Alternative explanations for the association between market values and stock-based compensation expenditure

Journal of Contemporary Accounting & Economics 5 (2009) 95–107

Contents lists available at ScienceDirect

Journal of Contemporary Accounting & Economics journal homepage: www.elsevier.com/locate/jcae

Alternative explanations for the association between market values and stock-based compensation expenditure Zoltan Matolcsy, Suzanna Riddell, Anna Wright * School Of Accounting, University of Technology, Sydney, Australia

a r t i c l e

i n f o

Article history: Received 25 March 2008 Revised 10 September 2009 Accepted 11 September 2009 Available online 27 September 2009 JEL classification: J33 G34 Keywords: Stock-based compensation Market values

a b s t r a c t The relation between stock-based compensation and market values has been tested previously in the literature, but the empirical findings are inconsistent: both negative and positive relations have been documented. The objective of this study is to provide an explanation for why both negative and positive relations between stock-based compensation expenditure and market values can be consistent with rational markets. We argue that stock-based compensation can be used either as a reward for past performance or as an incentive for future performance. We predict that there is a negative relation to market values when stock-based compensation is granted primarily as a reward to chief executives for past performance, while there is a positive relation when stock-based compensation is used to provide incentives for enhanced future performance. This prediction is tested on a sample of 259 firm-year observations for the period 1999–2004 using an instrumental variables approach, where the sample is classified into the ‘reward’ and ‘incentive’ groups on the basis of prior period performance and option characteristics. Our findings are that there is a positive association between stock-based compensation expenditure and market values for the ‘incentive’ group, but we find overall an insignificant relation for the ‘reward’ group. A number of sensitivity tests confirm the main findings. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction This study investigates the relation between stock-based compensation and market values. This relation has been tested in the literature, but the empirical findings are inconsistent, as both negative and positive associations have been found— negative in Aboody (1996) and Aboody et al. (2004b), positive in Rees and Stott (2001), Bell et al. (2002) and Brown and Yew (2002). The first objective of this study is to provide an explanation for why both the negative and positive relations between stock-based compensation and market values can be consistent with rational markets. We predict that a positive relation is expected when stock-based compensation is used to provide incentives for enhanced future performance. Alternatively, we predict a negative relation when stock-based compensation is granted primarily as a reward for prior years’ good performance. The second objective is to provide empirical evidence on this explanation for alternative relations between market values and stock-based compensation.

* Corresponding author. Address: School of Accounting, University of Technology, Sydney, P.O. Box 123, Broadway, NSW 2007, Australia. Tel.: +61 612 9514 3592; fax: +61 612 9514 3669. E-mail address: [email protected] (A. Wright).

1815-5669/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcae.2009.09.001

96

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

There are two motivations for this study. First, there has been much interest1 in the accounting for stock-based compensation,2 and the extant literature has primarily focused on the question of whether or not stock-based compensation should be expensed. This question has typically been addressed by testing the association between the value of stock-based compensation and market values. As previously noted, studies in this area have produced mixed results. The granting of stock-based compensation potentially affects share prices in two opposing ways. The stock-based compensation may dilute the value of the firm’s outstanding stock, but may also provide incentives to employees to increase the firm’s share price. Aboody (1996) and Aboody et al. (2004b) find a negative relation between estimated employee stock option expense and stock price. These two studies conclude that the market prices stock-based compensation as though it is an expense, and hence the negative association. In contrast, Rees and Stott (2001), Bell et al. (2002) and Brown and Yew (2002) document a positive relation, which implies that the market prices stock-based compensation as though it is an asset. These studies typically conclude that incentive effects are dominant in firm valuation. This is consistent with the view that aligning the interests of Chief Executive Officers (CEOs) with those of shareholders results in benefits that increase current and future earnings and offset the cost of dilution (Core and Guay, 1999; Guay et al., 2003). Plausible explanations for these conflicting results have not been addressed in the literature to date, but the question is important—are option grants an expense or an asset?3 The second motivation is that Australia offers a unique and powerful setting within which to study the association between market values and stock-based compensation. The extant academic literature has predominantly focused on stockbased compensation within the context of the United States (Aboody, 1996; Skinner, 1996; Rees and Stott, 2001; Bell et al., 2002; Espahbodi et al., 2002; Hanlon et al., 2003; Aboody et al., 2004a,b; Balsam et al., 2004). This research has limited generalisability to other settings because of the relatively homogenous use of stock options across firms in the US. Murphy (1999) reports that nearly all executive pay packages in the US contain stock options and that during the 1990s stock options became the single largest component of compensation in all industries except utilities. Murphy also reports that the overwhelming majority of those options have exercise price set at fair market value at grant date and they have 10-year terms and have no performance hurdles attached. In contrast, Australia offers a setting in which there is substantial variation in equity compensation. Matolcsy and Wright (2007) find that the option grant characteristics are extremely diverse with respect to the exercise price, option terms and presence of performance hurdles, especially in comparison to the US. This heterogeneity provides a unique opportunity to examine firms’ different possible motivations, such as using reward or incentives, for using stock-based compensation. This study provides empirical evidence on the association between market values and stock-based compensation expenditure for a sample of 259 firm-year observations for the period 1999–2004. The key results suggest that the market incorporates the value of stock-based compensation into price differently, depending on characteristics of the firm and the options granted. In particular, there is a significant positive association between stock-based compensation and market values for firms that grant options primarily to provide incentives to chief executives. The positive relation exists for firms that experienced negative abnormal market returns in the period prior to granting options and for firms that grant options at premium. This relation is predominantly insignificant for firms that grant options primarily as reward for past performance. A number of sensitivity tests confirm the main findings. This study makes a significant contribution to the literature on stock-based compensation and its association with market values as it investigates the nature of the incentive effects and dilution effects of stock-based compensation. Some firms are likely to be granting options to CEOs as a means to reward them for past performance (the ‘reward’ group) and, in doing so, the options are more of a compensation expense and therefore a cost to the firm. Other firms use options as a mechanism for aligning CEOs’ interests with those of existing shareholders in order to reduce agency costs in the future (the ‘incentive’ group). It is possible to classify firms into the ‘reward’ and ‘incentive’ groups based on various different measures, including firms’ prior performance and option grant characteristics. This study therefore helps in the understanding of the different results found in the literature by providing empirical evidence on how investors value stock-based compensation differently in each situation. The remainder of this paper is structured as follows. Section 2 provides some background and theory development. Section 3 provides the sample and research design, and Section 4 reports the main results. The discussion and conclusions of the paper are contained in Section 5. 2. Background and theory development The results from the literature investigating the relation between market values and stock-based compensation expenditure are mixed. Some studies document a negative relation between the value of stock-based compensation expense and stock price. For example, Aboody (1996) finds a negative relation using a broad sample of US firms. Aboody adopts an instrumental variables approach and performs a two-stage least-squares regression of the estimated value of employee stock options on price. This approach is necessary because of the positive mechanical relation between the value of stock 1 Johnston (2006) comments that in the United States, the debate over the accounting for stock-based compensation ‘was one of the most heated in the history of accounting setting’ (p. 399). 2 For recent examples see Chalmers and Godfrey (2005), Frederickson et al. (2006), Hodder et al. (2006), Johnston (2006). 3 The question of whether options are an asset or expense is also important to the popular press. For example, Warren Buffett is well known for his advocacy to treat options only as an expense, and his feelings on the matter have been described as ‘not just as a technical accounting question, but a matter of fundamental morality’ (http://money.cnn.com/2004/05/03/pf/buffett_qanda/page2). Academic evidence will also assist in this popular debate.

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

97

options and price.4 Similarly, Aboody et al. (2004b) document a negative relation between employee stock option value and price using an instrumental variables approach and a two-stage least-squares regression. The negative results could be driven by the composition of these studies’ samples in that they all use US data. Murphy (1999) reports that most US firms have an equity-based compensation scheme in place and he suggests that the majority of firms grant options as compensation for past performance, i.e., for reward purposes. This is partly due to the tax advantages in the US of issuing options to executives instead of fixed compensation5 (Core et al., 2003). It may be the case that the pooled regressions show negative coefficients on the value of employee stock options because the majority of sample firms are granting options as rewards, hence the dilution effects of the options are dominant in firm valuation for the majority of firms in the sample. In contrast to the above findings, other studies document a positive relation between employee stock options and price and/or returns. Rees and Stott (2001) document a positive association between employee stock option expense (as reported in pro forma disclosures) and annual stock returns. They also find that the relation is more positive for firms with high growth options. This result may suggest that firms with certain underlying economic characteristics may have different motivations for granting options. Bell et al. (2002) also find a positive relation in their study, however, they acknowledge that this has limited generalisability because their sample is restricted to profitable computer software firms for the period 1996– 1998. In the only published study on the association between price and employee stock options in Australia, Brown and Yew (2002) document a positive relation using a sample of 121 Australian firms.6 Australian firms do not enjoy tax advantages when issuing stock options and a much smaller percentage of Australian firms grant options than in the US (Matolcsy and Wright, 2007). Hence it is possible that the majority firms in Australia who grant options (but not all) do so primarily for incentive reasons. These important differences could explain why virtually the same research design as Aboody et al. (2004b) produced a different result in Brown and Yew (2002). In this study we use the Australian setting to separately identify firms using options as rewards and those using options for incentive purposes to investigate whether our argument offers an explanation for the different results found in the literature to date. Our perspective is based on agency theory, rather than the notion of rent extraction as proposed by Bebchuk et al. (2002) and others.7 This study starts with the argument that firms could use stock-based compensation for a variety of reasons, including reward for past performance, or to motivate CEOs to enhance future performance. Firms can be loosely classified into these two categories: ‘reward’ or ‘incentive’. The ‘reward’ group receives options in relation to past performance—the compensation is related to historical performance and is not necessarily related to future performance. In this circumstance, granting options can be thought of as a substitute for a simple cash bonus. However, it is important to note that even though these options may be thought of as expenses and not necessarily related to future firm performance they are still option grants and the reward is not payable immediately (an option usually cannot be exercised immediately on award). Thus these options could also be seen to be used as a way for the firm to retain high performing executives in the organisation, as options usually lapse on resignation.8 This argument is also consistent with the contract renegotiation literature, where option repricing helps to prevent executive turnover (for examples, see Carter and Lynch, 2001, 2004). In addition, as the options are not immediately exercisable, the executive still bears risk as the options could go underwater. Thus as well as a reward and employment retention measure, these options could also serve as a mechanism to retain the same level of effort from the CEO that achieved the past performance. In contrast, the ‘incentive’ group is granted options to mitigate agency costs in the future. Options reduce agency costs by aligning the interests of managers and shareholders. This alignment occurs since options provide incentives for the managers to take actions and choose investments that increase share price in the future (Jensen and Meckling, 1976; Core and Guay, 1999). The incentives work on the basis that the stock options become more valuable to the CEO as the share price increases. Both the CEO and the firm’s shareholders benefit from the increase in price. There a several circumstances which provide ex ante incentives for companies to utilise stock options as rewards (Core et al., 2003; Goodwin and Kent, 2004). If we accept that firms can grant options as a reward for past performance then it is necessary to explore the circumstances in which this is likely to occur. Rewards are, by nature, linked to historical performance. It is logical to expect that a firm that has experienced good performance during the previous period will reward its chief executive officer accordingly. A CEOs reward forms part of his or her total compensation package and may take a number of forms. A common example is an annual cash bonus, however, some firms may prefer to substitute cash with stock options (Murphy, 1999).9 In addition, firms that experience superior 4 The Black–Scholes model and other option-pricing models use share price as an input for valuing employee stock options. Consequently, there is a mechanical relation between the dependent and independent variables, which biases the results. The instrumental variables approach replaces the employee stock options value with new variables that are correlated with the value of the options and not correlated with the error term. 5 Tax legislation in the United States provides incentives for firms to grant options because non-performance based compensation to employees in excess of $US1m is not tax deductible whilst the value of employee stock options is tax deductible. In Australia the value of employee stock options is not, in most cases, tax deductible. 6 The sample contained 121 firms listed on the Australian Stock Exchange for the period 1997–2000. 7 The rent extraction arguments contained in papers such as Bebchuk et al. (2002) suggest that managerial power can lead to suboptimal compensation contracts that will, in turn, reduce shareholder wealth. Thus an alternate view of any expense to the firm could be that the options are the result of inefficient compensation contracts and rent extraction. 8 As a result these options will also have some benefit to them and as such may mitigate some of the negative association expected. 9 Firms may use options as a substitute for cash for a number of reasons. Firms that are highly cash constrained and are therefore unable to pay cash bonuses as part of the CEOs compensation scheme may use options instead. Alternatively, stock options may be used to attract and retain highly skilled CEOs in a competitive international labour market.

98

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

performance would already have a reasonable degree of alignment between CEOs’ interests and those of existing shareholders. Therefore, firms that grant stock options and have experienced good prior period performance are more likely to be in the ‘reward’ group where the stock options are being used to retain the CEO and to reward them for the past good performance. Granting options to CEOs as a reward for past performance is likely to be seen as a cost to the firm (in a similar manner to a CEOs salary or annual bonus). In addition, it is an expensive alternative to disbursing cash because of a lack of tax incentives in Australia and because of the costly process required to implement an equity-based compensation scheme at an Annual General Meeting.10 It is therefore likely that investors view the value of options as an expense to the firm.11 Hence, granting stock options primarily for reward purposes has a negative impact on share price. It is therefore predicted that: H1. There is a negative association between the value of stock options and market values for ‘reward’ group firms. On the other hand, some firms may use stock-based compensation more as a method of aligning the interests of CEOs with the firm’s shareholders. This alignment occurs because the value of the stock options increases as the firm’s share price increases, which benefits both the CEO and the shareholders and mitigates some of the agency costs associated with the manager–shareholder relationship (Jensen and Meckling, 1976; Murphy, 1999). Firms that are experiencing high agency costs (where managers’ interests are misaligned with shareholders’ interests) may use options as an incentive to enhance future performance. One can assume that a firm that has experienced poor performance in the prior period is unlikely to be rewarding its chief executive for past performance. In this situation there may be some misalignment between the interests of the CEO and those of existing shareholders as the CEOs past actions resulted in a decrease in shareholder wealth. The granting of CEO stock options can be used as a mechanism to mitigate this misalignment because the CEO and the shareholders will benefit in the future from increases in share prices and returns. Firms that have experienced poor prior period performance are likely to be in the ‘incentive’ group. The ‘incentive’ group of firms grants options because the perceived benefits (the reduced agency costs) outweigh the perceived costs (the dilution effect on shareholders equity). In this instance, we expect that shareholders recognise the net benefits associated with the options and value them as though they are intangible assets. It is therefore predicted that: H2. There is a positive association between the value of stock options and market values for ‘incentive’ group firms.

3. Research design 3.1. Sample and data The sample of firms used in this study is taken from the UTS ‘Who Governs Australia’ database. The sample includes firms that are listed on the Australian Stock Exchange (ASX) and are included in the ‘Top 500’ Australian firms (by market capitalisation)12 for the years 1999–2004. Investment trusts and managed funds have been deleted from the sample due to differences in governance and reporting requirements. Firms reporting in foreign currency, firms where management were paid by another company, firms with missing data and firms listed during the year were not included in the sample. Furthermore, firms are required to have two years’ prior data available for inclusion in the database and firms experiencing a CEO change are also not included. Since this study investigates stock-based compensation, firms that did not grant any options to the CEO during the year are deleted. Firms are also deleted where the CEO was granted two or more sets of options in a given year that had conflicting characteristics.13 The final sample comprises 259 firm-year observations. Sample selection is outlined in Table 1 below. Requirements in the Company Law Review Act of 1998 oblige companies to disclose details in their directors’ reports about the nature and amount of each element of the compensation received by individual directors and top executives. Therefore, companies’ annual reports provide us with the necessary information to estimate the value of the CEO stock options. Share price data is available from the SIRCA CRD14 database for the period 1997–2008. Financial statement data is available from the Aspect Huntley FinAnalysis database for the period 1997–2008. Stock options data that is required includes the grant date, exercise date, expiry date, whether there are performance hurdles attached to the options, the number of options issued, the options’ vesting periods and their term to expiration. Options granted during the year are valued using two methods. The first method values the options using the Black–Scholes formula, with adjustment for dividends.15 The second method that is used is simply to calculate 25% of the exercise price of the options. 10 According to ASX Listing Rules and the Corporations Act 2001, shareholder approval must be given at an Annual General Meeting before stock options can be granted to employees. Such approval must be sought for each and every option scheme and each and every employee. 11 However, we acknowledge that the retention of employment and effort effects discussed earlier may mitigate this negative relation. 12 As listed on Connect 4 (an electronic financial statement database of the top 500 Australian firms). Firms that drop out or join the Top 500 during the sample period are included in the analysis where they pass the sample requirements. 13 This conflict arises when the CEO of a firm was granted a number of options that fall into different option categories based on their exercise price, term and/ or performance hurdles. 14 Securities Industry Research Centre of Asia-Pacific Core Research Database. 15 The model used is the same as Murphy (1999) and Matolcsy and Wright (2007).

99

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107 Table 1 Sample selection. 1999

2000

2001

2002

2003

2004

Total

Original sample Reason for deletion No options granted Conflicting option characteristics Extreme observations

227

224

245

98

186

93

1073

164 4 1

157 4 5

213 1 2

47 – 4

143 – 2

67 – –

791 9 14

Final sample

58

58

29

47

41

26

259

Table 2 Firm and option characteristics that determine whether the primary motivation for granting options is for reward or incentive purposes. Method of classification

Measure

Reward group

Incentive group

Prior period performance Option characteristics

Prior period abnormal returns Exercise price

Positive abnormal returns Discount and fair market value

Negative abnormal returns Premium

3.2. Reward and incentive groups The full sample of firms is divided into two subsamples in order to test Hypotheses 1 and 2. The subsamples are: (1) the ‘reward’ group and (2) the ‘incentive’ group. A number of different proxies are used to categorise the full sample into the ‘reward’ and ‘incentive’ groups. The proxies used are summarised in Table 2 below. This study uses two primary methods for classifying the sample firms into the ‘reward’ and ‘incentive’ groups. As per Table 2, the first method of classification is a market-based measure of past performance. The proxy that is used to measure performance is prior period abnormal market returns.16 Since rewards are linked to historical performance, it is logical to expect that a firm that has experienced good performance during the previous period will want to retain and reward its CEO accordingly. In addition, firms that are experiencing superior performance would already have a reasonable degree of alignment between the CEOs interests and those of existing shareholders. Therefore, we argue that firms that have experienced positive returns in the period prior to granting options are likely to be granting those options for reward purposes. In other situations there may be some misalignment between the interests of the CEOs and those of existing shareholders since the CEOs’ past actions resulted in a decrease in shareholder wealth. The granting of CEO stock options can be used as a mechanism to mitigate this misalignment because the CEO and the shareholders will benefit in the future from increases in share prices and returns. Thus we expect that firms that have experienced poor prior period performance are likely to be in the ‘incentive’ group. The characteristics of the options that are granted may also indicate whether firms are granting options for reward purposes. Unlike US firms, Australian firms exhibit heterogeneity in the characteristics of the options that they grant (Goodwin and Kent, 2004; Matolcsy and Wright, 2007). The Australian setting provides us with a unique opportunity to gain an insight into the reasons for granting options by examining the very nature of the options themselves. The exercise price of options granted can indicate whether or not the options are being used as rewards for past performance. Options that are granted at discount (exercise price < grant date stock price) could be used as rewards because they are already in-the-money. This means that managers do not have to change their future actions in order for the options to be valuable. In fact, even a slight decrease in the stock price will mean that the options are still valuable. Similarly, options that are granted at fair market value (exercise price = grant date stock price) are likely to be used as rewards because a very small upward movement in price will put the options in-the-money. This also means that managers are unlikely to have to change their future actions in order for the options to be valuable. On the other hand, options that are granted at premium (exercise price > grant date stock price) are very likely to be used as mechanisms to align the interests of shareholders and employees in order to enhance future market performance. This is due to the fact that some effort is required on behalf of the option holder to increase the share price before the options are inthe-money. This provides more of an incentive to work hard than, say, a discount or fair market value option would. Therefore, firms that grant premium options are considered ‘incentive’ firms.17 Descriptive statistics for the pooled sample are reported in Table 3 below. The descriptive statistics displayed in Table 3 show that the data is highly right-skewed. The sample contains a number of very large firms and, as such, the sample violates the presumed condition of a normally distributed data set. This presents a problem within the context of this study since the regressions that are used assume that the population from which the data

16

Measured using the Market Model to calculate the firm-specific abnormal returns in the year prior to granting options. Unlike the United States where repricing of employee stock options is common (Murphy, 1999; Core et al., 2003), Australian firms rarely reprice options subsequent to the grant date (Coulton and Taylor, 2002). The incentives associated with premium options are larger than fair market value and discount options, especially in an environment where repricing is rare. 17

100

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

Table 3 Descriptive statistics. Pooled sample

ASSETSt EQUITYt REVENUEt MVEt OPATt OPATt1 TOTAL_COMPt SHARESt ROAt ROEt AB_PERFt1 MKT_BKt1 ROAt1

n

Mean

Median

Standard deviation

259 259 259 259 259 259 259 259 259 259 259 259 259

$ 15,367,430 $ 1,387,067 $ 1,567,892 $ 2,830,863 $ 187,851 $ 173,452 $ 758 378,454 0.0179 0.0611 0.1846 2.8903 0.0090

$ 352,000 $ 159,000 $ 323,503 $ 330,975 $ 17,700 $ 12,408 $ 360 155,290 0.0500 0.1182 0.0874 1.1356 0.0528

$ 60,476,161 $ 3,983,406 $ 3,913,204 $ 8,112,334 $ 618,243 $ 609,698 $ 1,171 874,836 0.1698 0.6068 0.8708 6.9001 0.2850

ASSETSt, book value of total assets, $’000; EQUITYt, book value of total equity, $’000; REVENUEt, total operating revenue, $’000; MVEt, market value of equity (calculated as average share price for 3 months after year end multiplied by number of shares outstanding), $’000; OPATt, total operating profit after tax, $’000; OPATt1, total operating profit after tax (t  1), $’000; TOTAL_COMPt, total CEO compensation, $’000; SHARESt, number of common shares outstanding as at year end, $’000; ROAt, return on assets; ROEt, return on equity; AB_PERFt1, abnormal performance (t  1), calculated using Market Model; MKT_BKt1, market-to-book ratio (t  1); ROAt1, return on assets (t  1).

is taken is normally distributed. To mitigate this problem, the data is analysed on a per share basis, and all variables are windsorised using the appropriate test in SPSS. 3.3. Experimental design Hypotheses 1 and 2 are tested using two alternative methodologies. The first method tests the association between stock price and the book value of equity, net income and the value of stock options for the pooled sample. The model is a variation on the model developed in Ohlson (1995), which relates the value of the firm to the information provided in the firm’s statements and includes accounting earnings and book value of equity. The second method tests the same association, but adopts an instrumental variables approach and performs a two-stage least-squares regression. 3.3.1. Method 1: the association between stock price and stock option value The association between stock options value and stock price is examined through a cross-sectional regression, where price is regressed on per share values of net book value of equity, net income and estimates of the value of stock options. Options are valued using the Black–Scholes option valuation model.

Pit ¼ a1it þ a2 BVEit þ a3 NIit þ a4 OPTit þ rit

ð1Þ

where Pit is the average share price for three months after year end (for firm i at time t); BVEit is the book value of equity per share (for firm i at time t); NIit is the net income per share (for firm i at time t); OPTit is the value of stock options per share – calculated using the Black–Scholes valuation model (for firm i at time t). The regression in Eq. (1) is expected to yield results that show a positive coefficient on the book value of equity, net income and the value of stock options for the pooled sample. The results from the regression in Eq. (1) demonstrate whether the value of stock options is related to market values. However, the results have a positive bias due to the positive mechanical relation between price and the value of options. Therefore, this method does not directly test Hypotheses 1 and 2, but it is useful as a comparison with results from prior studies. 3.3.2. Method 2: instrumental variables approach As Aboody et al. (2004b) point out, option values ‘are a positive function of the price of the underlying share, even when the option’s intrinsic value equals zero’. This has the consequence of a mechanical positive relation between share price and option value that is unrelated to investors’ valuation assessments. The results from running the regression in Eq. (1) are expected to demonstrate exactly this: the mechanical positive relation between options and share price. In order to address this endogeneity problem, it is necessary to adopt an instrumental variables approach to estimate the value of the options. This involves choosing instruments that are highly correlated with the value of the options, OPT, but not correlated with the error term, r. The instruments that are used include: the volatility of the underlying stock, the vesting period, the risk-free interest rate, the expected dividend yield and the number of options granted, as well as the other explanatory variables in Eq. (1) – BE and NI.18 The instrumental variables estimation model can be written as follows:

18

These instruments are consistent with those used in Aboody et al. (2004b) and Brown and Yew (2002).

101

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107 Table 4 Correlation matrix – all changed in this and following tables. Pearson correlation n = 259

P

Spearman’s rho P BVE NI PREDOPT OPTBS OPT25%

0.820*** 0.844*** 0.160*** 0.194*** 0.353***

BVE

NI

PREDOPT

OPTBS

OPT25%

0.668***

0.744*** 0.787***

0.040 0.060 0.061

0.127** 0.023 0.070 0.617***

0.173*** 0.143** 0.169** 0.293*** 0.547***

0.791*** 0.008 0.073 0.338***

0.025 0.110* 0.326***

0.405*** 0.263***

0.709***

P, average price per share for three months after year end; BVE, book value of equity per share; NI, net income per share; PREDOPT, estimated options value per share from 2SLS regression; OPTBS, options value per share measured using the Black–Scholes model; OPT25%, options value per share, calculated as 25% of exercise price  no. of options granted. * Significant at the 10% level (2-tailed). ** Significant at the 5% level (2-tailed). *** Significant at the 1% level (2-tailed).

OPTit ¼ b1it þ b2 VOLit þ b3 LIFEit þ b4 INTit þ b5 DIVit þ b6 NOPTit þ b7 BVEit þ b8 NIit þ tit

ð2Þ

where OPTit is the value of stock options per share – calculated using the Black–Scholes model; (for firm i at time t); VOLit is the share price volatility – % (for firm i at time t); LIFEit is the vesting period – grant date till vesting date in years (for firm i at time t); INTit is the risk-free interest rate – % (for firm i at time t) DIVit is the expected dividend yield (for firm i at time t); NOPTit is the number of options granted per share (for firm i at time t); BVEit is the BOOK value of equity per share (for firm i at time t); NIit is the net income per share (for firm i at time t). In the second-stage regression, Eq. (2) is estimated and the observed OPTit is replaced with the value of the options calculated in the first-stage regression, PREDOPTit, yielding the valuation model shown below in Eq. (3).

Pit ¼ c1it þ c2 BVEit þ c3 NIit þ c4 PREDOPTit þ eit

ð3Þ

where Pit is the average share price for three months after year end (for firm i at time t); BVEit is the book value of equity per share (for firm i at time t); NIit is the net income per share (for firm i at time t); PREDOPTit is the estimated value of stock options per share – the value from Eq. (2) above (for firm i at time t). The ordinary least-squares regression in Eq. (1) and the two-stage least-squares regression in Eqs. (2) and (3) are tested on the pooled sample and both the ‘reward’ and ‘incentive’ groups separately. For the pooled sample, the two-stage leastsquares regression is expected to yield positive coefficients on the book value of equity and net income, but the sign on the estimated value of the options cannot be predicted. For the ‘reward’ group, the two-stage least-squares regression is expected to yield positive coefficients on the book value of equity and net income, but a negative coefficient on the estimated value of the options. For the ‘incentive’ group, the two-stage least-squares regression is expected to yield positive coefficients on the book value of equity, net income and the estimated value of the options. 4. Results 4.1. Pooled sample Prior to testing Hypothesis 1 and 2 for the ‘reward’ and ‘incentive’ groups, the tests outlined above are performed on the pooled sample of 259 firm-year observations. These tests are performed on the pooled sample to facilitate a comparison with results documented in the prior literature (Aboody, 1996; Rees and Stott, 2001; Bell et al., 2002; Brown and Yew, 2002; Aboody et al., 2004b). The correlation coefficients of each of the variables of interest are reported in Table 4. It is clear that nearly all of the variables are significantly correlated with each of the remaining variables. However, it is interesting to note that the estimated value of options taken from the two-stage least-squares regression is not correlated with stock price based on the Pearson Correlation. On the other hand, both the Black–Scholes options value and the ‘25% of exercise price’ options value are significantly positively correlated with price. This is likely to be due to the mechanical positive relation between price and the values produced by these valuation methods. In addition, the estimated value of options is significantly positively correlated with the Black–Scholes options value and the ‘25% of exercise price’ options value, suggesting that the instruments chosen for the two-stage least-squares regression are appropriate. 4.1.1. Results from the ordinary least-squares regression19 The results from the ordinary least-squares regression are shown in Table 5. 19

All regressions reported in the paper include fixed-year effects as well as correcting for the standard errors using Huber–White corrections.

102

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

Table 5 Summary of the OLS regression coefficients for the relation between share price and option value measured using the Black–Scholes valuation model (pooled sample). The coefficients are based on the following equation: Pit ¼ a1it þ a2 BVEit þ a3 NIit þ a4 OPTit þ rit . Variables n

Predicted sign

Pooled sample 259

Intercept BVEit NIit OPTit

? + + +

0.206 (0.258) 0.953 (2.205**) 13.825 (5.066***) 118.535 (1.757*) 0.580

Adjusted R2 F ratio

45.535***

Pit, average price per share for three months after year end; BVEit, book value of equity per share; NIit, net income per share; OPTit, options value per share, measured using the Black–Scholes model. * Significant at the 10% level (2-tailed). ** Significant at the 5% level (2-tailed). *** Significant at the 1% level (2-tailed).

Table 6 Summary statistics from the OLS regression of Black–Scholes option value on instrumental variables (pooled sample). The coefficients are based on the following equation: OPTit ¼ b1it þ b2 VOLit þ b3 LIFEit þ b4 INTit þ b5 DIVit þ b6 NOPTit þ b7 BVEit þ b8 NIit þ tit . Variables n

Predicted sign

Pooled sample 259

Intercept VOLit LIFEit INTit DIVit NOPTit BVEit NIit

? + + +  + + +

0.003 (0.415) 0.004 (3.805***) 0.000 (0.039) 0.092 (0.643) 0.034 (3.217***) 0.323 (5.874***) 0.000 (1.159) .005 (3.683***) 0.336

Adjusted R2 F ratio

11.884***

OPTit, options value per share, measured using the Black–Scholes model; VOLit, share price volatility – % (for firm i at time t); LIFEit, vesting period – grant date till vesting date in years (for firm i at time t); INTit, risk-free interest rate – % (for firm i at time t); DIVit, expected dividend yield (for firm i at time t); NOPTit, number of options granted per share (for firm i at time t); BVEit, book value of equity per share; NIit, net income per share. *** Significant at the 1% level (2-tailed).

Table 5 displays the coefficients of the regression when the employee stock options value is measured using the Black– Scholes valuation model. As expected, the coefficient on the book value of equity is positive and significant (at the 5% level). There is also a positive coefficient (significant at the 1% level) on net income. The coefficient on the value of options is significantly positive at the 10% level for the pooled sample. This result is in line with expectations due to the mechanical positive relation between the dependent variable, price, and the variable of interest, the options’ value. This positive result is consistent with prior studies that use this model (Aboody, 1996; Rees and Stott, 2001; Bell et al., 2002; Brown and Yew, 2002; Aboody et al., 2004b). 4.1.2. Results from the two-stage least-squares regression Table 6 depicts the results of the first-stage regression. The coefficient of volatility (VOL) dividend (DIV), number of options granted (NOPT) and net income (NI) are all in the predicted direction and significant. The coefficient of the book value of equity (BVE) is in the right direction but not significant, whilst the vesting period (LIFE) and risk-free interest rate (INT) have the wrong sign and are insignificant. Overall the coefficient of determination (R2 ) is 0.336 and the F ratio is significant. The results from the second-stage regression for the pooled sample are shown in Table 7. Consistent with Table 5, the coefficient on the book value of equity is significantly positive at the 5% level for the pooled sample (and is significantly positive in 2001 and 2004). The coefficient on net income is significantly positive at the 1% level for the pooled sample (and is significantly positive in 1999, 2000, 2002, 2003 and 2004). The coefficient on the estimated value of the options is positive for the pooled sample only. This result may be driven by the distribution of firms across the reward and incentive groups during these years. The distribution of firms in 1999, 2000 and 2002 between positive and negative prior period profit and return on assets is very unbalanced. Similarly, the distribution in these years of firms between exercise price groups (discount, fair market value and premium) is also very unbalanced. The small number of

103

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

Table 7 Summary of the 2SLS regression coefficients for the relation between share price and option value measured using an instrumental variables approach (pooled sample and by year). The coefficients are based on the following equation: Pit ¼ c1it þ c2 BVEit þ c3 NIit þ c4 PREDOPTit þ eit . Variables

Predicted sign

n Intercept

?

Pooled sample

1999

2000

2001

2002

2003

2004

259

58

58

29

47

41

26

1.075 (1.321)

0.583 (0.466) 0.925 (0.864)

1.742 (2.215**) 4.911 (4.573***)

0.281 (0.295)

0.427 (0.662)

0.489 (0.508)

0.069 (0.082)

2.946 (0.517)

16.765 (2.375**) 291.517 (1.307) 0.612

0.714 (2.278** 1.225 (4.203***) 2.688 (1.933*) 72.205 (0.670) 0.650

22.016***

16.442***

0.185 (0.224) **

BVEit

+

0.939 (2.206 )

0.726 (1.260)

NIit

+

PREDOPTit

?

11.339 (2.978***) 164.706 (1.074) 0.544

18.681 (2.981***) 0.640 (0.002)

319.637 (1.168)

Adjusted

14.187 (5.208***) 262.193 (1.843*) 0.582

0.602

0.849

18.142 (2.552***) 560.430 (1.249) 0.601

R2 F ratio

45.879***

23.688***

29.783***

53.364***

24.084***

Pit, average price per share for three months after year end; BVEit, book value of equity per share; NIit, net income per share; PREDOPTit, estimated options per share, value from 2SLS regression. * Significant at the 10% level (2-tailed). ** Significant at the 5% level (2-tailed). *** Significant at the 1% level (2-tailed).

observations in each year may also contribute to the variation in results. Nevertheless, our pooled result is consistent with that of Brown and Yew (2002). 4.2. Prior period performance In order to test Hypotheses 1 and 2 it is necessary to classify the sample firms into the ‘reward’ and ‘incentive’ groups. The first method for classification is firms’ prior period performance. Prior period performance is measured using the Market Model to determine the firm-specific abnormal returns for the year prior to granting options to the CEO. The models are run for positive versus negative prior period abnormal performance and also for abnormal performance quartiles. The distribution of firms across these groups is shown in Table 8. Table 8 shows that more firms experience ‘good’ performance, in terms of market returns, during the prior year, as 61% of the companies fall into the reward group compared to 31% of companies in the incentive group. The results from the two-stage least-squares regression where firms are classified according to prior period abnormal performance are shown in Table 9. Overall the Coefficient of determination (R2 ) varies from 0.729 to 0.491, and all F ratios are significant. There is a positive association between the estimated value of options and price for firms that experienced negative abnormal performance in the period prior to granting options (incentive group). Firms that fall into the 1st quartile of abnormal performance demonstrate a similar positive relation (significant at the 5% level). The coefficient of the net income is positive and significant in all cases. The coefficient of the book value of equity is only significantly positive for the first quartile. Overall our results provide support for Hypothesis 2—there is a positive association between the value of stock options and market values for incentive group firms. The coefficient on the estimated value of options is insignificant for ‘reward’ firms, that is, firms with negative prior period abnormal performance. However, while firms in the 3rd quartile also have an insignificant coefficient on the estimated value of options, interestingly firms in the 4th quartile of abnormal performance demonstrate a positive relation, but this is only at the 10% level so should be interpreted with care. One possible explanation is that the retention effect of these options outweighs the reward effects—that is, these firms have all experienced high prior abnormal returns so to retain the CEO and to retain the same level of effort is of benefit to the firm. However, overall, the results do demonstrate that there are some differences in the way the market prices options for ‘reward’ and ‘incentive’ firms.

Table 8 Distribution of firms across reward and incentive groups based on prior period abnormal performance. 2000

2001

2002

2003

2004

Total

%

Prior period abnormal performance Positive (reward) 33 Negative (incentive) 25

1999

28 30

18 11

38 9

26 15

14 12

157 102

61 39

Total

58

29

47

41

26

259

100

58

Prior period abnormal performance = prior period firm-specific stock return measured using the Market Model.

104

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

Table 9 Summary of the 2SLS regression coefficients for the relation between share price and option value (prior period abnormal performance). The coefficients are based on the following equation: Pit ¼ c1it þ c2 BVEit þ c3 NIit þ c4 PREDOPTit þ eit . Variables

Predicted sign

n

Negative ABN performance Incentive 102

Positive ABN performance Reward 157

Quartile 1

Quartile 2

Quartile 3

Quartile 4

Incentive 65

Incentive 65

Reward 65

Reward 64

Intercept

?

1.416 (2.065**)

1.013 (1.054)

2.354 (2.355***)

0.729 (1.013)

0.281 (0.144)

1.259 (0.992)

BVEit

+

0.925 (5.269***)

0.859 (1.408)

0.200 (0.257)

1.240 (1.072)

0.895 (1.164)

NIit

+

4.281 (3.805***)

18.772 (4.628***)

17.486 (2.841***) 186.623 (0.553) 0.600

18.536 (2.334***) 315.842 (0.924) 0.617

16.342 (3.735***) 319.241 (1.9887*) 0.552

12.994***

13.875***

10.709***

186.012 (2.243 )

294.896 (1.601)

Adjusted

0.729

0.635

0.704 (2.703***) 3.816 (2.958***) 230.765 (2.179**) 0.491

R2 F ratio

35.019***

34.876***

8.720***

PREDOPTit

+ incentive  reward

**

Pit, average price per share for three months after year end; BVEit, book value of equity per share; NIit net income per share; PREDOPTit, estimated options per share, value from 2SLS regression. * Significant at the 10% level (2-tailed). ** Significant at the 5% level (2-tailed). *** Significant at the 1% level (2-tailed).

4.3. Option characteristics The second way in which this study classifies the sample firms into the ‘reward’ and ‘incentive’ groups is according to the characteristics of the options granted. The models are run for firms that grant options with discount or fair market value exercise prices versus firms that grant options with premium exercise prices. The distribution of firms across each of these groups is shown in Table 10. Table 10 shows that there is substantial variation in the characteristics of options granted by sample firms. Just over half the firms (53%) grant options with exercise price equal to grant date stock price (fair market value). However, a substantial portion of firms grant options that are already in-the-money (discount, 18%) or at premium (30%). This heterogeneity supports the notion of the importance of studying option grants in a unique setting such as Australia where such variation exists. The distribution of firms shown demonstrates the differences between Australia and the US, where Murphy (1999) reports that the overwhelming majority of firms grant options with at fair market value. In fact, only 5% of firms in Murphy’s sample of 853 option grants firms were at discount or at premium, while the remaining 95% were granted at fair market value. The results from the two-stage least-squares regression where firms are classified according to exercise price are shown in Table 11. The models display good fit, with an adjusted R2 of 44% and 61.9%, respectively, and both F ratios significant at the 1% level. There is a significant positive relation between price and the estimated value of options for firms that grant options with a premium (‘incentive’ firms). This relation is insignificant for firms that grant options with discount or at fair market value (‘reward’) firms. The coefficient on the book value of equity is positive for reward firms and the coefficient on net income is significantly positive for both groups of firms. The significant positive relation between the estimated value of options and share price for ‘incentive’ firms provides some support for Hypothesis 2, although the significance is low so the results should be interpreted with care. The absence of a significant negative relation between the estimated value of options and share price for ‘reward’ firms suggests that the empirical results do not provide support for Hypothesis 1. Overall these results confirm the findings of Table 9.

Table 10 Distribution of firms across reward and incentive groups based on option exercise price. 1999

2000

2001

2002

2003

2004

Total

%

Option exercise price Discount (reward) FMV (reward) Premium (incentive)

8 35 15

7 42 9

6 10 13

10 27 10

9 16 16

6 6 14

46 136 77

18 53 30

Total

58

58

29

47

41

26

259

100

Discount, option exercise price is less than the firm’s share price on grant date; FMV, option exercise price equals the firm’s share price on grant date; Premium, option exercise price is greater than the firm’s share price on grant date.

105

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

Table 11 Summary of the 2SLS regression coefficients for the relation between share price and option value (option exercise price). The coefficients are based on the following equation: P it ¼ c1it þ c2 BVEit þ c3 NIit þ c4 PREDOPTit þ eit . Variables

Predicted sign

Premium Incentive 77

Disc + FMV Reward 182

Intercept

?

0.523 (0.599)

0.171 (0.175)

BVEit

+

0.415 (0.602)

0.961 (1.879*)

NIit PREOPTit

+ + incentive  reward

10.242 (2.561**) 379.737 (1.894*)

16.216 (4.580***) 243.770 (1.296)

n

Adjusted R2 F ratio

0.440

0.619

8.461***

37.806***

Pit, average price per share for three months after year end; BVEit, book value of equity per share; NIit, net income per share; PREDOPTit, estimated options per share, value from 2SLS regression. * Significant at the 10% level (2-tailed). ** Significant at the 5% level (2-tailed). *** Significant at the 1% level (2-tailed).

Table 12 Subsequent performance of firms, groups based on exercise price splits. Incentive

ROAt+1 ROAt+2 ROAt+3 ROAt+4 ROAt+5 ROAt+6 AB_PERFt+1 AB_PERFt+2 AB_PERFt+3 AB_PERFt+4 AB_PERFt+5 AB_PERFt+6

Reward

n

Mean

Median

Standard deviation

n

Mean

Median

Standard deviation

67 66 59 55 40 27 67 65 59 53 35 23

0.064 0.055 0.062 0.072 0.088 0.089 0.159 0.050 0.147 0.006 0.002 0.100

0.061 0.059 0.061 0.072 0.072 0.081 0.174 0.002 0.026 0.052 0.064 0.068

0.062 0.068 0.074 0.068 0.085 0.061 0.389 0.407 0.528 0.312 0.268 0.460

157 145 135 120 101 73 171 158 151 143 126 99

0.043 0.035 0.036 0.032 0.013 0.001 0.120 0.001 0.021 0.120 0.031 0.105

0.057 0.057 0.062 0.066 0.072 0.060 0.095 0.048 0.112 0.026 0.041 0.033

0.102 0.116 0.116 0.145 0.190 0.222 0426 0.517 0.377 0.363 0.345 0.383

t-Test (t)

Mann–Whitney U

1.860* 1.515 1.868* 2.435** 3.233*** 3.164*** 0.656 0.702 1.925* 0.243 0.471 0.049

2050 4607 3880 2931 1787 698** 5395 4589 4342 3568 2035 1125

ROA, return on total average assets; ROE, return on total average equity; EPS, earnings per share; AB_PERF, abnormal performance, calculated using Market Model. Significant at the 10% level (2-tailed). ** Significant at the 5% level (2-tailed). *** Significant at the 1% level (2-tailed). *

4.4. Sensitivity tests Several sensitivity tests are used to investigate alternative methods of options valuation and to explore alternative proxies for classifying the sample into the ‘reward’ and ‘incentive’ groups.20 In order to test the robustness of the results, the ordinary least-squares regression in Eq. (1) is re-run with the options valued at 25% of exercise price. This option valuation method has been shown to have a relatively high level of correlation with the Black–Scholes valuation model (Core et al., 1999; Brown and Yew, 2002). The results support the main findings— there is a positive significant coefficient (at the 1% level) on the estimated value of the options for the incentive group, but the same coefficient in the reward group is insignificant. Several alternative proxies are used for classifying firms into the ‘reward’ and ‘incentive’ groups. As alternative measures of prior period performance this study also examines prior period profit and prior period return on assets. In these circumstances a firm is considered to be granting options for reward purposes if it reported a profit or experienced positive return on assets in the prior period, respectively. If a firm reported a loss in the prior period or experienced negative return on assets then it is classified into the ‘incentive’ group. The results from the two-stage least-squares regression based on prior period profit and prior period return on assets provide support for Hypothesis 2.

20

Results from sensitivity tests are available upon request.

106

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

4.5. Some further tests To confirm the validity of our classifications and to provide further evidence on the reward versus incentive groups of firms, we compare the grouped firms’ subsequent performance both in terms of accounting (return on assets, ROA) and market-based performance measures (abnormal market performance, AB_PERF).21 Table 12 depicts the results. Our results, based on ROA, provide some support for the view that incentives work as the incentive group of firm’s ratio is significantly higher than that of the reward group.22 The market-based results are less conclusive, as there is only one year (t + 3) when the abnormal returns of the incentive group is significantly higher than the reward group. 5. Discussion and conclusion This study has expanded on current knowledge by providing some evidence on how the relation between stock-based compensation expenditure (in the form of stock options) and market values differs, according to the primary reasons for granting the options. There are two main motivations for this paper. First, the extant literature that tests the association between the value of stock-based compensation and market values has produced inconsistent results (Aboody, 1996; Rees and Stott, 2001; Bell et al., 2002; Brown and Yew, 2002; Aboody et al., 2004b). These mixed results suggest that there may be an explanation for why both the negative and positive relations between stock-based compensation expenditure and market values can be consistent with rational markets. The second motivation is that Australia is a unique setting in which to study this association because of substantial heterogeneity in the characteristics of options granted by Australian firms (Coulton and Taylor, 2002; Matolcsy and Wright, 2007), which is very different to the homogeneity witnessed in the United States (Murphy, 1999). This study makes some key contributions to the literature as it provides some theoretical development on the nature of the incentive and dilution effects of stock-based compensation. Whilst the actual motivations behind firms granting stock options to their CEOs remain unobservable, it is possible to classify firms according to a number of proxies, including prior period performance and option characteristics. Limitations in the data reduce the generalisability of the results produced in this study. In particular, the study is limited due to the small sample size that is employed. This is due to the size of the Australian market and that only around one third of Australian top 500 firms grant options in any given year (Matolcsy and Wright, 2007). There are a number of implications for further research from this work. First, it would be interesting to investigate the differences between ‘reward’ and ‘incentive’ firms with larger samples over a longer period and/or in another national setting. Further refinement of these proxies and development of different proxies to classify the sample firms would be useful. Second, due to data limitations23 this study does not test the different types of performance hurdles in place and whether this places them in the ‘reward’ or the ‘incentive’ group. 5.1. Conclusion This study investigates the relation between stock-based compensation expenditure and market values. The paper provides a plausible explanation for why both negative and positive relations between stock-based compensation expenditure and market values can make economic sense. The results show that there is a significant positive relation between stockbased compensation and market values for firms identified as granting options to their CEOs as an incentive. However, our evidence does not support the hypothesised negative relation between option grants and market values for the reward group of firms but rather shows that on the whole, there is no relation between option grants and market values for the reward group of firms. This may indicate that the retention benefits of these options balance the expense effect of the option grants. Acknowledgements The comments of our anonymous referee, Dan Dhaliwal, Bin Srinidhi and Peter Wells and the participants of the UTS Research Seminars in Sydney, and the European Accounting Conference (2006) in Dublin are gratefully acknowledged. References Aboody, D., 1996. Market valuation of employee stock options. Journal of Accounting and Economics 22, 357–391. Aboody, D., Barth, M.E., Kasznik, R., 2004a. Firms’ voluntary recognition of stock-based compensation expense. Journal of Accounting Research 42 (2), 123– 150. 21

These measures based on Larcker et al. (2007). Although it is also possible that these groups of firms have greater incentive to manage their earnings. However, to check for earnings management is beyond the scope of this paper. 23 Full hurdle details are a voluntary disclosure in the financial statements. Since not all firms made this disclosure, the sample of firms that disclosed full details regarding the nature of the performance hurdles attached to options granted is extremely small in this date set. 22

Z. Matolcsy et al. / Journal of Contemporary Accounting & Economics 5 (2009) 95–107

107

Aboody, D., Barth, M.E., Kasznik, R., 2004b. SFAS No. 123 stock-based compensation expense and equity market values. The Accounting Review 79 (2), 251– 275. Balsam, S., Bartov, E., Yin, J., 2004. Disclosure versus Recognition of Option Expense: An Empirical Investigation of SFAS No. 148 and Stock Returns. Working Paper, New York University. Bell, T.B., Landsman, W.R., Miller, B.L., Yeh, S., 2002. The valuation implications of employee stock option accounting for profitable computer software firms. The Accounting Review 77 (4), 971–996. Bebchuk, L., Fried, J., Walker, D., 2002. Managerial power and rent extraction in the design of executive compensation. University of Chicago Law Review 69 (3), 751–842. Brown, P., Yew, E., 2002. How do investors regard ESOs? Australian Accounting Review 12 (1), 36–42. Carter, M.E., Lynch, L.J., 2001. An examination of executive stock option repricing. Journal of Financial Economics 61, 207–225. Carter, M.E., Lynch, L.J., 2004. The effect of stock option repricing on employee turnover. Journal of Accounting and Economics 37, 91–112. Chalmers, K., Godfrey, J.M., 2005. Expensing stock-based payments: a material concern? Journal of International Accounting, Auditing and Taxation 14, 157– 173. Core, J., Guay, W., 1999. The use of equity grants to manage optimal equity incentive levels. Journal of Accounting and Economics 28 (2), 151–184. Core, J.E., Guay, W.R., Larcker, D.F., 2003. Executive equity compensation and incentives: a survey. Economic Policy Review, 27–50. Core, J.E., Holthausen, R.W., Larcker, D.F., 1999. Corporate governance, chief executive officer compensation and firm performance. Journal of Financial Economics 51, 371–406. Coulton, J., Taylor, S., 2002. Option awards for Australian CEOs: the who, what and why. Australian Accounting Review 12 (1), 25–35. Espahbodi, H., Espahbodi, P., Rezaee, Z., Tehranian, H., 2002. Stock price reaction and value relevance of recognition versus disclosure: the case of stockbased compensation. Journal of Accounting and Economics 33, 343–373. Frederickson, J.R., Hodge, F.D., Pratt, J.H., 2006. The evolution of stock option accounting: disclosure, voluntary recognition, and management disavowals. The Accounting Review 81 (5), 1073–1093. Goodwin, J., Kent, P., 2004. The incentives of Australian companies to utilize executive stock option plans. Corporate Ownership and Control 1 (2), 137–149. Guay, W., Kothari, S.P., Sloan, R., 2003. Accounting for employee stock options. AEA Papers and Proceedings 93 (2), 405–409. Hanlon, M., Rajgopal, S., Shevlin, T., 2003. Are executive stock options associated with future earnings? Journal of Accounting and Economics 36, 3–43. Hodder, L., Mayew, W.J., McNally, M.L., Weaver, C.D., 2006. Employee stock option fair-value estimates: do managerial discretion and incentives explain accuracy? Contemporary Accounting Research 23 (4), 933–975. Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3, 303– 360. Johnston, D., 2006. Managing stock option expense: the manipulation of option-pricing model assumptions. Contemporary Accounting Research 23 (2), 394–425. Larcker, D.F., Richardson, S.A., Tuna, I., 2007. Corporate governance, accounting outcomes and organisational performance. The Accounting Review 82 (4), 963–1008. Matolcsy, Z.P., Wright, A.P., 2007. Australian CEO compensation: the descriptive evidence. Australian Accounting Review 43 (17 3), 47–60. Murphy, K.J., 1999. Executive compensation. In: Ashenfel, O., Card, D. (Eds.), Handbook of Labor Economics, fifth ed. North-Holland, pp. 2485–2563. Ohlson, J.A., 1995. Earnings, book values, and dividends in equity valuation. Contemporary Accounting Research 11 (2), 661–687. Rees, L., Stott, D.M., 2001. The value-relevance of stock-based employee compensation disclosures. The Journal of Applied Business Research 17 (2), 105– 116. Skinner, D.J., 1996. Are disclosures about bank derivatives and employee stock options ‘value-relevant’? Journal of Accounting and Economics 22, 393–405.