Disclosure Factors of Executive Managers Remuneration: A Probit Model

Executive managers’ remuneration has been an issue of debate for the last 30 years. Practitioners and academics are arguing for the mechanisms, mix, level, time horizon and goal. Disclosure or not of information regarding these issues preoccupies regulating, legislative authorities as well as capital market participants. The issue of remuneration is considered to be closely connected with financial performance (positively), firm size (positively), the organizational structure (negatively) and corporate governance mechanisms (negatively). Furthermore, a connection of ownership structure and executives’ remuneration has been well established (theoretically and empirically) in the literature (agency theory). The paper, using a Probit regression analysis, examines whether these relationships are valid in Greece. Greece hasn’t the characteristics of an Anglo-Saxon country. Overall the study has shown that remuneration levels in Greece are defined by a different set of factors than the ones that are prominent in an Anglo-Saxon country. The major factors that affect the disclosure of information about the remuneration levels are the adoption of mergers and acquisitions as the method to expand firm’s size, the investments risks that the firm is willing to take, stock market capitalization, board of directors size, capital to sales ratio, number of independent member of the board of directors dismissals, and the quality of corporate governance. These factors indicate that Greek firms are disclosing information about the remuneration levels when the investment effort and the quality level of corporate governance are high.


Introduction
Executive interest alignment is the focal point of agency theory. The basic proposed mechanism to solve the issue is remuneration (Grossman, and Hart, 1982;Jensen, 1986). In order to establish a rationality in formulating remuneration policies the firm has to establish a method to verify the value of managers (Petra, 2005), managers' and directors' contribution to financial performance (Letza and Kirkbride et al., 2008;Conyon et al., 1995;Gregg et al., 1993;Hassan, Christopher, Evans, 2003) and the overall value of the firm (Habib and Ljungqvist, 2005;Stulz, 1990). From the perspective of stewardship theory (Donaldson, 1990, p. 375) managers are not motivated only or mainly through remuneration. Other researchers (i.e. Petra, 2005) argue that it is necessary to enforce managers in order to enforce productivity. Remuneration control is exerted by the Annual Shareholders Meeting, the Board of Directors or by any committee that has been introduced to control and evaluate executive managers and their performance. The efficiency of these mechanisms has been the focal point of many studies (Petra, 2005;Conyon and Peck, 1998).
Some research findings (Bebchuk and Fried, 2004, p. 2) showed that managerial power has dominated the process of negotiation for remuneration levels. BoD is responsible for determining these levels and the schemes of remuneration. "In light of the historically weak link between non-equity compensation and managerial performance, shareholders and regulators wishing to make pay more sensitive to performance have increasingly looked to, and encouraged, equity-based compensation-that is, compensation based on the value of the company's stock" (Bebchuk and Fried, 2004, p. 7).
The theory that executives and directors should be motivated to align their interests with the shareholders' interests, has led to a quintuple (Cassidy, 2002) of executive remunerations in a decade (1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001) and the disclosure of frauds. On the other hand executives are willing to invest free cash flows ineffectively, to retain the capital assets within the firm, rather than to distribute them to shareholders (Hellwig, 1998). The basic motive for the executives is the dominance in corporate power game. Dominance guarantees high remuneration and entrenchment.
Agency theory addresses the issues that arise from organizational structure of firms that follow the Anglo-Saxon firm characteristics. There are major differences, relevant to remuneration, between the Anglo-Saxon system and the one in the Continental Europe (Weimer and Pape, 1999) are: a) markets for corporate control, capital and labor market for directors are more active and effective (although there is a growing discussion about how efficient they are). Furthermore, executive managers may entrench themselves in their positions, making it difficult to oust them when they perform poorly , b) As Shleifer and Vishny (1997) argue in Anglo-Saxon countries, capital providers need specialized human capital and executives need capital providers, because they do not have enough capital themselves. On the contrary, in Continental Europe countries executive directors are capital providers and in many cases, members of the dominant group of stakeholders and c) The presence of a large shareholder is likely to result in closer monitoring and reduce of executive directors' power to impose the pursuit of their interests (Shleifer and Vishny, 1986).
In Continental Europe countries the fact that major shareholders are members of the BoD, CEOs and Presidents of the BoD, reduces the possibility of monitoring and transparency. These members have triple attributes or roles (major shareholder -part of the dominant group, member of the BoD and CEO -President of the BoD). Greece is a typical Continental Europe system's country.

Corporate Governance status in Greece
Greek firms are mainly family or controlled by a group of stockholders (Mavridis, 2002). Free float is relatively small in percentage (20-50%) and the ability to achieve control through the capital market is limited. The members of the family or the controlling group are actively involved in management and normally, there is no distinction between management and ownership. The Board of Directors can be characterized as one tier. Managers that are not members of the family or the controlling group are closely connected with these groups and their decisions are subject to their control and monitoring. Institutional investors, although the catalyst for the adoption of CG mechanisms, have not actively been involved in management or in controlling and monitoring the decisions and actions of the controlling group.
Mertzanis (2001) (before the new law for the CG in Greece was enacted) noted: "the prevailing framework of corporate governance in Greece is not simply considerably outdated, but may cause potential problems, due to inadequate transparency and accountability, regarding the provision of cost-efficient finance that is required to increase investment and raise national competitiveness". So the Hellenic Capital Market Committee (2000) and the Committee on Corporate Governance have made 44 basic recommendations (compiled in seven main categories: rights and obligations of shareholders; the equitable treatment of shareholders; the role of stakeholders in corporate governance; transparency, disclosure of information and auditing; the board of directors; the non-executive members of the board of directors; Executive management. They have also proposed the adoption of IAS (now IFRS)). Only a small number of these recommendations have been adopted and introduced. Spanos (2005) notes that "the majority of medium and small capitalization (family-owned) companies have adopted the minimum mandatory requirements and lack further efficient CG mechanisms. As long as the competition for capital is increasing, listed companies have to realize that proper CG is a prerequisite in order to attract international capital. Moreover, corporate governance may meet one of the most significant challenges that family-run businesses face: management succession". The need for CG mechanisms is identified by all market participants as a substitute for trust (as a bonding and problem solving element) among the major stockholders or family members, but they cannot agree on what the mechanisms/processes will be. Also, there are strong resistive forces mainly by the major stockholders/family members who are not willing to pass power and information to "non-trust worthy" stockholders or professional executive managers. As a result the governing/administrative bodies do not function according to statutes or laws and the process that they provide, but according to the common will of the family members. Furthermore, an effective market for corporate control does not exist.
Τhe BoD is mostly acting as a passive body in the company where it follows the decisions of the management. Non-executive board members, rather than act as shareholders' agents, do not efficiently supervise the management (Schulze et al., 2003). This is the case in the majority of (family) public companies in Greece, where significant costs result from bias in favouring family interests over the firm's interests (such as non-family shareholders), because of loyalty toward the family (Schulze et al., 2003). Even though the rules mandate specific requirements regarding board independence, it's difficult in practice to identify whether the board meets these rules (Spanos (2005). In countries with concentrated ownership structure (continental Europe, Japan and other OECD countries), large dominant shareholders usually control managers and expropriate minority shareholders, in order to extract private control benefits. The question is therefore posed as how to align the interests of strong block-holders and weak minority shareholders (Spanos, 2005, p. 16;Becht, 1997).
On the other hand, investors usually use their exit options if they disagree with the management or if they are disappointed by the company's performance, signaling -through share price reduction -the necessity for managers to improve firm performance (Spanos, 2005, p. 16;Hirschman, 1970). The lack of market liquidity creates problems in the effectiveness of the shareholders exit option and governing problems (since the main governing body is the general shareholders meeting, but participation is not an easy task). The cost of involvement with management and control for the minor stockholder is greater than the cost of exit and so they may easily choose to sell their stock ("they vote with their feet") if they are not content with the managements' choices. The shareholders encirclement does not necessarily mean participation in the company administration. Moreover, family firms disclose less narrative information than non-family firms, where family-firms may disclose more information than non-family firms in some selected areas of interest, such as data information about share price policy and number of diagrams used in the interim report (Mavridis, 2002). In countries where business has traditionally been based on relationship and trust, corporate information is thought of as secret; and it is accepted practice to keep different sets of books, e.g. one for taxes, one for outside investors, and one for the majority shareholder (Fremond and Capaul, 2002, p. 18). There is a vicious circle whereby managers consider secrecy as imperative so that shareholders do not vote with their feet and through it they can cover up their lack of efficiency or impotence; minority shareholders (major shareholders already have the information because they are members of the BoD, management or the relevant cost for them is not too high) do not actively demand information because the cost of acquiring and processing it is too high for them.
Remuneration is considered to be closely connected with financial performance (positively), firm size (positively), the organizational structure (negatively) and corporate governance mechanisms (negatively). Furthermore, a connection of ownership structure and executives' remuneration has been well established (theoretically and empirically) in the literature (agency theory). The paper argues that the disclosure of remuneration levels is dependent on variables that are relevant to internal and external mechanisms of CG, performance and growth perspectives.

Sample -Methodology
The study's time horizon is from 2001 to 2006. Sixty firms, that are ranked in the two major stock indexes (FTSE-20 and FTSE-40) of the Greek Capital market and they are consider to be the biggest firms in terms of capitalization and with the highest free float, are used. Their annual reports are the basic source for the data collection. The data was supplemented by information collected by the corporate web sites. Total sample size is 303 observations. Although remuneration disclosure is mandatory, from the 303 available annual reports only 109 contain information about the executive board members. This is a strong indication of the trend to conceal "sensitive" information. As Bebchuk and Fried (2003) argue that executives have an incentive to "camouflage" their remunerations, in order to minimize the "outrage" of outsiders. In this case it's the major shareholders that conceal information.
To address the issue limited variable models (Probit) can be used. The Probit model has some significant statistical problems (normality of residuals, heteroscedasticity) and the usual measures of fitness are inefficient. One of the main advantages of probit models is that they allow the use of panel data and also they can take into account the factor of time. A correlation matrix has shown that independent variables are not correlated in a manner (-0,3>=r<=0,3) that may create problems of result reliability and colinearity. The use of panel variables helps to identify the quality variables that formulate the depended variable. Finally, marginal effects methodology can be applied. Four variables for the construction of panels were used in the present research (see Table 1).

Model construction
The dependent variable (E_REM) records the decision of the firm to disclose its executives' remunerations. What has been recorded in the study is the cash -salary payments made to the executives. No other way of remuneration (e.g. stock options) could be tracked through annual reports. This may result to the omission of some of the remuneration mechanisms. The omission of these mechanisms, although is important in the Anglo-Saxon countries, in countries like Greece these mechanisms are not widely used, which is common in the Continental Europe system countries (Kakabadse and Kakabadse, 2001, p. 65).
The model is:  (1) Where: i = 1 … Ν, t = 1 … T Disclosure of remuneration of the executives although it is an administrative, but not voluntary decision (Greek law (Law 3016/2002) has relevant provision), is a strong indication of transparency. Good performing firms and firms with growth prospective do not have the incentive to withhold this information. Firms that are managed by professionals have the incentive to link remuneration with performance, as the agency theory suggests.
The variables of CG quality index (for the construction of the index see Lazarides and Drimpetas, 2008) and the binary value of mergers and acquisitions should have a positive impact on disclosure of remuneration levels. BoD's size (BOD, number of members), independent BoD members are a measure for the determination of the monitoring efficiency. On the contrary, high numbers of executive members (BEXEC) of the BoD and secessions -resigns of board members (BDIS_P), to the total number of board members may lead to lower remuneration disclosure levels.
Firms depended heavily on debt to finance their operations, present high uncertainty and risk for future returns. This, in turn should lead to reduced levels of remuneration disclosures. Large firms have the tendency to depend more on professional managers and hence they have the tendency disclose more their policies and amount regarding remuneration. Concentrated ownership (variables OWN and HERF) has the opposite result. Family or high ownership concentrated firms do not need or they do not have the incentive to release this kind of information to the public. The same conclusions can me drawn for the variance OWNCEO (the biggest shareholder is the CEO).
High investments levels (INVP) are positively related with the disclosure of remuneration levels, if the assumption that investment can lead to better short-term financial results. Otherwise, the relation is negative. Younger firms tend to present higher risks and hence firms are reluctant to have relatively high remuneration levels. As the firm grows old it loses the initial family characteristics due to diffusion of shares (through IPO's, capital increases and succession) and so as firm's grow older there is a higher reliance on professional manages, rather than dominant stakeholders.

Findings -Statistical results
The sample was divided in two main categories. The first one is the observations of the firms that are ranked in the FTSE-20 index (the biggest 20 firms in terms of capitalization) of the Athens Stock Exchange and the second is the observations of the firms that are ranked in the FTSE-40 index (the next 40 firms in terms of capitalization). No significant differences were found.
Disclosure levels are higher in the Non Financial sector (see Table 2). Non Financial firms seem to disclose more information than the financial firms. As Table 3 depicts, firms with higher ownership concentration, better corporate governance level (Lazarides and Drimpetas, 2008) and better Tobin's Q, seem to disclose more information. Time variance of disclosure is also very interesting. CG seems to affect gradually disclosure. Another point that should be noted is that when disclosure is stratified with ROA then in years 2003 and 2004 disclosure is greater than in the previous and following years. This behavior can be explained if the investment schedules of the firms are taken into account. The fitness of the model, is satisfactory (McFadden Pseudo R-square is 0,2494, Log Likehood Function is -148,58 and Estrella is 0,3125). The model can predict the 61,468% (67/109) of the cases that there is disclosure (E_REM variable = 1) and the 86,08% (167/194) of the cases that no disclosure took place (E_REM variable = 0).
Chi square tests indicate that at least one of the independent variables is statistically significant (Χ 2 [18] (prob) = 98,71 (.0000)). The test to identify which independent variables are statistical significant (at the level a=0,05 or a=0,10) are shown in the Table 4 that follows. Twelve out of 19 variables seem to be statistical significant. An interesting statistical finding is that the Constant term of the Model is not statistically significant. This means that if all independent variables are equal to zero, then the dependent variable is most likely zero as well. This finding is important because it proves that if all other variables are constant there is a tendency to not release any information regarding the executives' remuneration. The model was further processed to reduce the independent variables, in order to contain only the variables that are statistically significant. The fitness of the final model, is satisfactory (McFadden Pseudo R-square is 0,22, Log Likehood Function is -154,33 and Estrella is 0,2776). The model can predict the 54,13% (59/109) of the cases that there is disclosure (E_REM variable = 1) and the 85,57% (166/194) of the cases that no disclosure took place (E_REM variable = 0). The combined total prediction capability of the model is 74,26%. Nine out of 19 variables seem to be statistical significant. Finally, the data were regressed using marginal effects. Marginal effect is the change of possibility due to the change of the independent variable by one unit. Two variables were used to measure their marginal effects on the dependent variable (OwnCEO, CEOCHAIR). The analysis of marginal effects is shown at Table 6. As the Table shows the combination of role of shareholder and CEO and the duality of roles of the chairman of the BoD and CEO contribute to disclosure of executives' remuneration. This finding is a paradox. The analysis of CG status, ownership concentration and organizational structure may lead the reader to the conclusion that the concentration of power and control when a person holds two position of power will give the opportunity and capability to these persons to withhold information and to create opacity for the firm's activities. The paradox can be explained by the fact that most of the executives belong to the dominant group or they are closely connected with it. So the firm hasn't anything to lose. On the contrary the firm has the opportunity to show to the other stakeholders that the firm is upholds the CG principles of transparency, responsibility and accountability.

Discussion -Conclusions
One major finding of this study is that only 36% (109/303), breaking the law because disclosure is mandatory, of the firms have disclosed in their annual reports the remuneration levels of their executive members. The selection of non disclosure is conscious. Major shareholders, groups of shareholders and families are the dominant stakeholders in the firm. They are unwilling to release information that may shake the status quo or question their power to make decisions.
The fact that nine variables were found to be statistically significant shows that the phenomenon of remuneration disclosure is complex. Statistically significant variables may be grouped into three major groups. The first group is firm's size (TA, SMCAP, OC_S2), the second the Board of Directors (BOD, BIND, BDISI_P) and the third (MERGER, INVP, CG) the external and growth factor group.
Financial performance was measured with many variables (PREPL, TQ, PE, ROA), but none of them were found to have significant impact on remuneration disclosure. The lack of connection between them is contrary to agency problem theory. Although it seems that stewardship theory might be more valid in the case of Greece, the fact that the majority of the executives are members of or closely connected with the dominant group and those that are not members of the group are directly dependent on it to retain their position; creates a top management team that its main concern is not to produce good financial results but to keep the status quo. This is contrary to the stewardship theory as well. Lazarides and Drimpetas (2009) in their empirical study have shown that the level of remuneration in Greece is strongly dependent on financial performance. There should not be any confusion about these findings. The decision to disclose is very different from the decision to regulate remuneration levels using financial performance as a benchmark to evaluate executives. Mergers and Acquisitions (MA) have a negative impact on remuneration disclosure. The reasons are the same as the ones of financial performance.
Time and corporate governance have a surprisingly positive impact on disclosure. Firms are cautiously honor their legal obligations and testing through time their capability to balance legal obligations, their will to retain status quo and simultaneously to exhibit to the other stakeholders their will to be transparent and accountable.
The Board of Directors, its composition and function play a crucial role in disclosure. The duty of the BoD is to monitor and control executives, to formulate strategy and to serve as an interface with the external environment. The study did not confirm these duties. Larger BoD (BoD0, bigger number of independent board members (BIND) doesn't seem to have the expected positive impact. The only BoD related variable that has the expected effect is the variable of independent board member dismissalresigns. Independent board members need time to affect firm's principles, policies and tactics. Firm's and especially dominant stakeholder group in their effort to comply with the law, have in their boards sufficient