-level investor protections. Specifically, across all proxies of audit commitment, we consistently find the interaction term between audit commitment and the strength of country-level investor protection is negatively associated with firms' voluntary disclosures. This finding is robust using an alternative proxy of investor protection such as the INVPRO measure from the Global Competiveness Index provided by the World Economic Forum.22
Taken together, the findings of Table 8 support a weakened (strengthened) role of commitment to higher levels of audit verification on firms' voluntary disclosures in countries with relatively more (less) developed institutions. In addition, by demonstrating how the positive relation between better audit verification and voluntary disclosure varies with country-level institutions, our study provides an explanation for the observed heterogeneity in the role of audits in countries around the world. Our finding is also in line with prior research that suggests that firm-level governance mechanisms are more important in countries with relatively weak institutions (e.g., Choi & Wong, 2007; Gomes, 2000; Lang et al., 2003, 2004, 2012).
4.4 Country-by-country analysis
In this section, as an additional robustness test, we estimate Equation 1 separately for each country without a country-level fixed effect. A concern related to this country-by-country analysis is the constrained statistical power in countries with fewer observations.23 Table 9 presents the results. These results show that, for example, in nine (15) countries of our sample countries, the estimated coefficient on Big4Auditor (AuditFee) is positive and statistically significant. Although not all countries show a positive association between various auditing proxies and management earnings forecast likelihood, the total number of observations of the countries with significant results together accounts for more than 50% of our sample size. This finding demonstrates that the findings presented in previous tables (Tables 5, 6, and 7 in particular) indeed indicate an international phenomenon.
1 | 2 | 3 | |||||
---|---|---|---|---|---|---|---|
Dev. variable | MF Occu | MF Occu | MF Occu | ||||
Test variable | Big4Auditor | AuditFee | ExcessFee | ||||
Model | Logistic | Logistic | Logistic | ||||
N | Coef | N | Coef | N | Coef | ||
1 | Australia | 10,442 | 0.268*** | 5,633 | 0.034* | 5,633 | 0.033* |
2 | Austria | 523 | 0.228 | 174 | 0.004 | 174 | 0.011 |
3 | Belgium | 925 | 0.504*** | 400 | 0.193*** | 400 | 0.188*** |
4 | Brazil | 2,168 | 0.170 | 296 | 0.365 | 296 | 0.366 |
5 | Canada | 16,977 | 0.199*** | 5,471 | 0.047* | 5,471 | 0.046* |
6 | Denmark | 1,147 | 0.175 | 702 | 0.280*** | 702 | 0.276*** |
7 | Finland | 936 | 0.574 | 606 | 0.132** | 606 | 0.130** |
8 | France | 4,811 | 0.184* | 2,235 | 0.103*** | 2,235 | 0.098*** |
9 | Germany | 6,193 | 0.037 | 2,450 | 0.128*** | 2,450 | 0.125*** |
10 | Greece | 1,276 | 0.008 | 254 | 0.433*** | 254 | 0.425*** |
11 | Hong Kong | 9,134 | 0.054 | 4,281 | 0.119*** | 4,281 | 0.115*** |
12 | Indonesia | 1,875 | −0.049 | 170 | 0.186 | 170 | 0.191 |
13 | Ireland | 335 | −0.312 | 262 | 0.198** | 262 | 0.197*** |
14 | Israel | 1,686 | −0.029 | 361 | 0.656** | 361 | 0.660** |
15 | Italy | 1,719 | 0.077 | 232 | 0.098 | 232 | 0.100 |
16 | Malaysia | 6,438 | −0.215** | 4,155 | 0.095*** | 4,155 | 0.094*** |
17 | Mexico | 695 | 0.408 | 104 | 2.446* | 104 | 2.437* |
18 | Netherlands | 922 | 0.791** | 517 | 0.064 | 517 | 0.064 |
19 | New Zealand | 787 | 0.655*** | 396 | 0.010 | 396 | 0.011 |
20 | Norway | 1,601 | −0.141 | 521 | 0.070 | 521 | 0.074 |
21 | Philippines | 1,258 | −0.365* | 371 | −0.185 | 371 | −0.187 |
22 | Poland | 2,583 | −0.162 | 262 | −0.075 | 262 | −0.072 |
23 | Singapore | 4,716 | 0.032 | 1,915 | 0.007 | 1,915 | 0.008 |
24 | South Africa | 1,790 | 0.312* | 922 | 0.086 | 922 | 0.085 |
25 | South Korea | 9,352 | 0.255 | 560 | −0.296 | 560 | −0.293 |
26 | Sweden | 2,941 | 0.094 | 992 | 0.084 | 992 | 0.082 |
27 | Switzerland | 1,926 | 0.440 | 1,123 | 0.108** | 1,123 | 0.107*** |
28 | Taiwan | 10,058 | 0.547*** | 735 | −0.581 | 735 | −0.580 |
29 | Thailand | 3,126 | 0.148 | 563 | 0.074 | 563 | 0.075 |
30 | UK | 11,574 | 0.085* | 6,223 | 0.127*** | 6,223 | 0.126*** |
119,914 | 42,886 | 42,886 |
- *** ,
- ** , and
- * indicate that the estimated coefficients are statistically significant at the 1%, 5%, and 10% levels respectively in two-tailed t-tests based on robust standard errors.
- This table examines the relation between a firm's commitments to audited financial reporting (as measured by Big4Auditor, AuditFee, and ExcessFee) and management earnings forecast likelihood across countries. Coef is the coefficient on Big4Auditor, AuditFee, or ExcessFee in our baseline model of Equation 1 after excluding country indicators. All firm-level continuous variables are winsorized at the 1st and the 99th percentiles.
5 CONCLUSIONS
In their study of the relationship between commitment to audited financial reporting and firms' voluntary disclosure, Ball et al. (2012) provide evidence that, in the USA, commitment to higher levels of audit verification of actual financial outcomes can play an important role in signaling the credibility of a firm's voluntary disclosure, which in turn affects the firm's voluntary disclosure practices. Motivated by the study of Ball et al. (2012), Chen et al. (2016) further show that the positive link between commitment to better audit verification and voluntary disclosure can also be generalized to voluntary nonfinancial disclosure. However, to our knowledge, there has been no research into whether the positive link between audit commitment and voluntary disclosure can be generalized to countries outside the USA, and, furthermore, whether and how the positive link varies with the strength of country-level institutions.
In our study, we extend the examination from a US setting to a cross-country setting, and examine whether the positive link between audit commitment and voluntary disclosure is an international phenomenon. Specifically, our findings suggest that the positive association between audit commitment (proxied by the choice of Big 4 auditor, the amount of actual audit fees, and excess audit fees) and voluntary disclosure (measured by the likelihood, frequency, and informativeness of earnings forecasts) indeed demonstrates a global pattern. More importantly, we find that the positive relation between audit commitment and voluntary disclosure decreases with the strength of country-level institutions, as measured by capital market development and level of investor protection. These findings provide further support for the notion that the effect of firm-level commitment to governance mechanisms on voluntary disclosure is contextual. Thus, the results of our study are relevant to accounting regulators, global investors, and academics.
ACKNOWLEDGEMENTS
We thank anonymous referees, Ilias G. Basioudis (the editor), the participants of the Joint International Conference of the Journal of International Accounting Research (JIAR) and Accounting, Organizations & Society (AOS) 2016 at Augsburg, Germany. Xiangting Kong acknowledges the financial support of The National Natural Science Foundation of China (Grant Numbers: 71790603, 71572206, 71272196), the Major Project of Guangdong Humanities and Social Sciences Key Research Foundation (Grant Numbers: 2012JDXM-0002), and The Basic Research Foundation for Assistant Professors of Sun Yat-Sen University (Grant Numbers: 16wkpy04).
ENDNOTES
- 1 Supporting this view, the literature suggests that the auditing service from auditors, especially Big 4 auditors, can have a more important corporate governance implication in countries with better developed capital markets and higher levels of investor protection, due to the higher perceived litigation and reputation costs associated with their audited financial outcomes (Ball, 2001; Francis & Wang, 2008).
- 2 For example, Cao, Myers, Tsang, and Yang (2017) find that management earnings forecasts have a stronger effect on a firm's cost of equity capital in countries with stronger investor protection, suggesting that country-level institutions per se have a positive role in a firm's voluntary disclosure. Also in line with this view, Lang, Lins, and Maffett (2012) document that committing to firm-level transparency has relatively limited incremental effect on liquidity in countries where investor protection and disclosure requirements are strong.
- 3 Additionally, our findings concerning the variations of the role of auditing in signaling the credibility of firms' voluntary disclosure will be informative to international investors when they attempt to evaluate the credibility of firms' voluntary disclosure.
- 4 Following the conceptual framework developed by Black (2001) that links investor protection and capital market development, in our study we examine the extent to which the relationship between audited financial reporting and voluntary disclosure varies with the level of capital market development (an output measure of the strength of country-level institutions) and strength of investor protection (an input measure of the strength of country-level institutions). In additional robustness tests, we also use alternative country-level institutional variables, including disclosure requirement and level of enforcement of stock exchanges, as alternative measures for strength of institutional infrastructure (Frost, Gordon, & Hayes, 2006), and find our conclusions unchanged.
- 5 According to our correspondence with CapitalIQ, which began collecting international management forecast data in January 2002, the data coverage is relatively complete only after 2004. We verify the completeness of the coverage for US firms using First Call. Specifically, we find that after 2004 the CapitalIQ coverage is slightly higher than that of First Call, and includes all First Call firms. Baginski, Hassell, & Kimbrough (2002, p. 32) document that about 2.4% of Canadian firms (estimated through dividing 219 forecasts by 9,075 firm-years and firm-quarters) provided management forecasts in the period 1993–1996. In our sample period, roughly 5.53% of Canadian firms issued management forecasts, which suggests that Canadian firms have significantly increased their forecasts in recent years.
- 6 We compared the original text and the translated version for a random sample of forecasts in foreign languages and found the translations to be accurate.
- 7 To ensure that our results are not sensitive to varying definitions of a firm's country, we redefined it based on the location of the firm's headquarters or its incorporation country rather than its primary listing country, and repeated all of our analyses. Changing the definition of a firm's country does not have any significant effect on either our results or our inferences.
- 8 We conduct tests on US sample only for robustness purposes.
- 9 While in our setting we assume an equally important role for Big 4 auditors in countries around the world, we acknowledge the possibility that this assumption may not hold in some countries. For example, in India (a country not included in our sample), Big 4 auditors are not allowed to operate independently without a partnership with a local auditing firm. We thank an anonymous referee for bringing this issue to our attention.
- 10 The estimated coefficients of the independent variables of the audit fees model are largely consistent with those in prior studies. For example, we find audit fees (AuditFee) are significantly and positively associated with LnAssets, Segment, Loss and Lag, and are negatively associated with ROA.
- 11 Untabulated results for robustness reveal that our conclusions are robust to using all unlagged determinants in estimating excess audit fees.
- 12 Because of the count nature of the management forecast frequency (MF Freq), a Poisson model is appropriate when the variable is used as a dependent variable in a multivariate regression model (Balakrishnan, Billings, Kelly, & Ljungqvist, 2014; Bamber, Jiang, & Wang, 2010; Lennox & Park, 2006). Our inferences are robust if we redefine MF Freq as loge 1 plus the number of forecasts and use an ordinary least squares (OLS) regression model.
- 13 Unlike the tests examining the incidence and frequency of management earnings forecasts, which are conducted at a firm-year level, tests examining the stock market reactions to management earnings forecasts are conducted at the forecast level using an OLS regression model.
- 14 A score of 1, 2, 3, or 4 is assigned to the forecast if it is a qualitative estimate, an estimate with a minimum or maximum value (i.e., open-range forecasts), a closed-range estimate, or a point forecast respectively. Thus, a higher score suggests more precise forecasts.
- 15 Because this variable requires the forecast to be either a point or closed-range estimate, the data for forecast error are available only for a small set of the subsample. As such, we only use these data as an additional control variable in examining stock market reaction to management earnings forecasts in the robustness test.
- 16 For robustness, we also cluster the standard errors by both country and year, or by both industry and year. In all of these settings, the results are qualitatively similar.
- 17 In an untabulated robustness test, we use an alternative measure of CAPMKT, the total number of listed domestic companies divided by the country's total population. Similarly, we obtain an alternative measure of INVPRO from the Global Competiveness Index provided by the World Economic Forum. We obtain similar results and conclusions using these alternative measures for country-level institutions.
- 18 Economic significance is estimated by exponentiating our coefficients of interest and subtracting one, which estimates the change in likelihood that a firm will issue a forecast from the logistic models.
- 19 Some inconsistencies in the estimated sign of control variables are noted as well. For example, results of Table 5 show that firms reporting a loss (Loss) tend to have a lower forecast likelihood for the full forecasts sample, while they also appear to have a higher forecast likelihood in the standalone forecasts sample. We attribute this inconsistency to the possibility that forecasts made by firms reporting a loss are likely to be more contextual and event driven (i.e., standalone forecasts) than forecasts made by firms that have not posted a loss (i.e., forecasts bundled with earnings announcement). In line with this view, Kothari, Shu, and Wysocki (2009) find that managers tend to delay bad news disclosure relative to good news.
- 20 In a robustness test (untabulated) on a subsample using available data on forecast error, we further add earnings forecast error (defined as the absolute difference between the forecasted earnings and actual earnings, scaled by the actual earnings) to the model as an additional control and find our inference unchanged.
- 21 Results based on the standalone earnings forecasts sample are similar, albeit weaker in some cases, which is probably due to the relatively small sample size. For example, the estimated coefficients for the interaction term of AuditFee × CAPMKT for forecast likelihood, frequency, and forecast informativeness test are −0.0004***, −0.0001***, and −0.0003 respectively.
- 22 In robustness tests, we also examine the association between forecasts and commitment to better audit verification with two subsamples, partitioned by country-level capital market development and investor protection, and find results consistent with the findings on the interaction variable. For example, the estimated coefficient on Big4Auditor is 0.295 (with Z-value of 6.44) for countries with stronger capital market development, CAPMKT, whereas it is 0.361 (with Z-value of 3.61) for countries with relatively weak capital market development.
- 23 In robustness tests, we retest the relationship between audit commitment and earnings forecasts using a subsample excluding the three countries with the largest number of observations (Canada, the UK, and Australia). The results of this test do not affect our main conclusion.
APPENDIX A
VARIABLES DEFINITION
Management forecast variables | |
---|---|
MF Occu | Forecast occurrence: an indicator variable equal to 1 if a firm issues a forecast in a given year and 0 otherwise. |
MF Occu SA | Forecast occurrence: an indicator variable equal to 1 if a firm issues a standalone forecast (i.e., a forecast not issued together with an earnings announcement) in a given year and 0 otherwise. |
MF Freq | Forecast frequency: the total number of forecasts issued by a firm in a given year. We set forecast frequency to zero if a firm issues no forecasts in a given year. |
MF Freq SA | Forecast frequency: the total number of standalone forecasts issued by a firm in a given year. We set forecast frequency to zero if a firm issues no forecasts in a given year. |
AbsCAR | Stock market reaction associated with management forecast: the absolute value of the 2-day cumulative market-adjusted return during the [0,1] forecast window, with day 0 equal to the management forecast date. |
AbsCAR SA | Stock market reaction associated with standalone management forecast: the absolute value of the 2-day cumulative market-adjusted return during the [0,1] forecast window, with day 0 equal to the management forecast date. |
MF Prec | Forecast precision: a precision score equal to 1, 2, 3, or 4 assigned to a qualitative, min or max, closed range, or point forecast respectively. |
MF Time | Forecast timeliness: the number of days between the release of a forecast and the earnings realization date (i.e., annual report filing date). |
MF Error | Forecast error: the absolute difference between the forecasted earnings and the actual earnings, divided by the actual earnings (as a percentage). |
Variables of interest | |
Big4Auditor | An indicator variable equal to 1 if a firm's auditor is a Big 4 Auditor and 0 otherwise. |
AuditFee | The natural logarithm of total assets in millions of US dollars. |
ExcessFee | The residual from a regression of log audit fees on the firm-level fee determinants. The choice of determinants is based on Ball et al. (2012), including LnAssets, ROA, Accruals, Current (the ratio of current assets to total assets), Foreign (the ratio of foreign segment sales to total sales), Segment (number of business segments), Leverage, Loss, Dec (an indicator variable for firms with December fiscal year-ends), and Lag (the lag between the fiscal period end and the earnings announcement data). |
Other firm-and industry-levels variables | |
NumEA | The total number of earnings announcements issued by a firm in a given year. |
LnAssets | The natural logarithm of total assets in millions of US dollars. |
ROA | Returns on assets. |
Accruals | A measure of firm-level financial opacity measured by country-, industry-, and year-adjusted total scaled accruals, based on Bhattacharya et al. (2003). Scaled accruals are computed using balance sheet and income statement information: |
ACCRUAL = (ΔCA − ΔCL − ΔCASH + ΔSTD − DEP + ΔTP)/lag(TA) | |
where ΔCA is the change in total current assets, ΔCL is the change in total current liabilities, ΔCASH is the change in cash, ΔSTD is the change in the current portion of long-term debt included in total current liabilities, DEP is depreciation and amortization expense, ΔTP is the change in income taxes payable, and lag(TA) is total assets at the end of the previous year. | |
Segment | The total number of business segments reported by a firm in a given year. |
Loss | An indicator variable equal to 1 if a firm reports a loss in the current period, and 0 otherwise. |
Analyst | The total number of analysts following a firm in a given year. The data are obtained from the Institutional Brokers' Estimate System. |
Institution | Percentage of shares (end of year) held by all types of institutional investors obtained from FactSet Ownership Data in Wharton Research Data Services. |
Insider | The percentage of a firm's common stock held by the chief executive officer. |
SalesGrowth | 24The percentage of a firm's sales growth from year t to year t+1. |
StockExch | The total number of actively traded stock exchanges on which a firm is listed. |
ADR | An indicator variable equal to 1 if a firm cross-lists its securities in the American Depositary Receipt, and 0 otherwise. |
Competition | A measure of competition defined as Herfindahl index × (−1), where the Herfindahl index is calculated as the sum of the squares of fractional market shares of firms within each two-digit Standard Industrial Classification (SIC) industry for each country year. |
HiTech | An indicator variable equal to 1 if a firm is in a high-tech industry (SIC 2833–2836, 8731–8734, 7371–7379, 3570–3577, and 3600–3674), and 0 otherwise. |
Country-level variables | |
CAPMKT | A country-year measure of the level of capital market development. It is defined as the total stock market capitalization of listed companies as a percentage of gross domestic product for each country-year, as obtained from the World Bank. |
INVPRO | A country-year measure of the strength of investor protection index (0–10), obtained from ‘Doing Business Indicators’ by the International Finance Corporation and the World Bank, at http://www.doingbusiness.org. |
APPENDIX B
REGRESSION RESULTS OF US SAMPLE
Table B1 reports the regression estimates of the relation between a firm's commitments to audited financial reporting (as measured by Big4Auditor, AuditFee, and ExcessFee) and management forecast likelihood in panel A, management forecast frequency in panel B, and stock market reaction to management forecasts in panel C for the US sample only. For brevity, only the estimated coefficient of the main variables of interest is reported.
Panel A: Audited financial reporting and management forecast likelihood | |||||||||
---|---|---|---|---|---|---|---|---|---|
Dep. variable Model | 1 | 2 | 3 | ||||||
All forecasts | |||||||||
MF Occu | MF Occu | MF Occu | |||||||
Logistic | Logistic | Logistic | |||||||
Coef | SE | Coef | SE | Coef | SE | ||||
Big4Auditor | 0.259*** | 0.03 | |||||||
AuditFee | 0.047*** | 0.01 | |||||||
ExcessFee | 0.046*** | 0.01 | |||||||
All other controls | Yes | Yes | Yes | ||||||
Country indicators | No | No | No | ||||||
Industry indicators | Yes | Yes | Yes | ||||||
Year indicators | Yes | Yes | Yes | ||||||
N | 56, 353 | 29,027 | 29,027 | ||||||
N (dep. var. = 1) | 17,757 | 12,721 | 12,721 | ||||||
Pseudo R2 (%) | 53.26 | 47.66 | 4766 |
Panel B: Audited financial reporting and management forecast frequency | |||||||||
---|---|---|---|---|---|---|---|---|---|
Dep. variable Model | 1 | 2 | 3 | ||||||
All forecasts | |||||||||
MF Freq | MF Freq | MF Freq | |||||||
Poisson | Poisson | Poisson | |||||||
Coef | SE | Coef | SE | Coef | SE | ||||
Big4Auditor | 0.192*** | 0.01 | |||||||
AuditFee | 0.317*** | 0.01 | |||||||
ExcessFee | 0.313*** | 0.01 | |||||||
All other controls | Yes | Yes | Yes | ||||||
Country indicators | No | No | No | ||||||
Industry indicators | Yes | Yes | Yes | ||||||
Year indicators | Yes | Yes | Yes | ||||||
N | 56, 353 | 29,027 | 29,027 | ||||||
Adj. R2 (%) | 41.43 | 37.63 | 3 |
Panel C: Audited financial reporting and market reaction to management forecasts | |||||||||
---|---|---|---|---|---|---|---|---|---|
Dep. variable Model | 1 | 2 | 3 | ||||||
All forecasts | |||||||||
Abs CAR | Abs CAR | Abs CAR | |||||||
OLS | OLS | OLS | |||||||
Coef | SE | Coef | SE | Coef | SE | ||||
Big4Auditor | 0.008* | 0.00 | |||||||
AuditFee | 0.026*** | 0.01 | |||||||
ExcessFee | 0.025*** | 0.01 | |||||||
All other controls | Yes | Yes | Yes | ||||||
Country indicators | No | No | No | ||||||
Industry indicators | Yes | Yes | Yes | ||||||
Year indicators | Yes | Yes | Yes | ||||||
N | 17,757 | 12,721 | 12,721 | ||||||
Adj. R2 (%) | 14.12 | 14.16 | 14.16 |
- *** ,
- ** , and
- * indicate that the estimated coefficients are statistically significant at the 1%, 5%, and 10% level respectively in two-tailed t-tests based on robust standard errors.
- All firm-level continuous variables are winsorized at the 1st and the 99th percentiles. All regressions include country, industry and year fixed effects. Refer to Appendix A for more detailed variable definitions.
Biographies
Rubing Liu is an assistant professor at the Accounting School of Guangdong University of Foreign Studies. His research mainly focuses on international accounting.
Xiangting Kong is an assistant professor at Sun Yat-Sen Business School, Sun Yat-Sen University, where she earned her PhD degree in accounting. Her research interests mainly include CEO compensation and corporate financial disclosure. She has published papers in the top 10 accounting journals.
Ziyao San is an accounting PhD student at the Schulich School of Business, York University, Canada.
Albert Tsang is an associate professor at the Schulich School of Business, York University, Canada. He earned three Master's degrees (MSc MIS; MSc Acct; MBA) and a PhD degree in accounting from the University of Texas at Dallas. He is a world-known scholar and researcher, and his research focuses on corporate social responsibility, voluntary disclosure, and international accounting.