Khaled Samaha Ph.D., M.Sc., CPA Assistant Professor Of Accoun ng School of Business, Economics and Communica on, The American University in Cairo (AUC)
The objective of this paper is to examine various determinants of implementing the EASs using a significantly larger and more random representative sample of 300 Egyptian listed annual reports for the period ending 2002. The analysis is extensive, including the use of relative total and partial indices. For each annual report; 1 overall and 5 partial indices (3 mandatory and 2 voluntary) are calculated for applicable EASs disclosures requirements. Also, for each annual report; 1 overall and 5 partial indices (3 mandatory and 2 voluntary) are calculated for applicable EASs measurement/presentation requirements.
Khaled Samaha obtained his PhD in Accounting from the University of Manchester â€“ UK in 2005. He is a Certified Public Accountant (CPA) from the Egyptian Society for Accountants and Auditors (ESAA), and is certified by the Egyptian Accounting Syndicate. He is currently an Assistant Professor of Accounting in the Department of Management at the American University in Cairo. He has several papers that are published in academic accounting journals, and are presented in academic and practitioners conferences. He has extensive practical experience in the application of International Financial Reporting Standards (IFRSs) and has recently published one paper about convergence with IFRS in Egypt. His research interests include harmonisation and compliance with IFRSs, financial reporting and corporate governance mechanisms, audit procedures and methodologies, and financial reporting on the internet.
The conceptual model underlying the determinants that may be associated with implementing the EASs in the current study focuses on a number of micro determinants that uses several developments referring to positive accounting theory such as signalling, agency, political process and capital need theories. The research design used is one of the cross-sectional analyses that is based on analyzing the various total and partial indices to assess the impact of the micro determinants on de facto compliance by comparing and testing for any relationship between any of them and firm specific characteristics using multivariate analysis that is based on binary logistic and transformed OLS regression. The regression analysis of the impact of micro level firm-specific determinants supports the proposition that these firmsâ€™ compliance with EASs is responsive to specific attributes of their environment. However, professionalism as measured by the type of auditor, internal corporate governance factor as measured by ownership concentration and structure as measured by market capitalization dominate with respect to the overall explanatory power for variations in compliance levels with EASs disclosure and measurement/presentation practices. In addition, the explanatory power of the EASs measurement/presentation models is lower than the EASs disclosure models indicating that firm specific characteristics motivate the compliance behavior with IASs disclosure requirements more than IASs measurement/presentation requirements in listed Egyptian companies.
1- Introduction In the 1990s Egypt began to move towards a guided free market economy, which involved the reactivation of the stock exchange market in 1995 and a privatisation program intended to increase the role of the private sector and shrink the public sector (Abdel Shahid, 2003; HassabElnaby et al., 2003; Hassan et al., 2006; Samaha, 2005). As one element of this reform process the Egyptian government pursued a policy of harmonisation with the International Accounting Standards Board (IASB) and required listed companies to prepare their accounts based on 22 newly harmonised Egyptian Accounting Standards (EASs) (MOEFT, 1997). While some countries adopt IASB regulations with little or no amendment, as a result of ministerial decision number 503, in October 1997, Egypt established a permanent committee to issue EASs that were to be based substantially on IASB regulation, though adapted for local conditions. This government led reform is an example of de jure harmonisation, which might be expected to lead to increased de facto compliance, although this is not necessarily the case (Chen et al., 2002). For example, although, in recent years, firms from countries around the globe have begun to cite their use of International Accounting Standards (IASs) in financial reports there is
evidence that many have failed to comply with some or all of the IASB requirements (Tower et al., 1999; Street and Gray, 2002; Glaum and Street, 2003). Specifically in relation to Egypt, Samaha and Stapleton (2008) have reported overall compliance levels of 50% on disclosure and 56% on presentation/measurement, which is similar to compliance levels reported in other developing nations (Taplin et al., 2002). The consequence is that the investor community generally believes that information available in published financial statements is of low quality (World Bank, 2001; World Bank, 2002; Dahawy and Conover, 2007; Samaha, 2005). Using theories, such as agency, signalling, political cost and capital needs, a number of studies (for example, Tower et al., 1999; Street and Bryant, 2000; Joshi and Al-Mudhahki, 2001; Street and Gray, 2002; Al-Shammari, 2005) indicate that factors at the micro level, that is, firm specific characteristics (Watts and Zimmerman, 1986, 1990) have explanatory and predictive power in relation to a firmâ€™s compliance with IASs. However, despite the likely adverse impacts of low levels of investor confidence there is little research in Egypt explaining why compliance levels are so low. In this paper we present evidence, which forms part of a larger study, about the firm specific determinants of compliance with the 22 EASs issued in 1997 that are in material terms Arabic translations of IASs. The evidence is based on a large representative sample of 281 Egyptian listed companies, in non-financial sectors and unlike previous studies the paper distinguishes between compliance in relation to disclosure and measurement/presentation requirements. The environment in which this study is conducted is important. Currently, there is no professional body in Egypt with the power to regulate the activities of the profession (World Bank, 2001). The local accounting body Egyptian Society for Accountants and Auditors (ESAA) was established in 1946 by a royal decree, reorganised in 1977 as a non-profit organisation and became a member of IFAC in 1983. It is an association of chartered accountants that develops educational and professional standards for its approximately 1,200 members (World Bank, 2001), about 785 of whom are actively involved in auditing practice, but membership in the society is voluntary (World Bank, 2002).
The ESAA does not have the power to license accountants and auditors or to establish auditing standards and it is often the case that external auditors are deeply involved in their client firms. This may include the preparation of financial statements, the disclosure of footnote items and even the decision making process in relation to year-end accounts (World Bank, 2002). Thus the environment of reporting is different to that of many countries in which IASs are applied. This paper is divided into six further sections. Section 2 provides a literature review and identifies research hypotheses in relation to micro level firm characteristic factors to be tested in the study. Section 3 explains the research methodology employed by describing the sample creation, the construct of the EASs checklist as the dependent variable and the measurement of the firm specific characteristics (independent variables). Section 4 provides the descriptive statistics for the independent variables and tests for multicollinearity. Section 5 provides the statistical findings, which are discussed in Section 6. Finally, the conclusion of the study, its limitations and the scope for future research, are considered in Section 7.
2- Literature, hypotheses independent variables
In this section we review literature which suggests an association between compliance with IASs and firm characteristics, based on theories of reporting, including agency, signalling, political costs and capital needs theories. We then consider the empirical evidence and draw out ten hypotheses for testing on the EASs. Compliance with IASs tends to provide more information (Ashbaugh, 2001; Leuz, 2003), because IAS may be considered as a form of expanded disclosure (Dumontier and Raffournier, 1998; Naser, 1998; Leuz, 2003). Moreover compliance with IASs considerably restricts the opportunity set of accounting choices (Ashbaugh and Pincus, 2001) and prohibits the creation of hidden reserves, giving IAS compliant financial statements greater credibility (Dumontier and Raffournier, 1998). This may have a number of different influences on the propensity to comply. First, compliance may reduce agency costs, making capital easier to obtain. Agency theory postulates a conflict of interest between members of an organisation. There is a characterisation of
weak owners and strong managers (Roe, 1994) so that owners must reduce agency costs and control opportunism. However, agency costs are not only a problem for owners, since they may translate into higher costs of capital for the firm. As a consequence constraints on behaviour may be self-imposed by managers, to avoid ex post misallocation with the intention of increasing the ex ante provision of finance (Shleifer and Vishny, 1986). Since poor disclosure impairs a common understanding between management and investors with respect to a company’s capital investment decisions, it creates information risk (Leuz and Verrecchia, 2004). Consequently managers may be willing to bond with owners by providing information to reduce the cost of capital. Leuz and Verrecchia’s (2001) empirical work supported the proposition that the use of IASs increased the level of disclosure and reduced the cost of capital. A similar argument applies to debt holders, who also suffer agency costs (Jensen and Meckling, 1976). Agency costs of this type are expected to be higher for companies with higher gearing because there is relatively more wealth for managers to transfer from debt holders. Thus highly geared firms can be expected to disclose more information to reduce agency costs by reassuring debt holders that their interests are protected. Second, since diffuse shareholders have limited incentives to conduct monitoring activities, agency costs are usually higher in companies where there is a wide dispersion of share ownership. Consequently, to reduce agency costs, these companies have an incentive to comply with accounting rules (Jensen and Meckling, 1976; Fama and Jensen, 1983), to lower the cost of capital. Third, compliance may create proprietary costs (Verrecchia, 1983). Irrespective of whether disclosure of information has a positive or negative impact on the value of the company, these costs are imposed if competitors, dissident shareholders or employees use the information in a way that is harmful to the company. Given that compliance with IASs increases the level of disclosure, smaller companies are more likely to be reluctant to comply with IASs (Dumontier and Raffournier, 1998; El-Gazzar et al., 1999; Leuz, 2003). Fourth, since gathering, generating and disseminating information in the form of an annual report is a costly activity (Buzby,
1975), additional disclosure is likely to fall more heavily on smaller companies, which may not be able to afford such costs (Dumontier and Raffournier, 1998; ElGazzar et al., 1999; Leuz, 2003). Fifth, the level of scrutiny by financial analysts, who typically follow large companies, may also be relevant. Since compliance with IASs enhances the level of disclosure it enables financial analysts to do a better job (Ashbaugh and Pincus, 2001), which may provide incentives to disclose if this reduces the cost of capital. This may also be linked to the extent of international activities since international companies are more likely to be subject to a broader range of regulatory authorities and to have diverse financiers, suppliers and customers. As a consequence, they are more likely to disclose more detailed information (Malone et al., 1993). Prior research (e.g., Choi, 1973; Zarzeski, 1996; Jaggi and Low, 2000; Archambault and Archambault, 2003) argues that firms competing for foreign resources tend to expand their financial and accounting disclosure as bonding for resource providers. This expanded disclosure is assumed to reduce resource providers’ uncertainty about transactions with the firm and in turn, enables the firm to obtain resources at lower costs. Sixth, in western countries an association has been identified between political visibility and the costs this may bring and the extent of disclosure and compliance with IASs. Politically visible companies generally are more exposed to attack in the form of greater regulation, such as price controls and the threat of nationalisation. In particular, although industry type can be a more appropriate proxy for political cost sensitivity than size (Ball and Foster, 1982), large earnings do render companies more vulnerable to regulatory intervention (Inchausti, 1997), so that large companies are more likely to be subject to wealth transfers as a result of government intervention (Watts and Zimmerman, 1986). Companies subject to such costs have a greater incentive to disclose more information in their annual reports, which may enhance their reputation and public image and lessen public criticism, thus reducing the threat of government intervention (Watts and Zimmerman, 1978; Holthausen and Leftwich, 1983). Furthermore, since adoption of IAS tends to lead to more conservative income, compliance with IASs signals not only greater earnings ‘quality’ and superior
performance to the market (Dumontier and Raffournier, 1998), but may also reduce political costs (Watts and Zimmerman 1986). Seventh, signalling theory predicts that managers of successful companies will wish to signal their success and strength to potential investors (Inchausti, 1997) and justify their compensation (Singhvi and Desai, 1971; Inchausti, 1997). This may be achieved by disclosing more information and by the related choice of external auditor. The signalling literature (Titman and Trueman, 1986; Craswell and Taylor, 1992) suggests that the choice of a large multinational auditor signals quality because they have greater incentives to resist clients’ pressure for limited disclosure to avoid adverse economic consequences to their reputation (DeAngelo, 1981). Furthermore, these firms probably have a competitive advantage in controlling the application of IASs because of superior staff training and the existence of economies of scale in the development of competence in IASs (Dumontier and Raffournier, 1998), so are less likely to be associated with low compliance firms. From this review of the international literature on compliance we draw out firm characteristics, establish hypotheses about their likely association with compliance with the substantially similar EASs in Egypt and identify our independent variables. We discuss each of the following characteristics in turn: profitability, liquidity, size, gearing, ownership diffusion, external auditor type, industry sector, share trading, legal form and internationality. Profitability – H1 Companies with larger profits have a higher level of compliance with the EASs. Although this is suggested by the literature above, it should be noted that Tower et al. (1999), Street and Bryant (2000) and Street and Gray (2002) found no association between profitability and the level of voluntary compliance with IASs. Similarly, Glaum and Street (2003) and AlShammari (2005) reported no relationship between profitability and the extent of mandatory compliance with IASs disclosure requirements. Nevertheless there are two reasons for this hypothesis. First, Dumontier and Raffournier (1998) argued that compliance with IASs by profitable companies is one way to signal their superior performance to the market. Since compliance with IASs could provide more information (Ashbaugh, 2001; Leuz, 2003),
it can be predicted that companies with larger profits are more likely to comply with IASs than companies with smaller profits. Second, Watts and Zimmerman (1986, p.235) further argued that larger companies that tend to earn the largest profits in absolute terms are more vulnerable to regulatory intervention. It may be that providing financial statements based on IASs enables a company to meet the demand of the government for information and thereby, reduce political costs. If so, then companies with larger profits have a greater incentive to comply with IASs and a similar hypothesis is thus proposed in relation to EASs. In this research, two measures of accounting profitability are defined in line with Dumontier and Raffournier (1998): ROE 1 is the ratio of net income before tax to total shareholders’ equity at the balance sheet date and ROE 2 is the ratio of net income to total shareholder’ equity at the balance sheet date. Liquidity – H2 Companies with lower liquidity ratios have a higher level of compliance with the EASs. Regulatory bodies, debtholders and shareholders are particularly concerned with the going-concern status of companies (Wallace and Naser, 1995) and the Egyptian Capital Market Authority refers to companies’ liquidity ratios to assess their ability to meet short-term commitments (CASE, 2006). Companies that seek external financing through lending agreements with banks are also subject to restrictive covenants related to liquidity, although relevant major debt holders usually have access to private information from companies (Craswell and Taylor, 1992). Demand for company information by the CMA and shareholders are expected to increase when liquidity is lower. As a result, a company with a lower liquidity ratio is more likely to disclose additional information in order to reduce agency costs and assure shareholders and the CMA that its financial position is sufficient to meet short-term commitments. In the current research, the variable LIQUID is measured by quick (acid test) ratio as it is a more stringent measure of corporate liquidity (Owusu-Ansah, 1998). It is defined as the ratio of current assets less stock to current liabilities. Size – H3 Larger companies have a higher level of compliance with the EASs. Although this hypothesis is suggested by the literature empirical, evidence is mixed.
Joshi and Al-Mudhahki (2001) have identified size as being positively associated with mandatory compliance with IAS 1 and Al-Shammari (2005) found it positively associated with compliance with mandatory disclosure requirements in the Gulf Council member states. However, other authors have found no such association (Tower et al., 1999; Street and Bryant, 2000; Street and Gray, 2002 in relation to voluntary compliance and Glaum and Street, 2003 in relation to mandatory compliance). The impact of larger companies on the Egyptian economy is substantial. They employ many nationals, consume a large quantity and value of raw materials and are largely responsible for imports (Carana, 2002; MOFT, 2007). Consequently, in Egypt they are politically visible so are more likely to be subject to government intervention and possible political costs. Also, as part of its economic reforms, Egypt has been implementing privatisation programs involving large, state-owned companies, so that larger companies are more likely to be closely followed by potential investors. The size of a company can be measured in a number of different ways. In the current study, three size variables will be considered in line with Street and Gray, (2002) namely: Total assets in Egyptian pounds (ASSETS), Total sales in Egyptian pounds (SALES), Market capitalisation (MCAP) in Egyptian pounds (as it is argued by Naser (1998) that financial disclosures by firms listed on stock exchanges can be influenced by total market capitalisation. Gearing – H4 Companies with higher gearing have a higher level of compliance with the EASs. The association between the extent of compliance with IASs and the level of gearing was investigated by Tower et al. (1999), who found no association. However, Al-Shammari (2005) found it positively associated with compliance with mandatory disclosure requirements in the Gulf Council member states. In Egypt the Cairo and Alexandria Stock Exchanges (CASE) and CMA (the regulators) are especially concerned about firms’ gearing and those with higher levels are subject to greater regulatory demand for detailed information so that the regulatory body can assess obligations and future cash flows. Following El-Gazzar et al. (1999), Murphy (1999) and Tower et al. (1999), the GEARING variable is measured as the ratio
of total long-term debt to equity at the end of the year. Ownership diffusion – H5 Companies with less concentrated share ownership structures have a higher level of compliance with the EASs. Prior research has found mixed results on the relationship between level of disclosure, ownership diffusion and cost of capital. A recent study by Glaum and Street (2003) found no relationship between mandatory compliance with IAS disclosure requirements and ownership diffusion, as measured by a company’s free float (percentage of equity capital that can be readily traded in the share market). Also, Al-Shammari (2005) found it to be insignificant in the GCC countries. In earlier studies Solas (1994), in Jordan, found no significant relationship between ownership diffusion and disclosure, whereas, Dumontier and Raffournier (1998), in Switzerland, showed a positive influence of ownership diffusion on voluntary compliance with IASs. By applying the Pedersen and Thomsen (1997) model to the Egyptian Stock Exchange, the ownership structure in listed Egyptian companies is characterised by medium to high concentration, which confirms findings by the World Bank (2001, 2003). In this model the concentration of ownership is considered to be: low where no entity or individual owns more than 20% of a company’s voting stock; medium where a single stockholder, whether an individual, a family, a government or a company, owns between 20%–50% and high if one person or family or the government owns the majority (i.e., more than 50%) of the voting stock. There were 1074 companies listed by end of 2002, of these around 600 are highly concentrated ownership companies (World Bank, 2001, 2003). According to the report by the Carana Corporation1 (2000), companies in this category are either effectively controlled by the State with the State as the largest shareholder or subsidiaries of one of the many industrial groups (holding companies) found in Egypt today or companies that were typically founded as private sector enterprises and have grown to comprise much of Egypt’s corporate sector. They tend to be family owned, controlled, managed and have a ‘patriarchal’ business culture in which most control and oversight functions are vested in the founder/family patriarch. The rest of the listed companies are medium ownership
concentration companies, of which 100 are actively traded companies (World Bank, 2001, 2003). The sample companies are categorised as above into either medium or high concentration only. According to this, the DIFFUSE variable is a dummy variable that is characterised by medium ownership concentration assuming to some extent wider ownership diffusion and therefore is coded one and otherwise is coded zero indicating high ownership concentration. Type of auditor – H6 Companies audited by a local audit firm with an international affiliation (Big Four) have a higher level of compliance with the EASs. Empirical evidence internationally is mixed. Street and Gray (2002) reported that the level of compliance with IASs disclosure and measurement requirements was positively associated with companies in developed countries being audited by big four auditing firms and Glaum and Street (2003) found a similar outcome in Germany. Al-Basteki (1995) and Dumontier and Raffournier (1998) also provide evidence of a significant positive relationship between being audited by a big four firm and voluntary compliance with IASs practices. Conversely, Murphy (1999) and Street and Bryant (2000) and Al-Shammari (2005) provide no evidence of this association. All the major international audit firms had a presence in Egypt at the time of the study, but there is a gap in expertise between these auditors and local firms. Many practicing auditors, who are not members of the ESAA, lack necessary professional competence for providing high-quality auditing services and as a result, do not play their expected role (World Bank, 2002). In this study the AUDIT variable is coded as follows: one if the firm is audited by a big four international firm and zero otherwise. Industry sector – H7 Manufacturing companies have a higher level of compliance with the EASs. Research provides a range of results in regard to the association between industry type and the extent of compliance with IASs. While Street and Bryant (2000) and Al-Shammari (2005) found no evidence of such an association, Street and Gray (2002) reported a positive association. It is also possible that industry type may capture sensitivity to political costs not captured by other proxies that differ by industry [Ball and Foster, 1982; Watts and Zimmerman, (1986), p.239].
Three broad industries were identified in Egypt, namely manufacturing, trade and services according to the US country report 2001 (Bureau of Economic and Business Affairs, 2002). The manufacturing sector exhibits extraordinary efficiency and productivity and together with the significant international exposure of this sector may have an effect on the extent of compliance with EASs in corporate annual reports that differs from other sectors (Bureau of Economic and Business Affairs, 2002). Since manufacturing is of fundamental importance to Egypt and in part some industries are price controlled (i.e., high political visibility for manufacturing companies), it is possible that levels of compliance with EASs in the corporate annual reports of manufacturing companies may be higher than those in other business sectors in order to reduce political costs. The INDUSTRY variable is coded as follows: one for manufacturing (industrial) and zero for trade or services (nonindustrial). Share trading – H8 Actively traded companies have a higher level of compliance with the EASs. Many previous studies (Dumontier and Raffournier, 1998; El-Gazzar et al., 1999; Murphy 1999; Street and Bryant 2000; Street and Gray, 2002) have looked at the influence of listing status or multi-listing, on the level of compliance with IASs. Multilisting could not be tested for Egyptian companies as only nine were multi-listed at the time of the study (CASE, 2000). Of the 1074 companies listed by end of 2000 around 400–500 were closed companies and rarely or not traded (World Bank, 2001, 2003). The rest are of medium ownership concentration and of these 100 are actively traded companies (World Bank, 2003). It is assumed that actively traded companies list to raise capital, while rarely or not traded companies list for tax exemption purposes. The SHARE variable is coded as follows: one if actively traded and zero if rarely or not traded. Legal form – H9 Private sector companies have a lower level of compliance with the EASs. In Egypt ‘public sector companies’ are state-owned and were listed as a preliminary to privatisation and at the time of the research most were partially privatised. They tended to be amongst the largest Egyptian companies (Abd-Elsalam, 1999) and since they intend to issue securities in
the near future capital needs theory is especially relevant. These companies are intensively in the public eye as the government and international agencies regularly evaluate the privatisation program. Companies described as ‘private sector companies’ are held entirely or in the majority by private sector institutions or individuals and were less likely to issue securities in the near future, as they were initially perceived as less attractive to investors. However, from 1998 a number of new large private sector offerings presented more investment opportunities and stimulated market sophistication as investors began to focus on private sector ‘growth’ stock investment as against public-sector ‘value’ stock trading. A shift from ‘old economy’ value stocks to ‘new economy’ growth stocks began to appear, as private sector companies with sound management skills and growth potential offered shares to the public (Abdel Shahid, 2003). This encouraged a higher level of sophistication in trading decisions and stimulated longterm investment. The legacy of centralisation and public sector domination of the economy is over, as the private sector now accounts for more than 76% of the Egyptian economy. In the absence of published studies of attitudes of Egyptian investors to compliance with EASs according to their legal form after the harmonisation with IASs in 1997, it is very difficult to predict the direction of association. However, our hypothesis was designed for ease of comparability with previous disclosure research in Egypt before 1997 by AbdElsalam (1999). The LEGAL variable is coded as follows: one for a private sector company and zero otherwise. Internationality – H10 Companies with relatively more international activities have a higher level of compliance with the EASs. Studies found mixed results with respect to the relationship between internationality and compliance with IASs. Glaum and Street (2003) found that the level of mandatory compliance with IAS disclosure requirements was positively associated with companies being cross-listed on US stock markets. Al-Shammari (2005) found it positively associated with compliance with mandatory disclosure requirements in the Gulf Council member states, but Street and Gray (2002) found no such association.
Zarzeski (1996) used total foreign sales of a firm to classify the firm as an international firm. Jaggi and Low (2000) used criteria based on the following: 1- ratio of foreign sales to assets 2- export, i.e., domestic production sold in foreign markets 3- being an MNC affiliate. The current study also employs a similar approach to capture a firm’s internationality and these include: foreign business transactions (foreign sales or exports), being an MNC affiliate trading of the firm’s shares on foreign stock exchanges (albeit very rare in listed Egyptian companies). If a sample company is classified as international the variable INTER is coded one, otherwise zero.
3- Method In this paper, we test hypotheses about the relationship between compliance with EASs and ten independent variables, using data about levels of compliance that has been previously reported (Samaha and Stapleton, 2008), so that in this section we provide only a brief review of compliance measurement and sample selection and focus on method as it applies to the independent variables to be measured here. That earlier paper constructed a checklist to measure Egyptian companies’ compliance with both the disclosure and the measurement/presentation requirements of the IASB in a way that is comprehensive but also appropriate to the specific environment of Egyptian accounting. The checklist was based on items used by previous authors but was designed to: focus compliance calculations on those regulations with which the Egyptian government intends companies should comply (i.e., mandated EASs); differentiate between the EASs which were part of the Egyptian accounting regulations prior to 1997 and those that were not, to provide a finer explanation of the effect of the companies’ attributes and their familiarity or otherwise with EASs and identify IAS regulations that are not mandated yet as EASs, but that appear relevant to Egyptian companies, as revealed by screening a sample of 30 annual reports. In this earlier analysis the standards with which measurement was to be measured were divided into six categories as follows: 1A group of nine IASs mandated as EASs in 1997, which were covered in Egyptian regulations prior to 1997 and for which there is no difference between the old and the new regulation hereafter referred to as mandated covered no difference
(MCND). This category relates to IAS 2, 10, 11, 16, 18, 21, 24, 25 and part of IAS 1 (excluding the statement of changes in equity). The transfer of IASs into EASs is not necessarily on a one to one basis. For example, IAS 1 is translated into three EASs – EAS 1, EAS 3 and EAS 9. 2A group of six IASs now mandated as EASs, which were covered in previous regulations but with differences between the old and the new regulations [mandated covered with differences (MCD)]. This category relates to: IAS 8, 20, 23, 27, 28 and part of 38 relating to research and development. 3A group of four IASs now mandated as EASs, which were not covered in previous regulations [Mandated New (MNEW)]: IAS 7, 31, 33 and 35. 4A group of four IASs not mandated as EASs in 1997, but which were covered in previous regulations although with differences between the Egyptian and the IAS regulations [not mandated covered with differences (NMCD)]. These are relevant because there appears to be an intention to converge with the IASB in the near future. This category relates to: IAS 17, 22, 37 and part of IAS 38 on pre-operating and set-up costs. Of the three IASs not mandated as EASs, two (on provisions and business combinations) are covered by previous Egyptian accounting regulations that differ from the relevant IAS. Only one EAS (leases) exists that is not based on the equivalent IAS 17. 5The part of IAS 1 that relates to the statement of changes in equity is a separate category as this was not mandated as an EAS, not covered in previous regulations not mandated not covered (NMNC) but again is relevant because there appears to be an intention to converge with the IASB in the near future. 6A final group of ten IASs, which were not mandated as EASs in 1997. In relation to eight of these (IASs 14, 15, 29, 32, 36, 39, 40 and 41) there was no Egyptian regulation at the time of the study and there appeared to be no intention to regulate in the near future because these IASs cover areas of accounting that are not common in Egyptian accounts (for example segmental reporting) or are not applicable to an economy with low inflation rates. Two standards IAS 12 and 19 were covered in previous Egyptian regulations but with differences and there appeared to be no intention to converge with the IASB in the foreseeable future.
Based on this, we focused on those standards that are applicable to Egypt, after the partial convergence policy with IASs in 1997, that is: the 19 IASs in Categories 1 to 3 above, which have been mandated as 21 EASs (i.e., MCND, MCD and MNEW)3, three full standards and two part standards in Categories 4 and 5 above (i.e., NMCD and MNNC) with which, at the time of the study, compliance was voluntary. The standards in Category 6 above were excluded from our calculations because the Egyptian government does not intend that companies should comply now or in the immediate future.
3.1 Sample selection The target population is the non-financial listed Egyptian companies on the CASE for the year ending 2002 of which there were 934. Three categorical variables were chosen for the stratification process to select the sample: whether the company is actively traded; whether the company is in the public or private sectors; whether the company is in the manufacturing or trade and services sector. An initial sample size of 300 companies, which represents 32% of the total population, was targeted following the Saunders’ et al. (2000, p.463) method, which resulted in a largest adjusted minimum sample size of 272 companies. This sampling frame was then divided into the discrete strata and all 79 actively traded companies were selected on the basis that they are the most visible companies on the CASE and they include all 15 companies partially owned by the government. All 22 of the remaining public sector companies listed for privatisation were also selected. This left 199 companies to be selected at random, using the random number generator provided by Excel, in proportion to the sector divide 59% manufacturing and 41% trade and services. All 300 sets of annual accounts were sought and eventually the researchers obtained 281 usable sets, the remainder being incomplete. This provides a sample size that is significantly larger than the sample sizes used in previous work in developing countries. Table 1 shows an analysis of the sample companies by categories. Four overall compliance indices were calculated: two for disclosure (DIS 1 and 2) and two for measurement/presentation (MEAS/PRES 1 and 2). DIS 1 and MEAS/PRES 1 assigned an equal weighting
to each EAS, whereas DIS 2 and MEAS/PRES 2 assigned equal weighting to each item of disclosure and treatment/presentation. In addition, we also calculated five partial indices for each of the disclosure and measurement/presentation requirements, corresponding to the Categories 1 to 5 identified above.
INTER + e The
regression was utilised for the remaining 12 continuous
regression routines were then tested using stepwise search, on SPSS (2000), in accordance with studies such as Street and Bryant (2000) and Street and Gray (2002).
3.2 Measurement of the explanatory (independent) variables
The OLS regression equation is:
Multivariate statistical analyses are used to test for any relationship between the firm specific characteristics and de facto compliance with EASs. Fourteen regression models were run. Two overall EAS disclosure and two overall measurement/presentation regressions enable strong conclusions to be drawn on which firm characteristics motivate overall compliance in listed Egyptian companies. Five partial EAS disclosure and five partial measurement/presentation regressions enable finer explanation to be drawn on whether determinants of compliance have the same explanatory power depending on whether the mandated EASs were part of the Egyptian accounting regulations prior to 1997 or not and on whether the IASs practices are mandated as EASs or not. This is relevant to the contribution of positive accounting theories in explaining compliance. The multivariate test used in this study is the binary logistic regression analysis for two dichotomous indices, the NMNC disclosure and measurement/presentation indices. In accordance with studies such as Cooke (1989a, 1989b, 1991) and Inchausti (1997) a stepwise model is used to determine which of the independent variables could best explain compliance. Since there were a few cases with continuous values these required reclassification and values below 0.8 were reclassified as zero (no compliance) and values above 0.8 were reclassified as one (full compliance). The binary logistic regression equation is:
Indexj = β0 + β1 ROE2j + β2 LIQUIDj + β3
Z= β0 + β1 ROE2j + β2 LIQUIDj + β3 ASSETj + β4 SALESj + β5 MCAP+ β6 GEARING+ β7 DIFFUSE+ β8 AUDIT +
β9IND +β10SHARE+ β11LEGAL+ β12
ASSETj + β4 SALESj + β5 MCAP+ β6 GEARING+ β7 DIFFUSE+ β8 AUDIT +
β9IND +β10SHARE+ β11LEGAL+ β12 INTER + e As advised by Cooke (1998) nonparametric OLS regression was used, based on the transformation into normal scores based on the Van der Warden approach4 to avoid violating the regression assumptions.
4 Descriptive statistics of the independent variables and tests for multi-collinearity Before testing hypothesis H1 to H10, corporate characteristics were screened. Their descriptive statistics are reported in this section. The firm characteristics include four continuous variables and six categorical variables. Descriptive statistics for the continuous variables (three size variables, liquidity, gearing and two profitability ratios) are shown in Table 2. The statistics on return on equity (ROE) indicate that a small number of firms exhibit negative ROE. Although, the maximum ROE is 177.0%, the mean ROE is 10.84% (ROE 1) and 8.84% (ROE 2), which shows that the sample firms on average were not financially sound for the study period. This may have implications for the relevance of signalling theory. The mean liquidity ratio is 1.2. The debt equity (gearing) ratio shows that on average the ratio of debt to equity is 36.27%, meaning that long-term debt is a substantial portion of the capital structure of sample firms. The average firm size in terms of market capitalisation is 150 million Egyptian pounds (approximately, 44 million US dollars) with the maximum market capitalisation being 7.5 billion Egyptian pounds (approximately, 2.2 billion US dollars). The range of the total market
capitalisation indicates that the sample firms are widely distributed. Descriptive statistics for the categorical variables are shown in Table 3. The mean of the type of auditor variable is 31.67%, indicating that the sample is dominated by local audit firms. The variable on ownership concentration shows that on average only 36.65% of the companies have medium ownership concentrations, implying that ownership structure is highly concentrated in listed Egyptian companies (63.35%). The mean of the extent of internationality variable indicates that 38.79% of the sample firms have a significant proportion of foreign revenues or are an MNC affiliation. Prior to performing the multivariate analysis, a multicollinearity test was performed. A bi-variate correlation above 0.80 is problematic (Field, 2003) and such a correlation was found between ROE 1 and ROE 2 (both measures of profitability). Consequently, the variable ROE 1, which had not been selected by any of the stepwise regression models, was eliminated. The next highest correlation coefficient was between the three size variables: market capitalisation and total assets (r = .64); total assets and total sales (r = .62); total sales and market capitalisation (r = .58). A moderate level of correlation was also found between: ownership concentration and share trading (r = .57); share trading and total assets (r = .55) and share trading and market capitalisation (r = .50). All these are, however, well below the 0.8 level for concern (Street and Bryant, 2000; Street and Gray, 2002; Field, 2003). It is also the case that none of the variable correlation is higher than 0.64, so that, after eliminating ROE 1, the problem of multicollinearity is minimal, giving support to the appropriateness of multivariate analysis.
5 Statistical findings In this section, we first briefly summarise the previously reported results on levels of compliance before presenting our new findings from the stepwise regression disclosure and measurement/presentation models.
5.1 Levels of compliance with EASs The findings, in Tables 4, 5 and 6 show that there are large variations in compliance with EASs among the sample companies in Egypt, but that the overall compliance level is very low at just 50% for EASs disclosure indices and 56% for EASs measurement/presentation indices.
Using descriptive statistics, as shown in Table 5, it appears that where preparers were familiar with the regulations compliance is higher than when the convergence with IASs brought about change to previous regulations or introduced new regulations. Table 6 shows the descriptive statistics of the compliance levels with disclosure and measurement/presentation partial indices in relation to the three non-mandated IASs and the non-mandated parts of IAS 38 and IAS 1. Although there are large variations in voluntary compliance, the mean level of compliance is minimal.
5.2 Multiple regression models and assumptions testing Tables 7, 8 and 9 show the results of six disclosure models (one overall [DIS 2] and five partial) and one measurement/presentation model (MEAS/PRES 2) used to test hypotheses H1 to H105. These results are not materially affected by robustness testing, which included the Kolmogrov-Smirnov (K-S) test, a visual inspection of the histogram and normal probability plot of the residuals in the 12 OLS models. This is to be expected since normal scores are normally distributed, implying normality of the error term (Camfferman and Cooke, 2002). To further examine the sensitivity of the results, the whole sample was divided into two random samples and the regression run on each separately, as recommended by Field (2003). The results on the split samples confirmed the results on the whole sample. A further check was conducted to see if the results are driven by outliers by removing the influential cases as recommended by Field (2003). The results are not materially affected by any of these robustness tests. As shown in Table 7, the multiple regression models for the overall levels of disclosure and measurement/presentation have reported F values of 137.676 and 132.730 respectively (both significant at the .0001 level) and adjusted coefficients of determination (R2) of 71% and 69% respectively. Thus, these were highly significant and had an explanatory power which is higher than those reported in earlier studies in an Egyptian setting (e.g., Abd Elsalam, 1999; Hassan et al., 2006). However, there were some apparent differences in the explanatory power between the partial models as shown by the R2 in Tables 8 and 9. These range from
48% in the case of MNEW to 72% for MCND, whereas the partial measurement/presentation models (not reported) varied from 47% for MCD to 67% for MCND. Thus the measurement/presentation models have lower explanatory power than the disclosure models, which may have implications for theory as will be discussed in Section 6.
5.3 Results of hypotheses testing In terms of the association between company characteristics and the extent of compliance Tables 7, 8 and 9 show that just five independent variables were found to motivate compliance with both the disclosure and measurement/presentation requirements of EASs. We consider each of these five in turn. First, there is a highly positive association between compliance with both overall disclosures and measurement/presentation and being audited by a big international firm (Table 7). The type of auditor alone explains more than 40% of the DIS 2 and MEAS/PRES 2 indices (both significant at the .0001 level), providing support for hypothesis H6 that companies audited by a local audit firm with an international affiliation (big four) have a higher level of compliance with the EASs. This is consistent with previous literature in other countries which has drawn on agency and signalling theories (Street and Gray, 2002; Glaum and Street, 2003), but is inconsistent with evidence from Murphy (1999), Street and Bryant (2000) and AlShammari (2005). Second, there is a positive association between medium ownership concentration and overall compliance (significant at the .0001 level) and all four partial indices shown in Table 8. Thus, hypothesis H5 is supported. This is consistent with agency theory suggesting that companies with widely held ownership are more likely to comply more than companies with closely held ownership structure (Dumontier and Raffournier, 1998), but this evidence contrasts with earlier studies (Solas, 1994; Glaum and Street, 2003; Al-Shammari, 2005). Third, size as measured by market capitalisation (significant at the .0001 level) is positively associated with overall compliance, although not with the partial indices MCD and MNEW, so that, hypothesis H3 is supported. Thus, both agency theory (e.g., Jensen
and Meckling, 1976; Fama and Jensen, 1983) and political process theory (Watts and Zimmerman 1986) are supported. Fourth, international companies are shown to comply more overall (significant at the .0001 level) although the association does not hold with MNEW. This provides support for hypothesis H10 in line with Glaum and Street (2003) and Al-Shammari (2005) and is consistent with Zarzeski’s (1996) argument that companies will disclose more information if they have large proportions of foreign sales. International firms may find compliance with recognised accounting standards reduces the chance of misunderstanding and misinterpretations, thereby strengthening confidence (Wyatt, 1997). Finally, actively-traded companies are shown to comply overall more (significant at the .0001 level), although this only holds for the one partial index MCND. This provides support for hypothesis H8 in line with many previous studies (Dumontier and Raffournier, 1998; El-Gazzar et al. 1999; Murphy 1999; Street and Bryant 2000; Street and Gray, 2002) that have looked at the influence of listing status or multilisting, on the level of compliance with IASs. In line with these findings, the lowest levels of compliance were as low as 16% on disclosure and 17% on measurement/presentation and were obtained by companies which did not have a large international audit firm, were closelyheld and were not international, whereas the highest levels of compliance at 90% on disclosure and 93% on measurement/presentation were obtained by companies audited by large international firms. In relation to the remaining five independent variables there was no association with overall compliance, although Table 8 does show some association with specific partial models. The lack of association between compliance measured by any of the indices in relation to profitability (H1) and liquidity (H2), while consistent with findings of prior studies (Tower et al., 1999; Street and Bryant, 2000, Street and Gray, 2002; Glaum and Street, 2003; Al-Shammari, 2005), is not consistent with signalling and political process theories arguments (Inchausti, 1997; Dumontier and Raffournier, 1998; Watts and Zimmerman, 1986). Similarly, the lack of association between compliance and gearing (H4), while consistent with the
findings of Tower et al. (1999), is not consistent with agency theory arguments (Jensen and Meckling, 1976). Moreover, while the finding that industry sector (H7) is insignificant in relation to compliance is consistent with Street and Bryant (2000) and Al-Shammari (2005), it is not consistent with political process theory arguments (Ball and Foster, 1982; Watts and Zimmerman, 1986). Finally, the legal form variable (H9) is found to be insignificant in relation to the overall level of compliance. Thus hypotheses H1, H2, H4, H7 and H9 are not supported in terms of overall compliance.
6 Discussion of findings The current study shows a change compared with prior disclosure research conducted in the first year of the reactivation of the CASE. In her earlier study, Abd-Elsalam (1999, p.262) concluded that: ‘positive accounting theories are too simplistic as they do not offer strong explanation for varying disclosure levels in Egyptian listed companies in 1995, associated with low R2 varying between 21%–46%’. In the current study the R2 associated with the firmspecific factors ranges between 48% and 72% across the various partial disclosure indices and is 71% for overall disclosure. The range is between 47% and 67% across the various partial measurement indices and 69% for overall measurement/presentation. This indicates that positive accounting theories, especially agency theory provide strong evidence for explaining differing levels of compliance with EASs. Firm specific characteristics have most explanatory power in relation to the MCND index and least explanatory power in relation to the MNEW index. That is compliance is generally highest where the EASs have introduced no change to preexisting Egyptian regulation and is generally lowest when a completely new regulation is implemented. This suggests that de facto compliance may involve other determinants beyond the typical micro determinants considered by agency, political process, signalling and capital need theories. Such other determinants might include education/training, since these findings are consistent with Cairns’ (2001) argument that the adoption of new financial reporting requirements brings additional problems, one of the most significant being the education and training of accountants.
Enforcement may be another determinant, since weak enforcement mechanisms permit a gap to develop between official accounting standards and actual practices (Lin and Liyan, 2001; Chen et al., 2002; Taplin et al., 2002). It is possible that the lowest levels of compliance associated with changed or completely new regulation may be related to weak enforcement of EASs. That is, the mechanisms used by the CMA to enforce and monitor corporate compliance may have been inadequate. Some support for this proposition is evident by the introduction of new listing rules effective from August 1, 2002 that aim to enhance administrative action against issuers that do not comply with EASs. It will be interesting to investigate in future whether these do increase the levels of compliance. The findings indicate that de facto compliance tends to be very low unless the company is connected with a large international audit firm or is publicly owned. Being audited by one of the large audit firms (as a proxy for high professionalism) is the dominant factor in this study contrasting with the earlier findings of Abd-Elsalam (1999) in Egypt. However, the current study is consistent with previous literature in other countries which has drawn on agency and signalling theories to show that companies which are audited by big four audit firms comply more with IASs and is consistent with Chen et al. (2002) who reported that lack of quality audit is an important contributing factor for non-compliance with IASs. Due to problems associated with the audit profession (Samaha and Stapleton, 2008), it was expected that international audit firms operating in Egypt would be more familiar with IASs leading to an expectation that Egyptian companies audited by one of the international firms will comply more closely with the EASs, as these are based on IASs. The findings in relation to ownership concentration are also consistent with previous literature which has drawn on agency theory to show that companies with wider ownership diffusion comply more with IASs. Compliance with EASs and nonmandated but applicable IASs by companies with medium ownership concentration may be aimed at reducing agency costs. Gray’s (1988) secrecy hypothesis also argues that where a firm’s shares are closely held, there is a preference for confidentiality so that disclosure is restricted to those who are
closely involved with the management and financing of the firm. The results show that five variables did not have the explanatory power expected and it is interesting to reflect on possible explanations. First, the general lack of influence of industry sector on compliance may be related to the fact that most listed Egyptian companies are exempted from taxes. Given that taxes are an important measure of political visibility (Salamon and Dhaliwal, 1980), such exemptions may imply there is no need for extensive compliance to reduce political costs, so that all industry sectors have reacted similarly. Second, although, previous research by Abd-Elsalam (1999) provided evidence of a positive association between compliance and public sector status, the current research did not find such an association nor was there support for a legal form effect. This variation in the results may be because market conditions have changed between the conduct of the studies. In 1995, the Egyptian stock market was revitalised by a process of deregulation and this played an important role in its development as a channel for divesting state-owned enterprises through public stock offerings. However, from 1998 a number of new large private-sector offerings presented more investment opportunities and stimulated market sophistication as investors began to focus on private sector ‘growth’ stock investment as against public-sector ‘value’ stock trading (Abdel Shahid, 2003). This might have encouraged greater private sector compliance removing the positive association between compliance and public sector status. Third, the role of the banks in Egypt may explain the results regarding the gearing variable. During 2000, corporate debt generally took the form of bank loans. Given that banks usually have direct access to information for monitoring purposes, such gearing need not imply a greater need for extensive compliance in published financial statements. Finally, the non-significant results for the profitability variable (as a proxy of signalling theory), may be due to fact that on average the companies in the sample were not financially sound for the study period as discussed in Section 4.
7 Conclusions In this paper we provide evidence about the firm specific determinants of compliance
with the 22 EASs issued in 1997 that are in material terms Arabic translations of IASs. Overall, the findings of this research are strongly consistent with the predictions of agency theory (as measured by the type of auditor, ownership concentration and market capitalisation). Although, the findings are consistent with the predictions of capital need theory (as measured by internationality and share trading), political process theory (as measured by size and internationality) and signalling theory (as measured by type of auditor and market capitalisation), the effect of these three theories is not clear. The analysis shows that lack of profitability, liquidity and leverage of the firm do not seem to affect the extent of its compliance with EASs, which limits the support for agency and signalling theories. Since compliance is not related to manufacturing and non-manufacturing sectors this limits the support for political process theory and as the performance of private and public sector firms is similar, this limits the support for capital need theory. This may indirectly suggest that the positive accounting perspective may not be entirely applicable for de facto compliance in Egypt. Hence, it cannot be entirely inferred that listed Egyptian companies use de facto compliance with standards that are close to IASs, as a means of raising capital, reducing political costs or signalling to the market that they are high quality firms. It is important to recognize that this research must be interpreted in light of some limitations. For example, there is a weakness in that we have depended on ownership concentration to act as a proxy for the existence of corporate governance mechanisms. A finer classification could have been used taking into consideration the board composition and existence of audit committees. However, in Egypt, audit committees are non-existent (Carana, 2000; World Bank, 2001) and information on board composition is neither described in the companies law nor in capital market law and is not required to be disclosed in the annual report as explained earlier (Carana, 2000; World Bank, 2001). But notwithstanding these limitations, the results are sufficiently robust to consider their implications for Egyptian policy makers and to warrant an extension to other emerging economies. Before considering some policy implications of this study it is worth noting that highly concentrated ownership companies listed on CASE have little
motivation to comply with EASs because the demand for public information is relatively weak in comparison with that of companies with wider share ownership. This structural feature of the Egyptian stock market provides a countervailing force to the growing pressures for internationalisation and global transparency. This is an important issue that needs to be factored into investor decisions and accounting standards setting policy at both the national and international levels. The policy implications suggested by this study should impact in particular the role of the auditor in enforcing compliance with standards, the education and training of local firms to undertake that role and because of the impact of ownership concentration, corporate governance arrangements may be significant. While further research is needed, such strategies might include: 1- Development of the capacity of the ESAA to establish quality control and quality assurance arrangements, in addition to taking measures to ensure compliance with international best practices (e.g., IFAC code on auditor’s professional ethics and independence). This should include implementing a peer review program for monitoring quality assurance arrangements especially in non-big four audit firms to ensure that auditors comply with applicable auditing and ethical standards and independence requirements. 2- Improvement of the arrangements for auditors’ education, training and examination in line with IFAC guidelines. The financial statements of listed enterprises to be audited by either a big four firm or a registered local auditor, who has acquired appropriate higher education, obtained relevant practical experience subject to verification and passed the ESAA examination on EASs and International Standards on Auditing (ISAs). Consideration might also be given to the licensing arrangements for registered auditors. 3-Improvement of corporate governance practices by introducing mandatory requirements based on the OECD principles on corporate governance, as well as enforcing these requirements. CASE should be one of the main driving forces towards establishing a ‘corporate governance code’ in Egypt together with other professional bodies such as the ESAA.
Table 1: Stratified sample by industry sector, share trading and legal form
Ac vel y traded (79)
Manufacturing Private sector
Trade and Services Private sector
Rarely or not traded (221): A. All 22 public sector (but 2 excluded)
14 (2 excluded) -------
B. Remaining (199) - Manufacturing 59% (12 excluded)
117 (12 excluded) 105
- Trade and services 41% (5 excluded)
82 (5 excluded) 77
TOTAL Usable Sample 281
Table 2: Descriptive statistics of the continuous explanatory variables Code in SPSS
Range Sta s c
Min. Sta s c
Max. Sta s c
Sta s c St
Mean d. Err or St
Std. as c
Table 3: Descriptive statistics of the categorical explanatory variables Code in SPSS
Std. Devia on
Table 4: Compliance level with Overall IASs indices for 281 listed Egyptian companies Overall indices DIS 1 DIS 2 MEAS/PRES 1 MEAS/PRES 2
.79 .74 .79 .76
.11 .16 .18 .17
.90 .90 .97 .93
Mean Sta s c
Standard devia on
.46 .50 .55 .56
.01 .01 .01 .01
.20 .17 .18 .17
.41 .45 .51 .52
Table 5: Compliance level with mandatory IASs Partial indices (MCND, MCD and MNEW) for 281 listed Egyptian companies
Mandatory Par al Indices
IASs MCND IASs MCD IASs MNEW
.73 .88 1.00
.24 .00 .00
IASs MCND IASs MCD IASs MNEW
.68 1.00 1.00
.32 .00 .00
Sta s c
Mean Std. Error
Disclosure .97 .60 .88 .23 1.00 .47 Measurement/presenta on 1.00 .79 1.00 .30 1.00 .40
Standard devia on
.01 .01 .02
.17 .21 .27
.56 .17 .50
.01 .02 .02
.15 .27 .27
.80 .25 .38
Table 6: Compliance level with voluntary IASs Partial indices (NMCD and NMNC) for 281 listed Egyptian companies Mean Voluntary Par al Indices
IASs NMCD IASs NMNC
IASs NMCD IASs NMNC
Standard devia on
Sta s c Disclosure .19 .25
Measurement/presenta on 1.00 .26 1.00
Table 7: Regression results: Overall OLS regression models
Overall Disclosure index (DIS2) t-test Es m a t ed β R2 (%) Value coeﬃ cient
Independent Variables Constant ROE2 (H1) LIQUID (H2) ASSETS (H3) SALES (H3) MCAP (H3) GEARING (H4) AUDIT (H6) DIFFUSE (H5) INDUSTRY (H7) SHARE (H8) LEGAL (H9) INTER (H10)
-.650 x x x x .226 x .846 .349 x .344 x .405
Model Assump ons Largest VIF value Lowest Tolerance value Highest condi on index
-12.995*** x x x x 5.582*** x 10.628*** 3.955*** x 3.780*** x 5.309*** 1.804 .554
x x x x 15 x 45 6 x 2 x 3
Overall MEAS/PRES index (MEAS/PRES 2) t-test Es m a t ed β R2 (%) Value coeﬃ cient -.647 x x x x .227 x .765 .448 x .355 x .362
-12.768*** x x x x 5.542*** x 9.485*** 5.013*** x 3.852*** x 4.680*** 1.804 .554
Probability of F
Overall adjusted R
The symbol (X) indicates that the variable did not enter the equa on
*** Signiﬁcant at the level les s than 0. 001;
x x x x 8 x 41 16 x 2 x 2
Table 8: Regression results: Partial Disclosure OLS regression models (4 models) MCND par al index
Constant ROE2 (H1) LIQUID (H2) ASSETS (H3) SALES (H3) MCAP (H3) GEARING (H4) DIFFUSE (H5) AUDIT (H6) INDUSTRY (H7) SHARE (H8) LEGAL (H9) INTER (H10) Model Assump ons Largest VIF value Lowest Tolerance value Highest condi on index F Probability of F 2
MCD par al index Es m a t ed β coeﬃ cient
-.469 x x
-8.857*** x x
292 x x
6.909*** x x
38 x x
Es m a t ed β coeﬃ cient
-.643 x x x
-13.171*** x x x
x x x
.593 x x
6.820*** x x
MNEW par al index Es m a te t-test dβ Value coeﬃ cien t -.443
NMCD par al index 2
x x 3.589***
x x 2
x x x
x x x
x x x
15 x x
* Signiﬁcant at the level les s than 0. 05 **Signiﬁcant at the level les s than 0. 01 *** Signiﬁcant at the level les s than 0. 001
The symbol (X) indicates that the variable did not enter the equa on Table 9: Regression results: Partial Disclosure binary logistic models (one model)
NMNC partial index
Es m a t ed β coeﬃ cient
Wald test Value
ROE2 (H1) LIQUID (H2) ASSETS (H3) SALES (H3) MCAP (H3) GEARING (H4) DIFFUSE (H5) AUDIT (H6) INDUSTRY (H7) SHARE (H8) LEGAL (H9) INTER (H10)
x x x x x x 1.668 2.823 x x x 1.479
x x x x x x 16.221*** 46.568*** x x x 12.592***
Chi-square goodness of ﬁt Model P Overall R2 Index classiﬁca on%
x x .170
Es m a t ed β coeﬃ cient
155.928 .0001*** 63 86
*** Signiﬁcant at the level les s than 0. 001 The symbol (X) indicates that the variable did not enter the equa on.
x x x x x x 4 50 x x x 9