COMPANY ATTRIBUTES AND EXTENT OF INTANGIBLE ASSETS DISCLOSURE

Transcription

1 CHAPTER VI COMPANY ATTRIBUTES AND EXTENT OF INTANGIBLE ASSETS DISCLOSURE The results of chapter 5 reveal that the extent of disclosure of intangible assets varies from company to company. It could be due to the influence of different corporate specific attributes. This chapter analyses the association between the extent of intangible assets disclosure and corporate attributes. The corporate attributes studied are size of a firm, its profitability, leverage, listing category, nature of industry, foreign activity and audit firm size. These explanatory variables have been divided into three groups, consistent with Lang and Lundholm (1993), Wallace et al. (1994) and Oliveira et al (2006). These groups are: i) structural variables, including firm size, leverage, and audit firm size; ii) iii) performance variables, including profitability; and market variables, comprising industry, listing status and foreign activity 6.1. HYPOTHESES DEVELOPMENT Based on the theoretical considerations and empirical research, several hypotheses have been developed that relate company-specific characteristics to the disclosure of intangible assets by companies in India. All hypotheses have been stated in both null and alternate form Size of a firm Firm size is perhaps the most consistent corporate specific characteristic which has been found to be associated with the level of intangible assets disclosure. Bozzolan et al (2003, 2006), Meca and Martinez (2005), Oliveira et al (2006), Guthrie et al (2006), Carlin et al (2006) and White et al (2007) found positive relation between size of a firm and extent of intangible assets disclosure. However, Kang and Gray (2006) and Kamath 116

2 (2008a) observed no correlation between intangible assets reporting in the annual reports and size of the companies. Larger companies may be hypothesised to disclose more intangible assets related information in the annual report than smaller companies for a variety of reasons. Firstly, the cost of disseminating and accumulating detailed information may be relatively low for the larger corporations than the smaller corporations (Cerf, 1961; Singhvi & Desai, 1971; Buzby, 1975 and Firth, 1979). The big companies have the resources and expertise to produce more information in their company s annual reports (Ahmed & Nicholls, 1994 and Hossain & Adams, 1994). Hence, little extra cost may be incurred to increase disclosure, i.e. lower incremental cost of producing information for large firms (Lang and Lundholm, 1993). In addition, larger corporations might need to collect more intangible assets information for their internal management system. Secondly, larger companies undertake more activities, and usually have different business units which may have different critical success factors and different long term value-creating potential (Hackstone and Milne, 1996). This means that more information needs to be disclosed to provide stakeholders with a complete picture of the company. Thirdly, larger firms tend to go to the stock market for financing more often than smaller firms and as a result may disclose more information in their annual reports for their own interest (i.e. capital need hypothesis; Cerf, 1961). Fourthly, smaller firms may feel that their intangible assets disclosure activities could endanger their competitive oppositions with respect to other larger firms in their industry, i.e. reluctance of small firm to inform their competitors (Singhvi & Desai, 1971; Buzby, 1975; Raffournier, 1995). As a result, smaller companies may tend to disclose less information concerning its intangible assets than large companies. Fifthly, Wallace and Naser (1995) state the impact of large size companies on the economy. These can be considerable as these companies account for a significant proportion of goods and services produced, consumption of raw materials and number of people employed. The large companies are likely to come under the scrutiny of various interested parties and therefore tend to disclose adequate and detailed intangible assets related information in their annual reports. 117

3 Sixthly, large corporations are likely to have a higher level of internal reporting to keep senior management informed and therefore likely to have relevant information available (Cerf, 1961; Buzby, 1975 and Owusu-Ansah, 1998; Depoers, 2000). Seventhly, it has been suggested that the annual reports of large corporations are more likely to be scrutinised by financial analysts than those of smaller firms and investors may interpret non-disclosure as bad news which could adversely affect firm value. Thus, larger firms may have an incentive to disclose more information related to its intangible assets than smaller firms. Finally, as far as political costs are concerned, bigger companies are more vigil in the public eye which will tend to make them exhibit greater disclosure than the smaller firms (Firth, 1979; Holthausen and Leftwitch, 1983; Watts and Zimmerman, 1990). Agency theory shows that a company has to satisfy the needs of creditors and investors. Therefore, it should provide detailed information in annual reports to avoid information asymmetries (Jensen & Meckling 1976). Thus, bigger firms make more disclosure to minimise agency and political costs. The priori expectation is that larger companies will have higher levels of intellectual capital reporting than smaller companies. Several measures of size are available in the literature. The total sales, total assets and market capitalisation are used as surrogates of size of a firm in this study. The foregoing discussions lead to development of the following null and alternate hypotheses: H 01 : H 1 : The size of a company as measured by its total assets or total sales or total market capitalization has no significant impact on its intangible assets disclosure score. The size of a company as measured by its total assets or total sales or total market capitalization has a significant impact on its intangible assets disclosure score Leverage of a firm Several studies investigated the relationship between leverage (book value of debt to shareholder s equity) and intangible assets disclosure. It is viewed that companies with a higher level of leverage have high level of financial risk (Patton and Zelenka, 1997) and thus disclose more information. Agency theory can also be used to explain the influence of leverage on the level of voluntary disclosure by a firm. When firms incur outside debt, 118

4 agency costs arise from conflicts of interest between equity and debt investors (Berger & Bonaccorsi di Patti, 2006). These agency costs are comprised of a reduction in value of the firm and increased monitoring costs owing to the fact that the manager will try and reallocate the wealth of the debt-holder to the firm. The greater the level of debt, the wider apart the two parties interests (Jensen & Meckling, 1976) and hence the need for higher monitoring costs. Increased disclosure by a firm can reduce these monitoring costs. Thus, firms with high leverage levels have an incentive to make voluntary disclosures in order to reduce agency costs. Media agenda-setting theory (Cohen, 1963) has also been used to explain why a firm s leverage level could be associated with its intangible assets disclosure levels. If a firm has good news, low levels of debt for example, then it will attempt to highlight this positive information to the market through making voluntary disclosures in its annual report (Sujan & Abeysekera, 2007). Media agenda setting theory therefore suggests the opposite of agency theory. If a company has low leverage levels it will increase its voluntary disclosures to inform the market of its strong position. Prior studies have provided mixed results on the association between leverage and intangible assets disclosure levels. Oliveira et al (2006), Bozzolan el al (2006) and Woodrock & Whiting s (2009) showed no association between the extent of voluntary intangible assets disclosure and leverage, whereas Meca & Martinez (2005) and White et al. (2007) observed a significant positive relationship between the two. The following null and alternate hypotheses are tested in this study: H 02 : H 2 : The leverage of a company as measured by its debt-equity ratio has no significant impact on its intangible assets disclosure score. The leverage of a company as measured by its debt-equity ratio has a significant impact on its intangible assets disclosure score Audit Firm Size Many authors have suggested that auditors play a role in defining the disclosure policy of their clients (Raffournier, 1995). Large audit firms are widely scattered across the world while small audit firms operate domestically. The classification of audit firms into two groups has been drawn on the assumption that the large firms have more concern 119

5 for their reputation and therefore are more willing to associate with firms that disclose more information in their published financial reports. On the other hand, small audit firms do not possess the power to influence the disclosure practices of their clients. Rather they attempt to meet the needs of their clients in order to retain them (Firth, 1979 and Wallace and Naser, 1995). Oliveira et al. (2006) argued that large auditing firms may encourage their clients to disclose more information as they want to preserve their reputation, develop their expertise, and ensure that they retain their clients. Chalmers and Godfrey (2004) argue that high profile auditing companies are more likely to demand high levels of disclosure to maintain their reputation and to avoid reputation costs. Dumontier and Raffournier (1998) observe that the auditors want their clients to comply with complex accounting standards for the sake of their reputation and in their own interest. They may use the information disclosed by their clients as a signal about their own quality (Inchausti, 1997). The other reason advocated by Patton and Zelenka (1997) is that the audit by big-six-audit firm is believed to bring enhanced credibility to the financial reports. This is also linked to the fact that major international auditing companies have more knowledge about International Accounting Standards (IAS) and so the costs of implementing and auditing them to their clients is lower than for smaller auditing companies. This hypothesis is based on the argument that companies audited by big-six audit firms have substantial agency costs. They try to reduce them by contracting with these auditing firms. Auditing is argued to be a way of reducing agency costs (Jensen & Meckling, 1976 and Watts & Zimmerman, 1983) and the companies that have high agency costs tend to contract high quality auditing companies. Thus, there is a positive relationship between agency costs and the disclosure decision (Inchausti, 1997). Oliveira et al. (2006) and Woodrock & Whiting s (2009) are the sole studies to examine the relationship between the audit firm size and the extent of its intangible assets disclosure. They concluded that companies with a Big Four auditor disclose more 120

6 intangible asset information as compared to companies with non-big four auditors. The following null and alternate hypotheses have been formulated and tested in this study: H 03 : H 3 : The audit firm size of a company has no significant impact on its intangible assets disclosure score. The audit firm size of a company has a significant impact on its intangible assets disclosure score. This has been used as a dummy variable. The companies being audited by Big-six audit firms (Price Waterhouse; A.F. Ferguson; S.B. Baltiboi; Delloitte; Haskins and Sells and B.S.R. & Co.) were assigned a score of 1 and others Profitability of a Firm A number of researchers have used profitability as an explanatory variable for differences in the disclosure levels. A company s profit comes from its continued investment in intangible assets. The companies which have more investment in intangible assets will tend to disclose more information about them. Therefore the more an enterprise can make profits, the more it will tend to disclose the information about intangible assets. The higher the performance of companies is, the more it will discloses intellectual capital to the public (Li et al, 2008). Singhvi and Desai (1971) pointed out that a company with high profits will disclose more information about intellectual capital. Meca and Martinez (2005) verified a positive relationship between intellectual capital disclosure and the corporate profitability. Companies having higher profitability may disclose more information relating its intangible assets in their corporate annual reports than the companies with lower profitability (or losses) for a number of reasons. Firstly, if the profitability of a company is high, management may disclose more detailed information in their corporate annual reports in order to experience the comfort of communicating it as good news. On the other hand, if profitability is low, management may disclose less information in order to cover up the reasons for losses or lower profits (Singhvi and Desai, 1971). It can be linked to theoretical framework of reporting practices also. Agency theory suggests that 121

7 managers of very profitable firms will use external information in order to obtain personal advantages. Therefore, they will disclose detailed information in order to support the continuance of their positions and compensation arrangements (Inchausti, 1997). Signalling theory suggests that profitable firms are more likely to disclose more information to the market to differentiate themselves from poorer performers. Wallace and Naser (1995) argue that a profitable company is more likely to signal its good performance to the market by disclosing more information in its annual report. Political process theory argues that firms with huge profits will be interested in disclosing more information in order to justify the level of profits. However, the empirical studies found mixed results. Researchers have used a number of measures to determine the associations between profitability and disclosure levels. The return on assets, returns on sales, return on net worth and return on capital employed have been used as a determinant of profitability and its association with the level of disclosure in the present study. The following specific null and alternate hypotheses have been formulated and tested. H 04 : H 4 : The profitability of a company as measured by its ROA or ROS or ROCE or RONW has no significant impact on its intangible assets disclosure score. The profitability of a company as measured by its ROA or ROS or ROCE or RONW has a significant impact on its intangible assets disclosure score Listing Category of a Company The listing category of a firm also influences the disclosure level of a firm. The relationship between a company s listing category and disclosure practices is based on agency cost and the signalling arguments (Lopes and Rodrigues, 2007). Every Indian company listed on a stock exchange has to comply with its listing agreement. The companies whose shares are actively traded have always been scrutinised sharply by the 122

8 market as a whole and investors in particular. Thus, to reduce agency, political and monitoring costs more disclosure is being demanded from A category firms than others. The other reasons being that the international investors have shown their interest in the Indian capital market after The companies listed under A category are assumed to be audited by big-six audit firms, have overseas operations and might have listing on foreign stock exchange and are actively traded on stock exchanges. Such companies are believed to be disclosing better and detailed information. This argument is even stronger if the listed company wants to raise its capital in foreign markets (capitalneed hypothesis, Cooke, 1989). Oliveira et al s (2006) study showed a significant influence of listing category on the level of intangible assets disclosure. So, the above discussion has lead to the formulation and testing of the following null and alternate hypotheses: H 05 : H 5 : The listing category of a firm has no significant impact on its intangible assets disclosure score. The listing category of a firm has a significant impact on its intangible assets disclosure score. The impact of listing category of a firm has also been examined by introducing dummy variable. The score of 1 has been given for companies falling under A listing category and 0 otherwise Foreign Activity Another possible motive for increased disclosures on intangible assets is the degree of foreign activity of the firm. With the increase in the level of foreign trade, a company needs to increase their disclosures relating to intangible assets to prove its credit worthiness to its foreign clients. Managers of companies operating in several geographical areas have to control a greater amount of information, due to the higher complexity of the firm s operations (Cooke, 1989b). They are prone to increase their disclosure to show their international presence to stakeholders as a perceived good signal (Cooke, 1989b, Raffournier, 1995 and Depoers, 2000).. Accordingly, the following null and alternate hypotheses have been framed: 123

9 H 06 : H 6 : The extent of foreign activity of a company as measured by its total exports to total sales ratio has no significant impact on its intangible assets disclosure score. The extent of foreign activity of a company as measured by its total exports to total sales ratio has a significant impact on its intangible assets disclosure score Nature of Industry The nature of industry and the degree of competition within an industry to which a company belongs may also influence its intangible assets disclosure level. This can be explained by stakeholder theory, signalling theory, legitimacy theory, proprietary theory and media agenda-setting theory. Stakeholder theory purports that shareholders have a right to be provided with information about how the organisation s activities affect them (Vergauwen & Alem, 2005), particularly if they are less powerful shareholders who cannot access information through private meetings (Holland, 2001). So in order to satisfy the stakeholders need for information, firms are forced to make voluntary disclosures about their intangible assets. Signalling theory suggests that, within an industry, any deviation from established corporate reporting practice may be perceived by the market as bad news (Giner, 1997). Additionally, mandatory financial reporting is claimed to be less informative in high technology industries which make larger investments in intangibles (such as R&D, human capital and brand development) (Collins et al., 1997; Francis and Schipper, 1999; Lev and Zarowin, 1999). Legitimacy theory asserts that organisations, as part of a social contract, will take action to ensure that their activities are perceived as legitimate (Lindblom, 1994; Wilmshurst & Frost, 2000). Firms with high levels of intangible assets are more likely to engage in voluntary intangible assets disclosure because they cannot legitimise their status through the traditional symbols of corporate success, the tangible hard assets (Guthrie et al., 2004). They need to communicate how the firm uses its intangible assets to generate value (Sciulli et al., 2002). 124

10 Proprietary costs also vary according to industry. Different industries have different characteristics relative to market competition, the type of private information, and the threat of entry of new firms into the market. These factors provide incentives for companies belonging to the same industry to disclose more information than firms in another industry. The fifth theory that has been used to explain the level of voluntary intangible assets disclosure is that of media agenda-setting theory. This theory suggests that firms (as a form of media) set the agenda for public opinion by emphasizing or highlighting certain issues. Therefore intangible assets intensive firms boast of their intangible assets to signal their superiority over competitors (Sujan & Abeysekera, 2007). Most studies have supported the hypothesis that firms that are high in intangible assets (e.g. high-tech., knowledge intensive industries) are more likely to engage in voluntary intangible assets disclosures (e.g. Bozzolan et al., 2003; Petty & Cuganesan, 2005; Bozzolan et al., 2006; Oliveira et al., 2006; Sujan & Abeysekera, 2007; Woodrock & Whiting, 2009) and some studies have chosen to base their research solely on samples of firms from high intellectual capital industries (e.g. Sonnier et al., 2007; White et al., 2007). Accordingly, the null and alternate hypotheses of the study are H 07 : H 7 : The nature of industry to which a company belongs has no significant impact on its intangible assets disclosure score. The nature of industry to which a company belongs has a significant impact on its intangible assets disclosure score RESULTS AND DISCUSSIONS The effect of different corporate attributes on the extent of intangible assets disclosure has been studied with the help of two-factor ANOVA test, Univariate and backward step-wise regression analysis. The results of these have been discussed in the following sub-parts: Two-factor ANOVA 125

11 Intangible assets disclosure trends have been analysed for each of the corporate attributes (size of the firm- total sales, total assets and market capitalization, leverage, audit firm size, profitability- ROS, ROA, ROCE and RONW, listing status and foreign activity) for both years of the study. Each attribute (except dummy variables) was divided into three broad categories. First category, High consists of disclosure scores of all those companies which fall above th percentile values for each respective attribute. Medium second category consists of disclosure scores of all those companies which fall between to th percentile for each respective attribute. Third category Low comprises disclosure score of companies which fall below rd percentile for each respective attribute. Table 6.1 shows mean and standard deviation values for categories of all the above corporate attributes for both years of the study. Also, in order to examine any significant difference in the mean disclosure score of each category (i.e. high, medium and low) over two years of study, two factor ANOVA test was conducted. In table 6.1, F- values across years highlights significant difference in disclosure score over both years of the study; F-values across categories illustrates any significant difference in intangible assets disclosure score across different categories; and F-values Year and category wise shows differences in disclosure scores when both years and categories are together taken into consideration. The ANOVA test pre-condition of equal variances was satisfied by conducting Levene s test of homogeneity of variance. The Levene s test s significance for all corporate attributes was greater than 0.05, which satisfied the above ANOVA test prerequisite. 126

12 Table 6.1 Descriptives of corporate attributes F-Values Mean SD Mean SD Sales High Medium Low Total Assets High Medium Low Market Capitalisation High Medium Low Return on Asset High Medium Low Return on capital employed High Medium Low Return on Sales High Medium Low Yearwise Categorywise Year and category wise * * * * * * * * * 4.585** * 10.87* Cont 127

13 Return on net worth High Medium Low Debt-equity ratio High Medium Low Foreign Activity High Medium Low Listing status A category Others Audit firm size Audited by Big Others * 3.631** * 6.249* 2.506*** * 8.214* * * * * Note: *, **, *** significant at 1%, 5%, 10% respectively. 128

14 ANOVA tests result reveals significant difference in disclosure score over both years of the study. F- values are significant at 1% level of significance for all the corporate attributes (sales, total assets, market capitalization, ROS, ROCE. RONW, ROA, leverage, foreign activity, listing status and audit firm size). ANOVA test results in table 6.1 also demonstrate significant differences among mean disclosure score of each category of corporate attributes. To find out which category differs significantly from others Fisher s LSD (Least significant difference) post-hoc test was also conducted. Table 6.2 displays its results. The Fisher LSD post-hoc test results on size of a firm measured by total sales attribute reveal significant difference (at 1% level) between category high and category medium & low. For size of the firm measured by total asset and market capitalization attributes significant difference (at 1% level) was observed among all the three categories (high, medium and low). This means firms with high sales, high total assets and high market capitalization are reporting more on their intangible assets than firms with medium and low sales, total assets and market capitalization. In case of corporate attribute of profitability measured by ROS, ROA, ROCE and RONW significant difference (1% level) was observed in mean disclosure score of category high as compared to category medium and low. This reveals that companies with high profitability disclose more about their intangible assets in their annual reports than firms with medium and low profitability. In case of corporate attribute of leverage, negative relationship can be seen between high and medium & low categories. The negative association is significant (at 1% level) in case of high and low category. This means companies with high leverage disclose less about their intangible assets. For corporate attribute of foreign activity, significant positive relationship was observed among category high and medium with category low. This reveals that companies with high level of foreign activity disclose more about their intangible assets as compared to companies with low foreign activity. 129

15 Sales Total Assets Market Capitalisation ROA ROCE ROS Table 6.2 LSD Post-hoc test results Category Category Mean Difference High Medium Low High Medium Low High Medium Low High Medium Low High Medium Low High Medium Low 130 Medium 4.13* Low 4.82* High -4.13* Low 0.69 High -4.82* Medium Medium 2.33* Low 3.99* High -2.33* Low 1.65* High -3.99* Medium -1.65* Medium 4.01* Low 7.05* High -4.01* Low 3.04* High -7.05* Medium -3.04* Medium 2.05* Low 3.40* High -2.05* Low 1.35 High -3.40* Medium Medium 1.47* Low 2.21* High -1.47* Low 0.74 High -2.21* Medium Medium 2.04* Low 3.40* High -2.04* Low 1.36 High -3.40* Medium Cont

16 RONW Leverage Foreign Activity Category Category Mean Difference High Medium Low High Medium Low High Medium Low Note: *, **, significant at 1%, 5%, respectively. Medium 0.9 Low 1.99* High -0.9 Low 1.09 High -1.99* Medium Medium -1.3 Low -2.60* High 1.3 Low -1.3 High 2.60* Medium 1.3 Medium 0.69 Low 3.00* High Low 2.31* High -3.00* Medium -2.31* 131

17 6.2.2 Regression Analysis Further, the detailed effect of different corporate attributes on the extent of intangible assets disclosure has been studied with the help of regression analysis. For regression analysis, it is essential to assess the validity of the Model (regression equation). Assessing the Validity of the Model (regression equation) is necessary to find out the impact of various corporate-specific attributes on the disclosure practices of sample Indian private sector companies for the financial years and respectively. It is pre-emptive to check the existence of multicollinearity among the explanatory independent variables before proceeding to the results of regression analysis. Multicollinearity or collinearity is the situation where two or more independent variables are highly correlated and can have damaging effects on the results of multiple regression. The correlation matrix is a powerful tool for getting a rough idea of the relationship between predictors. The suggested rule of thumb is that, if the pair-wise or zero-order correlation coefficient between two regressors is high say, in excess of 0.8, then multicollinearity is a serious problem (Gujarati, 2006, p.359). The solution to this multicollinearity drawback is to drop that variable and thereafter run regression analysis with rest of the variables. Another way to check the multicollinearity is to compute the average VIF (Variance inflation factor). As a rule of thumb, if the average VIF of a variable exceeds 1 which will happen if correlation coefficient exceeds 0.80, then that variable is said to be highly collinear (Gujarati, 2006, p. 362) Checking the Normality of Data It was essential to check the distribution of independent and dependent variables. Certain descriptives like skewness and kurtosis were calculated to test the normality of data. The disclosure score is found to be within range of skewness (-1 to 1) and kurtosis (-2 to 2) but the size of a firm measured in all explanatory variables is found to be not normal. Thus, natural log transformation has been done to make it normal (see, for e.g. Raffournier, 1995; Patton & Zelenka, 1997and Owusu-Ansah, 1998). 132

18 Correlation Analysis The Pearson product moment correlation (r) was computed to examine the correlation between the dependent and independent variables and with the dependent variables. A correlation matrix of all the values of r for the explanatory variables along with dependent variables was constructed and is shown in table 6.1 and table 6.2 for the years and respectively. Table 6.3 illustrates that multicollinearity is an issue of concern in the case of measures of profitability of a firm i.e. between RONW and ROCE. RONW and ROCE have coefficient of correlation (.856), above the rule of thumb (0.80), significant at 1% level. Thus, the problem of multicollinearity exists only between these measures of profitability for the financial year Rest of the variables do not have problem of multicollinearity among them. The results of table 6.4 bring to light that the problem of high value of correlation coefficient appears in the case of measures of profitability for the year It was found in the case of ROS and ROA (.935) and RONW and ROCE (.794) significant at 1% level. Thus the problem of multicollinearity exists among only these explanatory variables. The problem was checked by taking only one measure of profitability in a regression equation at a time. Rests of the variables do not have any problem of multicollinearity among them for the year

19 Table 6.3 Pearson s Product Moment Correlation Matrices ( ) Disclosure Score Log Sales Log Assets Log Mar cap ROA ROCE ROS RONW Leverage Foreign activity Disclosure Score 1 Log of Sales.383 * 1 Log of Assets.321 *.780 * 1 Log of Market Capitalisation.601 *.683 *.717 * 1 ROA ROCE.231 *.134 ** * ROS * RONW.194 *.145 ** * * Leverage * *.291 * Foreign Activity.171 * ** Note: *, **, significant at 1%, 5%, respectively. Table 6.4 Pearson s Product Moment Correlation Matrices ( ) Disclosure Score Log Sales Log Assets Log Mar cap ROA ROCE ROS RONW Leverage Foreign activity Disclosure Score Log of Sales.361 * 1 1 Log of Assets.313 *.785 * 1 Log of Market Capitalisation.459 *.611 *.735 * 1 ROA * ** 1 ROCE * * ROS * **.935 * RONW * * * Leverage *.135 ** * ** * 1 Foreign.137 Activity ** Note: *, **, significant at 1%, 5%, respectively. 134

20 Regression analysis has been performed in two parts. The first part deals with the application of Univariate regression analysis and afterwards backward step-wise regression analysis was run Univariate Regression Analysis To see the individual impact of various independent variables on the dependent variable the Univariate regression analysis is performed. This means if any change occurs in the dependent variable due to change in the proportion of different independent variables, then the researcher needs to analyse the variations in the dependent variable by the various independent variables individually. This helps in drawing better conclusions about the influence of different explanatory variables on the dependent variable. Table 6.5 and table 6.6 explain the results of Univariate regression analysis for the two years of the study. The results of table 6.5 and 6.6 reveal that the size of a firm as measured in the terms of three explanatory variables namely sales, or assets, or market capitalization is found to be significant at 1% level. All have positive association with the disclosure of intangible assets by selected companies in India. It implies hereby that the large size companies have higher disclosure of intangible assets. Such results are well evidenced by literature. The values of adjusted R 2 in case of sales (.143), and assets (.099) are far less than that in the case of market capitalization (.359) for Similarly the values of adjusted R 2 in case of sales (.126) and assets (.094) are also considerably lower than that in the case of market capitalization (.208) for the year The F values ( , ) are also significant at 1% level of significance. Thus, size of a company as measured by market capitalization explains better variations in disclosure of intangible assets by the companies in India for the two years of the study. 135

21 Table 6.5 Regression Analysis (Univariate Analysis ) Independent Adjusted Variable Constant Beta t-value R 2 R 2 Log of Sales * * Log of Assets * * Log of Market Cap * * ROS ROA ROCE * * RONW ** ** Leverage Foreign Activity * * Listing Status * * Auditing Firm * * ** ** *** *** *** *** Industry Note: *,**,*** significant at 1%, 5%, 10% respectively. F 1 2 1=Agri input and tobacco; 2=automotive; 3=Banking & financial services; 4=capital goods, industrial and engg products; 5=consumer goods, electronics, durables & FMCG; 6=construction & electricity; 7=Drugs & pharmaceuticals; 8= media and telecommunication; 9= petrochemicals, chemicals & plastic products; 10=steel & other metals and minerals; 11= textiles & apparel; 12=Transport, tourism, hotels & other diversified; 13= software, IT & ITES (Dummy). Univariate regression analysis results with industry as dummy variable. For details on industry as dummy variable, refer chapter III, Database and research methodology, pp

22 Table 6.6 Regression Analysis (Univariate Analysis ) Independent Variable Constant Beta t-value R2 Adjusted R2 Log of Sales * * Log of Assets * * Log of Market Cap * * ROS ROA ROCE RONW Leverage Foreign Activity ** ** Listing Status * * Auditing Firm * * *** *** Industry Note: *,**,*** significant at 1%, 5%, 10% respectively. F 3 4 1=Agri input and tobacco; 2=automotive; 3=Banking & financial services; 4=capital goods, industrial and engg products; 5=consumer goods, electronics, durables & FMCG; 6=construction & electricity; 7=Drugs & pharmaceuticals; 8= media and telecommunication; 9= petrochemicals, chemicals & plastic products; 10=steel & other metals and minerals; 11= textiles & apparel; 12=Transport, tourism, hotels & other diversified; 13= software, IT & ITES (Dummy). Univariate regression analysis results with industry as dummy variable. For details on industry as dummy variable, refer chapter III, Database and research methodology, pp

23 The profitability of a firm is found to have positive and significant association with disclosure score when measured in the terms of ROCE and RONW for the year It is positively associated in the terms of ROA for the year and ROS, ROA, ROCE and RONW for the year But a negative association has been found in the case of profitability of a firm measured in the terms of ROS for the year Hence, no conclusive result could be found out and is advocated by the literature. However, profitability of a company as measured by ROCE has a positive effect on disclosure score for the two years of the study. The value of adjusted R 2 of ROCE (.049) is more than RONW (.034) with F value (13.475) significant at 1% level for the year implying hereby that ROCE is a better indicator of overall profitability position of a firm. It means that the more profitable firms will disclose more information. The results of table 6.5 and 6.6 show, that the audit firm size and listing category of a firm have positive association with the disclosure scores at 1% level of significance. It implies hereby that the firms being audited by big-six audit firms and listed as A category in the Indian stock exchanges have more extent of intangible assets disclosure. The level of foreign activity of a firm is also found to have positive association with disclosure score at 1% and 5% level of significance for the years and respectively. It shows that the firms having high exports disclose more information about their intangible assets to their investors. Leverage of a firm was found to be negatively associated with its intangible assets disclosure score for both years of the study. Thus the firms with high debt in their capital structure disclose less on their intangible assets. The nature of industry influences the intangible assets disclosure level of selected companies in India. As per the first approach 5, except construction, pharmaceutical and steel industry for the year and transport industry for the year , others industrial sectors could not significantly influence the disclosure score individually. The pharmaceutical industry has more extent of intangible assets disclosure while construction, steel and transport industry has less extent of disclosure as compared to software industry. But these sectors significantly influence the variations in the intangible assets disclosure score for two years of the study. The second approach also shows industry to be significant attribute (at 1% level of significance) that has an impact on the level of intangible asset reporting by companies. 5 Industry attribute has been studied under two approaches. For more details refer Chapter III, Database and Research Methodology, pp

24 Backward Step-wise Regression Analysis The regression equation analysing the impact of various corporate specific attributes on the intangible assets disclosure score for two years of the study can be framed as follows: Y = β 0 +β 1 X 1 +β 2 X 2 +β 3 X 3 +β 4 X 4 +β 5 X 5 +β 6 X 6 +β 7-18 X ε Where Y = Disclosure score; X 1 = Size of a firm (log of market capitalization); X 2 = Profitability of a firm (ROCE); X 3 = Leverage of a firm; X 4 = Foreign Activity; X 5 = Listing category of a firm; X 6 = Audit firm size of a firm; X 7-18 = Industry type 6 ; β = Slopes of the independent variables while β 0 is a constant or the value of Y when all values of X are zero; ε = ε t ~ (0, N). Initially, multivariate regression was performed by taking into consideration all independent variables hypothesised to have association with the disclosure score. This could not reveal any conclusive result. Afterwards backward stepwise regression analysis was run. It appears to be the preferred method of exploratory analyses where the analysis begins with a full or saturated model and variables are eliminated from the model in an iterative process. The fit of the model is tested after the elimination of each variable to ensure that the model still adequately fits the data. The analysis is complete when no more variables could be eliminated from the model. The results of backward step-wise regression analysis results for the year and , including the adjusted R 2, F-statistic values and significant corporate attributes, of all regression equations has been shown in table 6.7 and 6.8 respectively. 6 1=Agri input and tobacco; 2=automotive; 3=Banking & financial services; 4=capital goods, industrial and engg products; 5=consumer goods, electronics, durables & FMCG; 6=construction & electricity; 7=Drugs & pharmaceuticals; 8= media and telecommunication; 9= petrochemicals, chemicals & plastic products; 10=steel & other metals and minerals; 11= textiles & apparel; 12=Transport, tourism, hotels & other diversified; 13= software, IT & ITES (Dummy). 139

25 Table 6.7 Backward Stepwise Regression Analysis ( ) Variables/Equation Constant (0.950) (0.944) (0.991) (0.442) (-1.031) (-1.708) (-2.157) (-2.960) (-3.261) (-3.346) (-3.385) (-3.633) (-4.183) (-4.421) (-4.536) (-4.857) (-5.051) log market capitalisation 8.777* 8.813* 9.605* * * * * * * * 9.933* * * * * * * (2.201) (2.202) (2.251) (2.328) (2.266) (2.263) (2.245) (2.262) (2.266) (2.230) (2.187) (2.209) (2.264) (2.276) (2.274) (2.292) (2.286) ROCE (0.018) (0.018) Leverage (-0.007) Foreign activity (-0.024) (-0.024) (-0.024) (-0.025) Listing dummy (1.821) (1.821) (1.721) Audit firm size 2.451** 2.472** 2.458* 2.666* 2.865* 3.098* 3.162* 3.499* 3.426* 3.498* 3.482* 3.510* 3.405* 3.485* 3.490* 3.661* 3.924* (1.988) (1.990) (1.974) (2.119) (2.271) (2.417) (2.473) (2.683) (2.629) (2.683) (2.674) (2.697) (2.605) (2.662) (2.672) (2.792) (2.960) ** ** * * ** *** (-5.364) (-5.370) (-5.554) (-5.553) (-3.991) (-3.366) (-2.832) (-2.241) * * * * ** ** *** (-5.730) (-5.733) (-5.807) (-5.618) (-3.892) (-3.241) (-2.683) (-2.070) (-1.782) ** ** ** ** ** *** *** (-7.332) (-7.356) (-7.547) (-7.458) (-5.722) (-5.063) (-4.529) (-3.872) (-3.582) (-3.313) (-3.036) * * * * * * ** ** *** (-5.543) (-5.545) (-5.678) (-5.425) (-4.017) (-3.392) (-2.870) (-2.262) (-1.959) (-1.702) (-1.423) (-1.304) * * * ** *** *** (-4.758) (-4.762) (-4.715) (-4.404) (-2.994) (-2.317) (-1.769) * * * * * * * * ** ** ** *** *** (-7.030) (-7.033) (-7.277) (-6.925) (-5.168) (-4.490) (-3.935) (-3.266) (-2.985) (-2.684) (-2.378) (-2.266) (-2.046) (-1.890) (-1.765) (-1.569) * * * * * ** ** *** (-4.843) (-4.846) (-4.843) (-4.895) (-3.808) (-3.172) (-2.618) (-2.031) (-1.738) (-1.418) *** ** ** ** *** (-5.448) (-5.454) (-5.771) (-5.559) (-3.940) (-3.303) (-3.115) (-3.118) (-3.272) (-3.234) (-1.863) * * * * * * * ** ** ** *** *** (-6.498) (-6.503) (-6.660) (-6.470) (-5.225) (-4.554) (-4.010) (-3.343) (-3.058) (-2.774) (-2.484) (-2.364) (-2.119) (-1.961) (-1.837) * * * * * * ** ** ** *** *** (-6.818) (-6.822) (-7.028) (-7.008) (-5.824) (-5.181) (-4.638) (-4.022) (-3.729) (-3.436) (-3.121) (-3.017) (-2.784) (-2.643) * * * * * ** ** *** *** (-5.819) (-5.824) (-5.988) (-5.688) (-4.604) (-3.947) (-3.405) (-2.759) (-2.470) (-2.186) (-1.890) (-1.775) (-1.529) R Square Adjusted R Square F * * * * * * * * * * * * * * * * * DW Note: *,**,*** significant at 1%, 5%, 10% respectively 7 1= Agri input and tobacco; 2=automotive; 3=Banking & financial services; 4=capital goods, industrial and engg products; 5=consumer goods, electronics, durables & FMCG; 6=construction & electricity; 7=Drugs & pharmaceuticals; 8= media and telecommunication; 9= petrochemicals, chemicals & plastic products; 10=steel & other metals and minerals; 11= textiles & apparel; 12=Transport, tourism, hotels & other diversified; 13= software, IT & ITES (Dummy). 140

26 Seventeen cross-sectional regression equations have been framed for the financial year for the final analysis. The result of the multivariate regression analysis (first regression equation) shows the R 2 at It signifies that 48.4% variation in model is explained by all independent variables jointly. The adjusted R 2 (.436) states that 43.6% variation in the disclosure score is explained by variations in independent variables. The variables namely size of a company (positive at 1% level of significance), audit firm size (positive at 5% level) and nature of industry (automotive industry, capital goods industry, consumer goods industry, construction industry, pharmaceuticals industry, steel industry, textiles industry and transport industry at 1% level of significance; agri input industry, banking industry at 5% level of significance; and media and telecommunication industry at 10% level of significance) influence the disclosure of intangible assets. Rest of the variables could not significantly influence the disclosure score. The value of F (10.046) is significant at 1% level which shows the fitness of the model. The importance of good regression equation lies in the fact that the model is correctly specified. The regressors have the correct (i.e., theoretically expected) signs and that (hopefully) the regression coefficients are statistically significant at lowest possible level (one or five percent) of significance (Gujarati, 2006, p.260). Therefore, variables were removed one by one and afterwards the model having highest adjusted R 2, signs of independent variables (moving in expected directions) and significant at the lowest possible level of significant has been chosen for two years of the study. The cross sectional regression equation namely 17 in table 6.7 has been selected for the final analysis. The results of the backward step-wise regression analysis explain that the regressors namely size of a firm (positive at 1% level of significance) and audit firm size (positive at 1% level of significance), when regressed jointly explain 41.2% variations in the disclosure score. The F value (74.948) is significant at 1% level explains that model is best fit. All independent variables move in the expected directions. It shows that the size of a firm and audit firm size are found to be significantly affecting the disclosure of intangible assets in the companies in India and is in predicted direction. The final model has been selected on the basis of lowest possible level of significance of regressors with predicted directions and highest adjusted R 2 and F value. 141