1 Introduction

Accounting for share-based remuneration represents one of the most controversial issues in the accounting and corporate governance research literatures. Financial reporting and corporate governance are indeed highly interrelated: in particular, financial reporting constitutes an important element of the corporate governance system, as it may potentially reduce the information asymmetry between corporate outsiders and insiders such as executive directors and controlling shareholders (see, for example, Whittington 1993; Melis 2004; Di Pietra 2005).

The recognition of share-based remuneration as a cost that is expensed in the Profit and Loss Account is the recent outcome of a long debate between standard-setters and the preparers of accounts (Guay et al. 2003). In the US, the final result has been the issue of a revised version of SFAS 123R (2004); in Europe, it is the mandatory adoption by listed companies of IFRS 2 (2004). Each of the two standards has required the recognition of the cost of share-based remuneration, to be measured at the fair value of the equity instruments at the grant date.

Accounting is concerned with how economic actors process information and make decisions. It cannot be considered simply as a neutral technique for economic decision-making as it is able to sanction the distribution of wealth among corporate stakeholders, including shareholders (e.g. Horngren 1973; Rappaport 1977). Both the issue of SFAS 123R in the US and of IFRS 2 by the IASB have followed significant lobbying activity by constituents (Zeff 2002; Shelton and Stevens 2002; Giner and Arce 2007). This level of interest in the process is due to the potential economic consequences that the regulation of share-based remuneration could have on the wealth of corporate stakeholders, and on corporate governance in general.

The main purpose of this study is to explore the impact of the mandatory adoption of IFRS 2 for accounting for share-based remuneration. Based on a sample of Italian nonfinancial listed companies, the paper will investigate the effects of recording the cost at its fair value in terms of its impact on diluted earnings per share (EPS), and it will also consider the extent of disclosure in corporate annual reports. In addition, the analysis will assess the implications for corporate governance in terms of reducing information asymmetry between corporate insiders and outsiders.

2 Developments in the regulatory framework

In Italy, the adoption of IFRS 2 has completely changed the nature of accounting for share-based remuneration. Beforehand, the relevant Italian accounting standards (CNDC-CNR 2001; OIC 2007) had not included any pronouncements on such payments. Nevertheless, Italian companies could have voluntarily expensed the costs involved, as there is no impediment within the Italian framework to the recognition of such costs in the Profit and Loss Account (Corbella and Florio 2007).

The diffusion of share-based remuneration in Italy is relatively recent. Fixed stipends have been the mainstay of directors’ remuneration packages, and share-based schemes were rarely adopted in the past by Italian companies, even during the 1990s (Melis 1999; Zattoni 2003). Their present popularity seems to be dependent on factors such as their favourable fiscal treatment by comparison with cash and other in-kind remuneration (Di Pietra and Riccaboni 2001; Quagli et al. 2006) and the accounting treatment of these transactions (Quagli 2006).

Italian tax law has provided a favourable fiscal treatment since 1998. Share-based remuneration has been considered as an incentive mechanism able to foster a long term relationship between the firm and its high-qualified personnel and was tax-extempted. The 2000 legislationFootnote 1 eliminated the tax exemption on equity-based remuneration, by imposing the recipients to pay a 12.5 per cent rate on the capital gain. This rate was still favourable as it was lower than the one applied to cash-based remuneration (Autuori 2001; Di Pietra and Riccaboni 2001).

Before 2000, any information regarding the share-based remuneration packages that were then in force was not publicly available. A change in the accepted practice was initiated in 2000 by Communication No. 11508 issued by the Italian Securities Exchange Commission (the Commissione Nazionale per le Società e la Borsa, CONSOB) which required Italian listed companies to disclose the details of their share-based plans for directors and senior managers in the notes to the accounts as well as to adjust shareholders’ equity when the options were exercised by the holders. However, no clear sanction was imposed to cover those cases in which companies did not meet this provision.

The recording of the cost of share-based remuneration was voluntary for listed companies before 2005. Share-based remuneration was basically an off-balance sheet operation, as almost none of the companies that gave share-based options to their senior management (or other employees) as part of their compensation recognised the cost in their income statement. This creative accounting practice led to an understatement of recorded costs, with potential economic consequences, given the differential impact on financial statement users between the disclosure of a cost in the notes and its full recognition in the Profit and Loss Account (see Barth et al. 2003).

In 2005, the adoption of IFRSs has obliged Italian listed companies to recognise the fair value of share-based remuneration as an expense in their Profit and Loss Account (IFRS 2 2004). This cost is to be measured at the fair value of the option that is granted multiplied by the number of options that the company estimates will be exercised. IFRS 2 (paragraphs 44–52) also requires companies to disclose information about the kind of option pricing model used and the inputs to that model, including the weighted average share price, exercise price, expected volatility, option life, expected dividends, risk-free interest rate, and any other inputs to the model.

3 Research design and method

3.1 Sample

This study focuses on nonfinancial listed companies in Italy, which, as elsewhere in the EU, have been required to adopt IFRSs from 2005 onwards. Banks, insurance companies and other financial institutions have been eliminated in view of the sector-specific regulation of the financial industry. We also excluded listed companies that were not required to adopt IFRSs in 2005, in accordance with CONSOB regulations (CONSOB 2005, para 81bis, 82bis), as well as non-domestic companies listed on MTA International,Footnote 2 as the latter were not required to prepare and present their consolidated financial statements in accordance with Italian GAAP. From the remaining set of Italian nonfinancial companies listed in 2004 and 2005, the annual reports of those companies which have share-based plans in place (69 companies) have been analysed for these 2 years.

Following the adoption of IFRS 2, the majority of the 69 companies charged the cost of share-based remuneration against income. The only exceptions were amongst those companies that had schemes in place before 7 November 2002, and which were able to avoid expensing the costs involved by adopting one of the transitional provisions that allowed for non-adoption (IFRS 1 2004, para 25b–c; IFRS 2 2004, para 53–58).Footnote 3 For this reason, a further 23 companies were eliminated from the sample. Hence, 46 companies comprise the final sample. Summary statistics that describe this sample are presented in Table 1.

Table 1 Sample firms characteristics (31 December 2004)
Full size table

3.2 Data collection and analysis

The annual report is the most comprehensive document that is made available to the public by firms and is generally treated by researchers as the main disclosure vehicle (see Marston and Shivres 1991; Bassett et al. 2007). Empirical support for reliance on annual reports in this respect is provided by Lang and Lundholm (1993), who demonstrate that the level of disclosure therein is strongly and positively correlated with the amount of corporate information communicated to the market and to stakeholders using other media (Tables 2, 3).

The data required for the present study were gathered from annual reports for the years 2004 and 2005. The period considered was selected in order to allow us to conduct a natural experiment exploring the impact of current regulatory change on the practices surrounding share-based remuneration. This is because the 2004 results have been measured twice: all of the companies within the sample have been required to prepare and present their own 2005 consolidated financial statements in accordance with IFRSs, including the comparatives for 2004, whilst the original accounts for 2004 were all prepared in accordance with national GAAP (e.g. Andrei 2006; De Jong et al. 2006; Callao et al. 2007; Daske et al. 2007; Jaruga et al. 2007).

As discussed earlier, the research question that this paper seeks to answer is: ‘has the mandatory adoption of IFRS 2 improved the disclosure of the cost of share-based remuneration in Italian nonfinancial listed companies’ annual reports?’ To this end, the ability of IFRS 2 to alter the perceptions of financial statement users is explored not only in terms of disclosure levels, but also the impact on one of the key financial performance ratios, diluted EPS (see Botosan and Plumlee 2001; Street and Cereola 2004; Chalmers and Godfrey 2005).

We compare the diluted EPS calculated using the actual 2004 reported financial figures with the adjusted ratio after expensing share-based remuneration. The difference is equal to the change in the cost of share-based remuneration from Italian GAAP to IFRS, divided by the weighted number of shares, using hand-collected data from companies’ annual reports.

The extent of disclosure has been measured on a quantitative basis using disclosure indices. The use of indices to measure disclosure has a long tradition in accounting research. Although it is argued by some that disclosure is an abstract concept that cannot be measured directly (Marston and Shivres 1991), indices are able nevertheless to provide an indirect, although to some extent subjective, measure of the underlying concept. Despite this intrinsic limitation, indices have proved to be a valid research tool in empirical research, as their validity has been demonstrated not only by the consistent results documented by previous studies on the determinants of the disclosure level (see Ahmed and Courtis 1999), but also by specific analyses carried out to assess their ‘internal’ and ‘external’ validity (Botosan 1997).

A disclosure index is usually constructed as a function of the number and, sometimes, the relevance of the items provided in the financial statements (e.g. Botosan 1997; Prencipe 2004). Here, in order to limit subjectivity, the selection of items is based on the disclosure requirements under IFRS 2 (see Bassett et al. 2007). IFRS 2 does not prescribe a particular model to be utilised when determining the fair value of the options, but instead requires companies to disclose which option-pricing model is used (IFRS 2 2004, para 47) and to describe the inputs to the model. Accordingly, for this study, the disclosure index has been constructed from the following specified items: (1) cost of share-based remuneration, (2) option-pricing model, (3) exercise price, (4) life of the option, (5) expected volatility, (6) expected dividends, (7) risk-free interest rate, and (8) weighted average of the share price. All eight items are relevant to all the companies included in the sample.

We adopt both an unweighted and a weighted index to measure the extent to which the mandatory adoption of IFRS 2 has improved the disclosure of the cost of share-based remuneration. The use of both weighted and unweighted disclosure indices is widespread. Unweighted indices score each item equally (see Cooke 1989, 1992; Raffournier 1995) while others allow for differences in their perceived relevance (see Buzby 1974, 1975; Botosan 1997). As subjectivity cannot be completely avoided when determining weights, unweighted and weighted indices are often used together and the results compared in order to see the extent to which the indicing is sensitive to the assigned weights (see Chow and Wong-Boren 1987; Prencipe 2004).

For the eight items considered here, the unweighted index is calculated as

$$ T_{j} = {\frac{1}{8}}\sum\limits_{i = 1}^{m} {d_{i} } $$

where T j is the total disclosure score for company j, d i is 1 if item i is disclosed and 0 otherwise, and m is the maximum number of items (eight in this study).

For the weighted index, in order to assign weights to the different items, we consider whether or not the relevant information is publicly available. More specifically, we categorise variables as either market-based or company-based, and then ask how critical the information may be in appreciating the remuneration scheme. The risk-free rate (RFR) is given the lowest weight as it is publicly available, market-based information. We also assign the lowest weight to the option-pricing model (OPM), as, although it is company-based, the application of different models produces similar results in options whose time-horizon is as short as the options considered here. We weight the expected dividends (ED), the expected volatility (EV) and the weighted average share price (WSP) more highly. For expected dividends, although publicly available, we take into account the fact that they derive from company policy choices. Similarly, although the expected volatility is a market variable, the time-frame used in its calculation is decided upon subjectively by the company. The weighted average share price is also a market variable, but it can only be appreciated by an investor if the grant date of the option has been disclosed by the company. Hence we consider these three variables to be partially market-based and partially company-based. The exercise price (EP) and the life of the option (L) are weighted most heavily, as these variables are determined solely by the company itself in the context of its governance arrangements (e.g. Bernhardt 1999; Bebchuk et al. 2002; Zattoni 2007).

In view of the above, the weighted index is calculated as follows:

$$ T_{j} = {\frac{1}{2}}C + {\frac{1}{2}}\left[ {{\frac{1}{14}}\left( {3{\text{EP}} + 3L + 2{\text{ED}} + 2{\text{WSP}} + 2{\text{EV}} + {\text{RFR}} + {\text{OPM}}} \right)} \right] $$

In this case, we introduce a three-level scale of disclosure for the cost of the option (C): full disclosure, adequate disclosure and inadequate disclosure. Full disclosure of the cost of share-based remuneration is either inclusion as a separate item in the Profit and Loss Account, or as a clear statement of the cost in the notes to the accounts. Disclosure is treated as adequate when the company only recognises the corresponding increase of the cost in equity (equity-settled share-based payments) or liabilities (cash-settled share-based payments). Inadequate disclosure refers to those cases in which companies failed to disclose the fair value within the annual report.

Table 2 Weighted and unweighted disclosure indices scoring
Full size table

4 Results

Our findings show that the great majority of the companies which recorded the cost of share-based remuneration in their Profit and Loss Account in 2005 used to treat share-based remuneration as an off-balance sheet operation beforehand, as only five out of 46 companies had recorded the cost in their Profit and Loss Account before IFRS 2 required them to do so.

The impact of expensing the cost of share-based remuneration on diluted EPS is moderate, although sometimes material.Footnote 4 Whist the average decrease in diluted EPS is 12.9%, the median is only 1.9%. Our findings suggest that this decrease is material for 28.3% of the Italian nonfinancial listed companies with share-based plans. However, this result underestimates the impact, as our estimates are based on those costs that have actually been disclosed in annual reports. In fact, in 51.5% of those cases where the impact is deemed not to be material, the firms adopted one of the transitional provisions (IFRS 1 2004, para 25b–c; IFRS 2 2004, para 53–58) that allowed them to avoid expensing.

Table 3 Disclosures on share-based remuneration
Full size table

The analysis also shows that none of the companies analysed recorded the cost of share-based remuneration as a separate item in the Profit and Loss Account. The great majority (82.6%) recorded it within personnel costs. Among these, 63.0% disclosed the cost of share-based remuneration as a separate item in the notes to the accounts, while 19.6% disclosed it indirectly by providing the amount of the stock option reserve in the notes. 10.9% of companies preferred to record the cost of share-based remuneration under other headings (e.g. extraordinary costs and administrative costs).

As in previous studies on the subject (Botosan and Plumlee 2001; Street and Cereola 2004; Chalmers and Godfrey 2005; Bassett et al. 2007), we document cases of non-compliance with the option value disclosure requirements. Only 47.8% of companies provided full disclosure in their annual report, i.e. they disclosed the cost, the option pricing model used and all six inputs. In accordance with the IASB (1989), non compliance cases are allowed when the value is deemed as not material. However, by using a Bravais-Pearson test, we found no correlation (ρ = −0.1035) between the nonmateriality of the cost and the level of disclosure.Footnote 5

Despite some evidence of creative compliance with respect to those remuneration plans that were set up before 7 November 2002, we are able to confirm that the introduction of IFRS 2 has increased the level of disclosure on share-based remuneration (as measured by our disclosure indices) and has thus contributed to revealing the ‘true’ cost of share-based remuneration to minority shareholders and other investors (see Table 4). The results for the unweighted and weighted indices are very similar in this respect, which is in line with the findings of a number of previous studies on the extent of disclosure (e.g. Choi 1973; Firth 1980; Chow and Wong-Boren 1987). This finding counterbalances the potential lack of objectivity in the choice of the weight given to each item.

Table 4 Impact on disclosure
Full size table

5 Implications for corporate governance and concluding remarks

This paper reports on an exploratory study in Italy of the consequences of the adoption of IFRS 2, in terms of its impact on diluted EPS and disclosure levels in annual reports. Disclosure is at the core of corporate governance (e.g. Cadbury 1999; Mallin 2002), and in this respect it is evident that IFRS 2 has significant implications, especially regarding share-based remuneration. If share-based remuneration does not negatively affect the Profit or Loss, companies have strong incentives to offer such remuneration to their directors and employees instead of cash. Indeed, it has been argued that the use of share-based remuneration has arisen to some extent because of the favourable accounting treatment, which has made the perceived cost of a share-based option much lower than its economic cost (Murphy 2002; Hall and Murphy 2003).

Despite evidence of creative accounting concerning the substance over form principle for remuneration plans set up before 7 November 2002, the mandatory adoption of IFRS 2 has considerably improved the disclosure of share-based remuneration. Thus, it has allowed minority shareholders and other stakeholders to enhance their understanding of the ‘true’ cost of share-based remuneration. As the information asymmetry between corporate insiders (i.e. executive directors and controlling shareholders) and outsiders concerning the cost of share-based remuneration has been lowered, corporate outsiders may better safeguard their interests by fostering the accountability of senior management. Future applications of IFRS 2 are likely to increase the level of disclosure further, as transitional provisions will no longer be in force, and all share-based plans will have to be expensed.

Finally, certain limitations of the study are acknowledged here, as the paper has focused on a single country analysis. Whilst this choice has contributed to internal validity, the extent to which the results of the study may be applied elsewhere is potentially limited. Future research could investigate the impact of the mandatory adoption of IFRS 2 across a range of different settings, or conduct a longitudinal study in order to assess if and how lapses in disclosure may have been addressed in subsequent applications of the appropriate accounting regulations.