International Journal of Economics and Financial Issues Vol. 3, No. 4, 2013, pp.885-900 ISSN: 2146-4138 www.econjournals.com 885 Board of Directors Independence and Firm Value: Empirical Evidence Based on the Bucharest Stock Exchange Listed Companies Georgeta Vintilă The Bucharest University of Economic Studies, Department of Finance 6, Romana Square, 1st district, postal code: 010374, postal office: 22, Romania. E-mail: vintilageorgeta@yahoo.fr Ştefan Cristian Gherghina The Bucharest University of Economic Studies, Department of Finance 6, Romana Square, 1st district, postal code: 010374, postal office: 22, Romania. E-mail: stefan.gherghina@fin.ase.ro ABSTRACT: This paper examines the influence and causal relationship between board of directors independence, CEO duality, and firm value. By estimating multivariate regression models for panel data, unbalanced, for a sample of companies listed on the Bucharest Stock Exchange, there resulted a positive influence of the percentage of independent directors on firm value, but down to a threshold of their representation of 47.23 percent, whereupon their influence becomes negative. When we employed fixed-effects models, the relationship previously mentioned was not statistically validated. However, the results provide support for a lack of statistically significant relationship between the percentage of non-executive directors and firm value. Besides, by estimating fixed-effects models we found a positive influence of CEO duality on industry-adjusted Tobin’s Q ratio, but not statistically significant when estimating models without cross-sectional effects. The causal relationships between board independence and firm value identified based on Granger causality are not robust. Keywords: independent directors; non-executive directors; firm value; panel data regression models; Granger causality; vector autoregression JEL Classifications: G32; G34 1. Introduction Corporate governance covers all the procedures which ensure the fact that management operates in order to satisfy shareholders’ interests (Kanagaretnam et al., 2007). Thereby, based on its importance within modern corporations, board of directors as internal mechanism of corporate governance was slightly disputed (Iwasaki, 2008; Jensen, 1993), but widely researched and criticized within the pale of specific literature (Adams et al., 2010; Ezzamel, 2005). Agency theory reflects the prevailing approach as regards the investigation towards the relationship between board independence and firm value. This prevalent theory suggests that information asymmetry, as well different aims between principals (shareholders) and agents (management) impose costs on principals when an agent holding discretionary authority pursues his objectives rather than shareholders’ interests. Bebchuk and Fried (2003) noticed that there are not a priori reasons in order to assume the fact that managers’ aim is shareholders’ wealth maximization. According to Byrd and Hickman (1992) and Fama and Jensen (1983), the companies are mitigating their agency costs through implementing an appropriate monitoring system such as efficient oversight of managers by board members. Furthermore, Fama and Jensen (1983) underlined the efficiency towards managers’ monitoring when corporate boards are dominated by outside independent members. Besides, Hossain et al. (2000) mentioned that the value of outside members among boards is determined through their ability against the assessment of corporate performance, whereas insiders cannot often own this virtue. As much, we ascertain a narrowed efficiency of insiders as corporate monitors. International Journal of Economics and Financial Issues, Vol. 3, No. 4, 2013, pp.885-900 886 The aim of this research consists in the empirical investigation of the influence and causal relationship between board of directors independence assessed through the percentage of independent directors, as well the percentage of non-executive directors and firm value. Moreover, we will investigate the impact and causal relationship between CEO duality and industry-adjusted Tobin’s Q ratio employed as proxy for firm value. This study is important since it provides the first empirical results for a sample of companies listed on the Bucharest Stock Exchange (BSE) over 2007-2011. The novelty of this study is depicted by the examination of board independence based on a sample of companies listed on the Romanian Stock Exchange. Notwithstanding, Romania is acknowledged as post-communist country from Eastern Europe, nevertheless scanty regulated concering corporate governance. With respect to the privatisation process, Boycko et al. (1996) advocated the postive effects of this course, respectively the decline of agency problems and the growth of efficiency by way of improvement the monitoring systems. Thus there are provided the required incentives to the agents in order to act properly. Besides, the privatisation process of Romanian companies owned by the State during the communist regime, initiated after 1990, was not entirely accomplished. Likewise, the act of privatisation is endless litigated and bounded to corruption. According to Stulz (2006), when State behaves improper the individuals who control the State actuate in own interests rather than enforcing property rights and facilitating contracting among private parties. The remainder of this paper is organized as follows. Section two presents the regulations towards board independence within the companies listed on the Bucharest Stock Exchange. Section three reviews the literature on the relationship between board independence and firm value and shows the research hypotheses. The research sample alongside all the employed variables and empirical methods are described in Section four, whilst Section five provides the empirical results. Last section concludes the paper. 2. Board Independence Regulations within the Bucharest Stock Exchange Listed Companies We emphasize the prevalence of unitary boards within the companies listed on the BSE. Besides, Law 31/1990 on trading companies points out that corporate boards of directors shall have the following main competencies that may not be delegated to directors: to establish the main directions of activity and development of the company; to establish the accounting and financial control system and to approve the financial planning; to nominate and dismiss directors and to establish their remuneration; to supervise directors’ activity; to prepare the annual report, to organise the General Meeting of Shareholders and to implement its decisions; to introduce the request for opening the insolvency proceedings of the company according to the law on insolvency proceedings. However, directors shall be nominated by the Ordinary General Meeting of Shareholders, except for the first directors which are appointed through the Constitutive Act. The nominees for the positions of directors shall be appointed by the current members of the board or by the shareholders. Moreover, for the duration of the term of office, directors may not conclude an employment contract with the company. If directors were nominated among the employees of the company therefore the individual employment contract shall be suspended over the term of office. In Romania, Corporate Governance principles and recommendations are expressed within the Bucharest Stock Exchange Corporate Governance Code (2008). Thus, the companies admitted to trading on the regulated market of the BSE shall adopt and comply with the provisions out of the Romanian Corporate Governance Code, but on a voluntary basis. It is mentioned that the board structure of the companies should ensure a balance of executive and non-executive directors or independent directors in a flawless framework, insofar that no individual or small group of individuals can dominate the board’s decision making process. Consequently, an adequate number of non- executive directors shall be independent by considering the fact that they do not maintain, nor have recently maintained, directly or indirectly, any business relationships with the issuer or persons linked to the issuer, of such significance as to influence their autonomous judgment. The independence of non-executive directors shall be evaluated based on the following criteria: a non-executive director is not an executive director of the company or of an entity controlled by it and has not been in such a position for the previous five years; he is not an employee of the company or of an entity controlled by it and has not been in such a position for the previous five years; he does not receive and has not received significant additional remuneration from the company or of an entity controlled by it, apart from a fee received as non-executive director; he is not and does not Board of Directors Independence and Firm Value: Empirical Evidence Based on the Bucharest Stock Exchange Listed Companies 887 represent in any way a significant shareholder of the company; he does not have and has not had within the last financial year a significant business relationship with the company or of an entity controlled by it, either directly or as a partner, shareholder, director, or employee of a body having such a relationship; he is not and has not been within the last three years a partner or an employee of the present or former external auditor of the company or of an entity controlled by it; is not an executive director in another company in which an executive director of the company is a non- executive director; he has not served on the board as a non-executive director for more than three terms; he is not a close family member of an executive director or of persons in the situations above mentioned. The Guide for implementing Corporate Governance Code (2010) notice the fact that at least half of the total number of directors should be non-executive and at least a quarter of them shall be independent. Usually, the Chairman of the board is independent. 3. Literature Review and Hypotheses Development Although the relationship between the monitoring function related to corporate boards of directors and firm value has been extensively researched (Berle and Means, 1932; Jensen and Meckling, 1976; Fama and Jensen, 1985), there are not congruent points of view towards the sense of relation between the monitoring role exerted by outside directors and its effect on firm value (Hermalin and Weisbach, 2003; Denis and McConnell, 2003). Thereby, Morck et al. (1988), Hermalin and Weisbach (1991), Mehran (1995), and Klein (1998) reported a lack of a statistically significant relationship between board independence and firm performance, whereas Agrawal and Knoeber (1996) and Bhagat and Black (2001) concluded a negative relationship. Contrariwise, Rosenstein and Wyatt (1990) noticed positive share-price reactions at director appointments on board. However, whilst insiders are an important source of company-specific information for boards, we distinguish different aims against shareholders’ wealth maximization due to the private benefits and lack of independence towards CEO (Raheja, 2005). As opposed Raheja (2005), outside directors confer an independent monitoring, but they own reduced information as regards the constraints and opportunities of the companies. Therefore, as benefits (costs) of monitoring rise, the boards will perform a better (lower) monitoring through the rise (reduction) of outside members. Hermalin and Weisbach (1988) suggested that outside directors are potential sources of counselling, thus being able to enhance the proficiency and expertise of boards. Likewise, Mariolis (1975), Koenig et al. (1979), and Mace (1986) mentioned that outside directors can serve as source of prestige by offering new business contacts. We notice two competing theories regarding board characteristics emphasized by Boone et al. (2007). The inefficient board hypothesis highlights the inefficient organization of boards unless regulations force them to a more efficient size and composition. In fact, boards may be structured either haphazardly or perversely. Unfortunately, if boards are perversely inefficient, its structure will not increase firm value, but will facilitate the extraction of private benefits by managers at shareholders’ expenses. The economic hypothesis or the efficient board hypothesis emphasizes that board structure is endogenous driven by unique specific characteristics to every company and the tradeoff between costs and benefits. Besides, Boone et al. (2007) notice that board structure is endogenous, being required its adjustment according to the characteristics of the environment where the company operates. By investigating the forces that drive board size and composition, Boone et al. (2007) established three non-mutually exclusive testable hypotheses. Thus, the views of Fama and Jensen (1983), Coles et al. (2008a), and Lehn et al. (2009) are reflected through the scope of operations hypothesis which involves that board structure is driven by the scope and complexity of the firm’s operations. The measures of firm scope and complexity including firm size, age, and number of business segments are positively related to board size and the proportion of independent outside directors. As the companies are growing, its boards of directors are extending due to the net benefits of monitoring and specialization by board members. The monitoring hypothesis defined out of research employed by Demsetz and Lehn (1985), Raheja (2005), Gillan et al. (2011), and Harris and Raviv (2008) emphasizes that board size and composition are positively related to private benefits of managers and negatively related to the cost of monitoring. Thereby, board size implies a tradeoff between the benefits and costs of monitoring. From the studies of Hermalin and Weisbach (1998) and Baker and Gompers (2003) was developed the negotiation hypothesis according to which board International Journal of Economics and Financial Issues, Vol. 3, No. 4, 2013, pp.885-900 888 composition emerges from a negotiation between the CEO of the company and its outside board members. According to Duchin et al. (2010), we highlight three broad views concerning the way how boards operate. Thereby, window-dressing view supported by Romano (2005), also called co-option (Coles et al., 2008b), shows the establishment of numerical targets as regards independent directors through the medium of regulations. However, this circumstance will not improve corporate governance since managers can select directors that meet the criteria of independence according to regulatory definitions, but are still excessively congenial to management. Bhagat and Black (2001) reported that low-profitability companies decide to increase the independence of boards, but there is no evidence that this strategy works. Entrenchment view advocates that managers dislike independent boards and look for their dismissal from oversight. Optimization view shows that managers trade off the strengths and weaknesses of inside and outside directors towards counselling and monitoring in order to maximize shareholders’ wealth. Based on these considerations, we draw the first and the second hypotheses which will be tested within current study: H1: The percentage of independent directors has a positive influence on the value of listed companies on the Bucharest Stock Exchange. H2: The percentage of non-executive directors has a positive influence on the value of listed companies on the Bucharest Stock Exchange. CEO duality occurs when the same person holds the functions of CEO and Chairman of the board of directors (Rechner and Dalton, 1991). Therefore, the Cadbury Report (1992), titled Financial Aspects of Corporate Governance emphasized the concentration of power when the roles of CEO and Chairman are combined in one person. Anyway, the Cadbury Report (1992) recommended a division of responsabilities at the head of the company in order to ensure a balance of power and authority, such that no one individual has sovereign powers of decision. Agency theory supports the negative effect of CEO duality due to the limitations of CEOs’ towards the settlement of decisions in shareholders’ interests (Jensen and Meckling, 1976; Fama and Jensen, 1983), whereas stewardship theory maintains the positive effect related to CEO duality and argues the improvement of organizational efficiency thereupon shareholders’ wealth (Miller and Friesen, 1977; Stoeberl and Sherony, 1985; Anderson and Anthony, 1986; Donaldson and Davis, 1991; Finkelstein and D’Aveni, 1994; Dahya et al., 1996; Brickley et al., 1997; Bhagat and Black, 2001). According to Berg and Smith (1978), the CEOs which hold the position of Chairman of the board can select board members among the individuals with a similar attitude. Thus, Jensen (1993) argued that CEO duality provides an excessive power to the CEO as against decision taking process, as well the possibility to accomplish their own aims. Iyengar and Zampelli (2009) mentioned that CEO can lower the monitoring power related to the board, following the fulfillment of his objectives without taking in account the costs bore by shareholders. On the contrary, Miller and Friesen (1977) stressed the ability of the companies towards the quick act at outside events, respectively the efficiency of decision taking system if both functions are held by a single person. Likewise, Finkelstein and D’Aveni (1994) promoted the unification of CEO and Chairman roles in order to avoid confusions or ambiguities related to multiple authorities. On the basis of these considerations we state the third hypothesis which will be tested within current study: H3: The separation of roles related to the CEO and Chairman of the board has a positive influence on the value of listed companies on the Bucharest Stock Exchange. 4. Data and Research Methodology 4.1 Sample selection and variables description Our initial sample comprised all the companies listed on the Bucharest Stock Exchange on all three tiers between 2007-2011. Subsequently, we removed from our sample the companies from financial intermediation sector (eleven companies) including credit institutions (three banks), financial investment companies (five SIFs), and financial investment services companies (three SSIFs), since these companies are regulated by specific rules. Likewise, we dropped from the initial sample the companies out of ‘Unlisted’ tier (twenty five companies) and the companies out of ‘International’ tier (two companies). Therefore, our final sample shows the following distribution: 63 companies in 2007, Board of Directors Independence and Firm Value: Empirical Evidence Based on the Bucharest Stock Exchange Listed Companies 889 67 companies in 2008, and 68 companies between 2009-2011, counting 334 statistical observations. The industry membership of selected sample is sundry as following: wholesale/retail, construction, pharmaceuticals, manufacturing, plastics, machinery and equipment, metalurgy, food, chemicals, basic resources, transportation and storage, tourism, and utilities. Table 1 describes all the variables employed in the empirical research. Table 1. Description of variables Variable Definition Variable regarding firm value QAdj Industry-adjusted Tobin’s Q ratio. Tobin’s Q ratio was computed as the market value of assets divided by the book value of assets, where the market value of assets equals the book value of assets plus the market value of common equity less the sum of the book value of common equity. Variables regarding corporate board of directors independence IND The ratio between the number of independent directors on corporate board of directors and the total number of directors on board (%). IND2 The percentage of independent directors on corporate board of directors squared (%). NED The ratio between the number of non-executive directors on corporate board of directors and the total number of directors on board (%). NED2 The percentage of non-executive directors on corporate board of directors squared (%). Variable regarding CEO duality CEODual Dummy variable If the CEO holds simultaneously the positions of CEO and Chairman = 1; If the CEO does not hold simultaneously the position of CEO and Chairman = 0. Control variables FS Firm Size, as the annual total assets (logarithmic values). Lev Leverage, computed as debt/book value of assets. SGrowth Sales Growth, as the relative increase of sales from the previous year (%). Listing The number of years since listing on the BSE (logarithmic values). Source: Author’s processing. All data was hand collected, besides the source of it being represented by the annual reports disclosed by the companies. The value of selected companies will be measured through Tobin’s Q ratio, but industry-adjusted, similar Eisenberg et al. (1998), in order to account for the varied industry membership. Thereby, we employed the method described by Kaplan and Zingales (1997), Gompers et al. (2003), and Bebchuk et al. (2009). Nevertheless, we have not considered the market value of debt at the numerator, respectively the replacement cost of assets at denominator. Further, after we have computed Tobin’s Q ratio for each company, we have adjusted it according to industry membership. Thus, the difference between Tobin’s Q ratio of a certain company and industry’ median Tobin’s Q ratio is ∆Q, while industry-adjusted measure of Tobin’s Q ratio (QAdj) is defined as follows: QAdj = sign(∆Q)*sqrt(|∆Q|), where sign(∆Q) is the sign of difference between Tobin’s Q ratio of a certain company and industry’ median corresponding to Tobin’s Q ratio, whereas sqrt(|∆Q|) is the square root of absolute value of ∆Q. We decided to use median instead of mean because our data did not follow a normal distribution. The annual total assets (logarithmic values) will be employed as proxy for firm size. In fact, large companies require an additional managerial effort and a multifarious expertise, thus registering a higher size and independence related to corporate board of directors. Therewith, large companies benefit of more outside contractual relationships (Booth and Deli, 1996) than small companies, being emphasized a higher degree of transparency within these types of organizations (Lang and Lundholm, 1993; King et al., 1992). According to Coles et al. (2008a), Guest (2008), and Linck et al. (2008), board size and independence are positively related to several firm characteristics as indebtedness level, the age of the company, and industrial diversification. There is showed that highly leveraged companies with a long history and significant diversification are more complex and request superior experience and skills (Fama and Jensen, 1983; Guest, 2008). We will employ debt/book value of assets in order to control for the indebtedness level, considering the use of debt in order to restrict managerial discretion by reducing the size of cash-flow. The growth opportunities are proxied by the relative increase of sales from the previous year. The cost related to managers’ monitoring increases International Journal of Economics and Financial Issues, Vol. 3, No. 4, 2013, pp.885-900 890 with growth opportunities (Smith and Watts, 1992; Gaver and Gaver, 1993), being argued the fact that faster-growing companies will record a reduced size of boards and a higher percentage of insiders due to higher monitoring costs (Lehn et al., 2005). This circumstance is confirmed by Linck et al. (2008), who identifed that faster-growing companies registered a reduced size of board, as well a lower board independence. Moreover, we will use the logarithmic values as regards the number of years since listing on the BSE. Black et al. (2006) and Balasubramanian et al. (2010) mentioned that younger firms are likely to be faster-growing and perhaps more intangible asset intensive which can lead to higher Tobin’s Q ratio. 4.2 Empirical framework The hypotheses stated within Section three will be empirically tested by estimating multivariate regression models for panel data, unbalanced, both without cross-sectional effects and fixed-effects models. Besides, in order to detect potential nonlinear relationships we will estimate several polynomial regression models. Similar Baltagi (2005), we consider the following general form of panel data regression model without cross-sectional effects: y = α + X β+ u i = 1, ..., N, t =1, ..., T (1) where y is the dependent variable (industry-adjusted Tobin’s Q ratio), X is the vector of explanatory variables (variables regarding board of directors independence and CEO duality, as well control variables), the i subscript denotes the cross-section dimension, respectively the companies listed on the BSE, whereas t subscript denotes time, respectively the period 2007-2011. According to Baltagi (2005), most of the panel data applications employ a one-way error component model for the disturbances as following: u = μ + υ , where μ shows the unobservable individual-specific effect, whilst υ shows the remainder disturbance. As well, we will consider the following general form of the fixed-effects model, where μ are assumed to be fixed parameters to be estimated, whereas the remainder disturbances stochastic υ independent and identically distributed IID(0, σ ): y = (α + μ ) + X β + υ i = 1, ..., N, t =1, ..., T (2) The causal relationship between board independence and firm value, employing additionally CEO duality, will be investigated by the instrumentality of Granger (1969) approach. Thus, by considering two time series X and Y, it is said that X Granger-cause Y if a prediction of Y based on a set of information which comprises the history of X is better than a prediction which disregards the history of X. There will be estimated the following regression equations: y = α + α y + … + α y + β x + … + β x + ε (3) x = α + α x + … + α x + β y + … + β y + u (4) The null hypothesis stipulates that H0: β = β = ... = β = 0, respectively X does not Granger- cause Y in the first equation, whereas Y does not Grager-cause X in the second equation. In order to test for stationarity related to selected time series we will employ the test of Dickey-Fuller (1979) by estimating the following regression model: ∆y = α + γy + ∑ β ∆y + ε (5) As well, we will perform the test of Phillips-Perron (1988) by estimating the following regression model: ∆y = γy + u , because it corrects for serial correlations and heteroskedasticity in the errors. The null hypothesis considers that time series are non-stationary H0: γ = 0, while the alternative hypothesis states that time series are stationary H1: γ < 0. Subsequently, we will employ the vector autoregressive technique (VAR). Therefore, by considering y the vector of studied variables and ε the vector of innovations, we will estimate a VAR(p) model as following: y = A y + … + A y + Bx + ε . Moreover, the model VAR(p) can be represented as following: ∆y = ∏y + ∑ Γ∆y + Bx + ε , by considering that ∏ = ∑ A − I, whereas Γ = - ∑ A . The potential long-term connections will be empirically tested by employing Johansen (1991) cointegration test. This procedure consists in estimating the matrix Π out of an unrestricted VAR model, respectively testing if the restrictions required by the reduction of the matrix Π rank can be rejected. Board of Directors Independence and Firm Value: Empirical Evidence Based on the Bucharest Stock Exchange Listed Companies 891 5. Empirical Results 5.1 Descriptive statistics Table 2 shows descriptive statistics regarding all the variables employed within empirical research. The mean percentage of independent directors is only 13.77 percent. In fact, we notice that the recommendation out of the Guide for implementing Corporate Governance Code (2010) which states that at least a quarter of the total number of directors shall be independent is not followed. Moreover, the mean percentage of non-executive directors (54.43 percent) highlights that the balance between executive and non-executive members recommended by the Bucharest Stock Exchange Corporate Governance Code (2008) is accomplished. Table 2. Summary statistics Variable N Mean Median Min Max Std. Dev. QAdj 334 0.08928 0.00000 -0.81178 1.87060 0.570688 IND 334 0.137700 0.000000 0.000000 0.800000 0.174208 NED 334 0.544390 0.600000 0.200000 0.888890 0.176366 CEODual 334 0.389221 0.000000 0.000000 1.000000 0.488305 FS 334 8.241298 8.193217 6.977173 10.52934 0.610849 Lev 334 0.387540 0.353737 0.006916 1.940834 0.285651 SGrowth 334 0.070588 0.045353 -0.913607 2.503076 0.356558 Listing 334 0.968339 1.041393 0.000000 1.204120 0.253036 Source: Author’s calculations. Description of the variables is provided in Table 1 Table 3 exhibits the frequency table of the percentage of independent directors and non- executive directors, while Table 4 shows the frequency table of CEO duality for the companies listed on the BSE.For independent directors we remark that the threshold of 50 percent is almost not exceeded, whereas the percentage of non-executive directors registers the highest frequency between 30 percent and 40 percent. Likewise, in average, within 60.95 percent out of the selected companies, the roles of CEO and Chairman of the board are performed by different persons. By comparison, De Andres et al. (2005) reported the following average values as regards the percentage of outside directors on corporate boards of directors: UK (48 percent), USA (79 percent), Canada (74 percent), Belgium (76 percent), Spain (75 percent), France (81 percent), Italy (74 percent), Switzerland (90 percent). Table 3. Frequency table as regards the percentage of independent directors and non-executive directors for the companies listed on the Bucharest Stock Exchange Variable 2007 2008 2009 2010 2011 N % N % N % N % N % 0%<=IND<=10% 33 52.3809 34 50.7462 36 52.94118 36 52.9411 36 52.9411 10%