Proprietary costs, governance and the segment disclosure decision

The Enhanced Business Reporting Consortium (EBRC) is a collaborative, market-driven initiative promoted by the AICPA that provides an opportunity for users and providers of capital to work together in the public interest to improve the quality of information provided to capital markets. The Consortium promotes greater transparency by developing an internationally recognized, voluntary framework for presentation and disclosure of value drivers, non-financial performance measures and qualitative information. More information can be found at: www.​aicpa.​org.

Segment reporting affects share prices (Kinney 1971; Collins and Simons 1979; Tse 1989: Thomas 2000) and reduces errors in analysts’ predictions (Kinney 1971; Collins 1976; Emmanuel and Gray 1978; Baldwin 1984; Swaminathan 1991; Boatsman et al. 1993; Ahadiat 1993; Berger and Hann 2003).

Other documented benefits are related to an increase in the stock’s liquidity in capital markets, correction of potential mis-valuations of the firm’s stock or an increase in the interest of institutional investors (Healy and Palepu 2001).

The enactment of the Transparency Act in 2003 (26/2003) improved the requirements of transparency and information on corporate boards. Since its enactment, firms are required to file a corporate governance report, offering detailed information on their board’ structure. Boards must comply with the recommendations of the Corporate Governance Code that requires the presence of at least one-third of independent directors (Código Conthe 2006).

The Fourth and Seventh Directives stated that companies have to disclose their sales by “category of activities” and “geographical markets” as part of the Notes to the Financial Statements.

The 437/1998 Royal Decree.

The IBEX-35 is the main index of the Madrid Stock Exchange, including the 35 most liquid stocks.

The 2007 National GAAP apply to all non-listed companies and to the individual accounts of listed companies. Listed companies in Spain must adhere to IAS/IFRS in the preparation of their consolidated annual financial statements.

Small firms are considered as those reporting sales under 11.4 million Euros, and an asset figure lower than 22.8 million Euros and less than 250 employees.

The empirical analysis in this paper is based on IAS 14, which was superseded by IFRS 8 from 2009 Financial Reports onwards.

A similar analysis was carried out by the IASC in 1994, obtaining similar results for IAS 14.

The Report of the AICPA Special Committee on Financial Reporting, Improving Business Reporting—A Customer Focus (1994) refers to the need for the following improvements to the standard: (a) more information about segments, (b) segmentation based on the internal management organization, (c) consistency of the information with other parts of the annual report and finally, (d) more segment disaggregation for some companies.

Prencipe (2004) points to many other firm-specific or institutional characteristics that have been claimed to affect the cost-benefit equilibrium associated with disclosure policy, including company size, listing status, industry sector, the presence in international markets, managers’ compensation plans, ownership structure and litigation costs.

Companies lobbying the SFAS 131 exposure draft were those significantly affected by the new requirements on segment reporting. The new standards resulted in the loss of the reporting flexibility allowed under the previous SFAS 14. Companies’ primary concern was that costs would be imposed by competitors. The new more informative reporting requirements would allow competitors to identify competitive opportunities in the most profitable business segment (Ettredge et al. 2002).

Leung and Horwitz (2004) and Wan-Hussin (2009) extend the analysis into an Asiatic context. However, these authors do not place emphasis on proprietary costs, but on the impact of governance characteristics (i.e. director ownership, independent directors, family firms). McKinnon and Dalimunthe (1993) in Australia and Bradbury (1992) in New Zealand look at the firm-specific determinants of voluntary segment disclosure in line with the general voluntary disclosure literature. Bradbury (1992) only finds a positive association between firm size and leverage. McKinnon and Dalimunthe (1993) in Australia observe that, together with the impact of the size variable, others related to the agency relationship within the corporation also affect the disclosure decision. The ownership structure and the percentage of minority interests are singled out as positively affecting the level of segment disclosure.

This is the argument that the IASB uses in favor of the new management approach on segment disclosure reporting under IFRS 8.

However, majority shareholders may also have an interest in increasing disclosures. When strong legal mechanisms are in place, these firms will try to improve their reputation as highly transparent firms (Patelli and Prencipe 2007; Lim et al. 2007). In addition, majority shareholders are more involved in the daily operations of the firm and therefore, have access to more value relevant information for decision-making. For this reason, a strong presence of majority shareholders can also be related to higher levels of disclosure. However, most of the empirical evidence reports a negative relationship between ownership concentration and disclosure (Broberg et al. 2010; Patelli and Prencipe 2007; Cheng and Courtenay 2006). In the segment disclosure literature McKinnon and Dalimunthe (1993) and Leuz (1999) reveal the positive relationship between ownership diffusion and segment disclosure, while Patelli and Prencipe (2007) found no evidence in the Italian capital market.

All the sample firms are above this threshold.

The latter is the percentage of affiliated directors on the Board but is not considered as an explanatory variable in the empirical model.

Using a panel data set, the common assumption of independence in regression errors is generally violated. This residual dependence across firms or time can result in biased standard error test statistics and misspecified significance statistics in OLS regressions. As Petersen (2009) explains, in panel data analysis, when residuals are correlated, the OLS standard errors are underestimated leading to confidence intervals that are too small. Here, we only report the results of the OLS and the firm-fixed effects regressions. The results of the time-fixed effects estimates are similar to those of the OLS, so the results are not reported. However, the fixed effects approach has certain caveats and limitations. On one hand, it does not produce estimates for the effects of variables that do not change over time (i.e. N-SIC). Apart from this, its estimates may be imprecise for explanatory variables that vary greatly across individuals but have little variation over time for each individual. Controlling for fixed effects when explanatory variables have little within-firm variation leads to substantially larger standard errors, higher p-values and wider confidence intervals, that is, non-significant results. The best situation for a fixed effects analysis is when all of the variation on a time-varying predictor is within-firms (Allison 2006). The ANOVA analysis allows us to know the within-firm variation for the different explanatory variables in our model. For these, most of the variation is between firms, and within-firm variation is low: DifROE = 36 %, %_IND = 27 %, CCAP = 21 %, N_COMPT = 10 %. Results from the fixed effect analysis should be interpreted with caution.

Firm-fixed effects models do not produce any estimates for variables that do not change over time. The industry diversification variable remains stable for the whole sample and therefore cannot be considered in the estimation of the firm-fixed effects model.