1 Introduction

Corporate governance (CG) has received growing attention in recent years, and literature and international organizational pronouncements have recognized that “good” CG is an instrument with which to achieve business aims and, particularly, with which to oversee and monitor their compliance (Cañibano 2006). However, the current debate about the role of CG revolves around the protection of stakeholders’ interests, leading to the emergence of the concept of stakeholder corporate governance (Donaldson and Preston 1995; Freeman 1984; Kaufman and Englander 2011; Sternberg 1997). Accordingly, good governance practices should focus not only on the interests of shareholders but also on the interests of a wide group of stakeholders (those agents whose utility is affected by the decisions of the firms, e.g., employees, managers, shareholders, customers, suppliers, creditors and the local community).

In line with the concept of stakeholder corporate governance, Daily et al. (2003) define CG as “the determination of the broad uses to which organizational resources will be deployed and the resolution of conflicts among the myriad participants in organizations” (p. 371). Specifically, Donaldson and Preston (1995) argue that the firm should pay attention to stakeholders for two main reasons: first, as stakeholders contribute to the firm’s wealth creation, they have legitimate claims that need to be met (normative approach to stakeholder theory); and, second, due to the influence some stakeholders have on the firm, this group should be taken into consideration because of their possible impact on firm profitability (instrumental approach of the stakeholder theory). Based on the above arguments, researchers have linked two literature streams, CG and stakeholder theory, showing that the orientation of CG systems is important for corporate social responsibility (Ayuso et al. 2014; Ricart et al. 2005) and for enabling firms to face up to the conflict of interests between stakeholders, particularly those conflicts resulting from the distribution among different stakeholders of quasi-rents generated by the firm (Ayuso and Argandoña 2009). Additionally, the design of the CG structure and its composition “induce or force management to internalize the welfare of stakeholders” (Tirole 2001, p. 4). In this sense, because managers and directors of the firm are in charge of the decisions on maximization and distribution of quasi rents, the board of directors has to play an important role in reconciling the interests of the different stakeholders with regard to value-added distribution.

Although there is a broad theoretical debate regarding the stakeholders’ perspective as a way of focussing the role of boards of directors (see, among others, Van Buren 2001; Ayuso and Argandoña 2009), we have very limited insight into whether board structures and composition can truly have an impact upon the stakeholders’ interests. Several papers analyse the influence of an effective board on firm performance from the traditional shareholders’ perspective (Agrawal and Knoeber 1996; Nicholson and Kiel 2007; Azim 2012; Cabrera-Suárez and Martín-Santana 2015; Lefort and Urzúa 2008, among others) without taking into account the effect of these boards on the protection of the stakeholders’ interests. Furthermore, a dichotomous point of view regarding the complementary or substitutional role of the board in attending to the interests of shareholders and other stakeholders has arisen among corporate governance researchers. First, some authors have made the point that an effective board of directors from a shareholder’s perspective should also address the different stakeholders’ interests (Donaldson and Preston 1995), which would result in a complementary relationship between the effectiveness of the board and the interests of shareholders and other stakeholders. However, recently, García-Torea et al. (2016) indicated that the change to a CG system based on the stakeholders’ view “requires revisiting the conceptualization of board effectiveness under the shareholder perspective to assess its validity for the stakeholder perspective” (p. 249), which would lead to the existence of a substitute relationship between board effectiveness and the interests of both (shareholders and stakeholders). In other words, efficient corporate governance mechanisms for the protection of shareholders’ interests could damage the other stakeholders’ interests. Consequently, greater progress regarding the consequences of different CG structures on other stakeholders’ interests is necessary to gain a better understanding of the concept of stakeholder CG.

Considering the importance that stakeholders have acquired in recent times both from the points of view of institutions (national and international initiatives highlight the role of stakeholders in today’s society) and of business (companies want to differentiate themselves by satisfying as many stakeholders as possible), this paper contributes to bridging this gap by analysing whether the effective structure of the board, from the traditional point of view of the shareholders (including shareholder/owners, non-independent directors, CEO and Chairman that holding shares), is also valid to defend the interests of other primary stakeholders (workers, creditors and government) or if, on the contrary, it is necessary to review the traditional effectiveness of the board, as is proposed by some authors. Therefore, an investigation of the boards’ characteristics as an important driver in the stakeholders’ corporate governance may provide insights to improvements in corporate governance practices. This paper specifically investigates aspects of corporate governance linked to the distribution of value added among stakeholders. The research is motivated by the OCDE (2015, p. 46) recommendations addressing corporate governance challenges that advise that the board “take into account the interests of stakeholders”. This study includes an examination of various aspects of the effectiveness of boards of directors, including board size, separation of chairman and CEO figures, board independence and board ownership, in relation to the value-added distribution to stakeholders.

We have considered only primary stakeholders because they are the ones that maintain an economic stake in the company and are therefore those that could be directly affected in the distribution of wealth (measured by value-added distribution). Considering that the main problem of the primary stakeholder approach is in the distribution of wealth, we focus our approaches on value-added distribution as a proxy for the primary stakeholders’ interests. In doing so, this paper contributes to the understanding of how the CG models and stakeholders’ theory interact in the firm with regard to the issue of distribution of wealth.

Thus, the specific research questions addressed in this study are the following: Do board size, separation of the roles of CEO and Chairman, independent directors and board ownership have an impact on the value-added distributed to stakeholders who are differentiated as shareholders and other primary stakeholders? Accordingly, we ask the following question: Do the characteristics of the board contribute to a more equitable distribution of value added among stakeholders?

To test our hypotheses, we have chosen the Spanish context, as the characteristics that it presents (high concentration of ownership and a unitary board system) (see for more details, Acero and Alcalde 2013; De Miguel et al. 2004; Manzaneque et al. 2016) make it an ideal scenario to study the role played by the board of directors in satisfying the interests of the different stakeholders. From the sample obtained, we conduct a panel data analysis of 438 observations of Spanish listed firms during the period 2007–2012. The contribution of this research is both theoretical and empirical. From an empirical point of view, the results show that the characteristics of the board of directors have an impact on the primary stakeholders’ interests in terms of value-added distribution. In particular, although the separation of the roles of chairman and CEO and board independence are effective from a stakeholder perspective, other characteristics of the board of directors, such as board ownership and board size, might not serve in the interests of primary stakeholders. From a theoretical point of view, our study offers a greater insight into the impact of corporate governance practices on the primary stakeholders’ interests and to the reconciliation of the dichotomized approach of CG (shareholders/stakeholders CG approach). In addition, our study offers arguments in favour of companies moderating the size and shareholding of the board in order to reduce the difference in value-added distribution between stakeholders, especially in the context of ownership concentration.

The article is structured as follows. In the next section, the literature about the shareholders’ and stakeholders’ approach to CG is summarized, and the development of the hypotheses is presented. After this review, the following section describes the data, variables and methodology. The third section discusses the results of the statistical models. Finally, the last section presents the discussions and conclusions.

2 Theoretical background

2.1 Literature review

Currently, the debate on CG models contrasts two approachesFootnote 1: the approach from the perspective of the shareholder and the approach from that of the stakeholder. The traditional shareholder perspective is based on the idea that “corporate governance comprises the set of mechanisms that induce the managers who control corporations to make decisions that maximize the value of the shareholders who own those corporations” (Denis 2016, p. 468). Following this approach, the created value is measured by the wealth that shareholders receive, that is, the investment returns (Shleifer and Vishny 1997), and CG is the main mechanism to maximize that value, as it can prevent the expropriation of the shareholders’ wealth due to the managers’ self-interest (Jensen and Meckling 1976). In this regard, and from the empirical perspective, previous studies concerning CG systems have focused on the relationship between board characteristics (mainly, CEO/Chairman duality, board independence, board size, and board ownership) and corporate value creationFootnote 2 for shareholders (shareholder perspective). However, the obtained results are mixed. For example, Jackling and Johl (2009) and Nicholson and Kiel (2007) show that greater board size has a positive impact on value creation by the firm, while O’Connell and Cramer (2010) and Arosa et al. (2013) report a negative relationship. In relation to CEO duality, Arosa et al. (2013) and Cabrera-Suárez and Martín-Santana (2015) report a positive relationship with value creation, while other studies find no relationship (Jackling and Johl 2009; Villanueva-Villar et al. 2016). In addition, the percentage of independent directors on the board has been found to be positively related to value creation (Azim 2012; Villanueva-Villar et al. 2016). The evidence on the impact of board ownership is also mixed. Fahlenbrach and Stulz (2009) report a positive relationship between board ownership and firm performance, but Sheikh et al. (2013) show a negative relationship.

One possible explanation for these mixed findings could be that the approaches do not consider all the stakeholders’ interests. In fact, although the shareholders’ model has dominated CG literature, a new stakeholderFootnote 3 oriented view has been attracting attention since the late 20th century, considering that stakeholders represent all the agents whose utility is affected by a firm’s decisions (Charreaux and Desbrieres 2001), for example, employees, customers, suppliers, communities, government, and society in general. According to the stakeholder theory, the aforementioned traditional representation of CG based on shareholders’ interests is incomplete since the firm’s decisions involve consequences for all stakeholders (Charreaux and Desbrieres 2001; Tirole 2001) and, consequently, all stakeholder voices should be included in the CG process (Van Buren 2001), and their interests should be taken into account (see seminal words of Freeman (1984) or Blair (1995), among others). In short, stakeholder theory proposes that management decisions should address the needs, expectations and values of all stakeholders.

Taking into account stakeholders interests and needs, the economic perspective of the stakeholder theory has been linked with the social responsibility perspective of CG, given that CG fulfils a social function in attending to all stakeholders’ claims. First, according to the legal and economic perspective of stakeholder theory, the firm has contractual responsibilities with the stakeholders, and these stakeholders have legitimate interests in the corporate activity of the firm. Specifically, Clarkson (1995, p. 110) posits that the purpose of firms should be to “create and distribute wealth or value sufficient to ensure that each primary stakeholder group continues as part of the corporation’s stakeholder system”. According to that perspective, each individual stakeholder is important to the survival of the corporation because they bring resources into a company that enable greater profits to be achieved, which is a critical objective for the future of the firm (Freeman and Reed 1983; Freeman 1984). As firms gain a higher level of interdependency with primary stakeholders, their survival and profitability depend on their ability to create wealth and value to satisfy those primary stakeholders’ interests.

In line with the above arguments and as a result of the union of stakeholder approach and CG theory, the concept of stakeholder corporate governance has emerged. According to this relatively new CG perspective, CG should act in the interest of all stakeholders rather than in the interests of only its shareholders (Denis 2016). In this sense, effective stakeholder CG should provide the following: protection of the legitimate interests of all stakeholders, disclosure of information in the most transparent way, mediation between the interests and demands of all types of stakeholders, and strategic and ethical guidance for firms (Aguilera et al. 2015). Under this perspective, since stakeholder CG is expected to be exercised through internal control mechanisms rather than through external control markets (Letza et al. 2004), the board of directors should be the main mechanism that defends the interests of shareholders and other stakeholders. Hence, it is necessary to take into account the board characteristics (size and composition, among others) (Ayuso and Argandoña 2009) since these will determine the board’s effectiveness in performing its monitoring and strategic advisory roles over management (De Andrés et al. 2005) and will affect the satisfaction of all stakeholders’ interests and demands.

Taking this perspective, the existence of dominant interests could influence the impact of corporate governance on the firm’s value-added distribution and, consequently, on the satisfaction of stakeholder interests. In other words, while it is likely that shareholders control corporate governance practices, it is also probable that decisions about value creation and distribution are oriented to their interests. Accordingly, some corporate governance mechanisms may contribute to reducing the differences among stakeholder interests and, consequently, contribute to a more equitable value-added distribution between shareholders and other stakeholders.

2.2 Hypothesis development

2.2.1 Board size

According to the shareholders’ paradigm, an appropriate board size contributes to greater board effectiveness and quality with regard to decision-making, although no consensus exists about the ideal size of the board. It is claimed that having more members on the board contributes to meeting the diversity criteria and improves efficiency (Dalton et al. 1999) and independence (Pearce and Zahra 1992). Adams et al. (2005) argue that having a greater number of members on board contributes to the diversity of the board membership and of opinions and experiences, which increases the monitoring capacities of the group. In other words, the board has more information, the advantage of more expertise (Dalton et al. 1999), a greater knowledge base (Coles et al. 2008) and a broader range of experience (Xie et al. 2003). In addition, more members on the board can contribute to directors having greater opportunities to exercise their independence from the CEO and to exercising their power in governing the firm (Pearce and Zahra 1992). Additionally, larger boards contribute to the mitigation of distributional conflicts between insider and minority outsider owners’ wealth (Allegrini and Greco 2013). Nevertheless, it is also accepted that small boards have significantly less coordination and information problems because of the small board’s greater speed and efficiency in the decision-making process (Lipton and Lorsch 1992; Jensen 1993). Taking all these views into account, it would appear that boards (large or small) have both advantages and disadvantages. Consequently, an adequate board size could contribute to protect the shareholders’ interests and increase the shareholders’ wealth.

From the point of view of stakeholders, there is some evidence that more members on boards can contribute to a diversity of criteria, opinions and experiences, thereby increasing the monitoring capacities of the group (Adams et al. 2005), as the board members enjoy greater diversity among their members (Klein 2002). As the literature about board diversity maintains, boards comprising different types of members should provide more effective problem solving between stakeholders (Carter et al. 2003) and greater social capital—relationships with business networks and stakeholders (Dang et al. 2014). Additionally, a greater number of members on a board allows the incorporation of stakeholders’ representatives. According to Huse and Rindova (2001), in stakeholder-oriented CG systems, since the interest of the board is to protect all stakeholders’ interests, representatives of their interests should be present on the board. Consequently, it is argued that more members on board can contribute to responding to the challenge of the diversity of stakeholders’ demands.

These arguments suggest that a greater number of members on the board may contribute to defending the interests of both shareholders and other primary stakeholders, which will result in a smaller difference of the added-value granted to each one of them; therefore, we propose the following hypothesis:

Hypothesis 1

Board size impacts the stakeholders’ value distribution (differentiating between shareholders and other primary stakeholders), promoting smaller differences between the shareholders and other primary stakeholders in the value distributed.

2.2.2 Separation of Chairman and CEO

Previous empirical studies suggest that the presence of CEO duality (the same person is the Chairman and the CEO) reduces the boards’ allocation of attention to monitoring (Tuggle et al. 2010) and negatively affects the effective control of the board (Morck et al. 1988; Gul and Leung 2004). CEO duality implies more power on the board for the Chairman (Nahandi et al. 2011) and, consequently, less board effectiveness in pursuit of achieving the shareholders’ interests (Hermalin and Weisbach 2003), negatively influencing the value creation of the company (Pucheta-Martínez 2015). In other words, CEO duality contributes to managerial opportunism (Blackburn 1994) and thus leads to decision-making for the benefit of management (Jensen and Meckling 1976) and to the detriment of the company’s shareholders (Coles et al. 2001; Jensen 1993). In line with this argument, in the interest of shareholders, most research on CG systems emphasizes the desirability of separating these positions (CEO and Chairman).

In the same way, CEO duality, which implies that all decision making power falls to the same person, may result in CEOs making decisions that only benefit themselves (Jensen and Meckling 1976) and, therefore, that may also be detrimental to the interests of other stakeholders. In fact, Webb’s (2004) study found that socially responsible firms, which consequently have more possibilities of responding to stakeholder interests (García-Torea et al. 2016), have less CEO duality than do non-socially responsible firms. Therefore, we propose that the separation of the roles of the CEO and the Chairman has a positive effect on the distribution of value added to shareholders and other primary stakeholders, which will generate a smaller difference of the added value granted to both.

Hypothesis 2

Separation of the roles of Chairman and CEO impacts the stakeholders´ value distribution (differentiating between shareholders and other primary stakeholders), promoting smaller differences between shareholders and other primary stakeholders in the value distributed.

2.2.3 Independent directors

Board independence (the number or proportion of outside directors on the board) is one of the most analysed issues in relation to examining a board’s effectiveness for protecting shareholders’ interests (e.g., Fama and Jensen 1983; Jensen 1993; Dahya et al. 2002). According to Mangena and Tauringana (2007), since directors should be independent, they tend to side with the investors’ information needs and encourage managers to satisfy them. Additionally, in the context of concentrated ownership, as is the case in most continental European countries, including Spain, and Asian countries such as Japan, the problem known as principal–principal (large against minority shareholders) is more frequent. Therefore, the role of the independent directors in controlling the large shareholders’ actions may be essential for defending the interests of minority shareholders (Pucheta-Martínez 2015). In fact, board independence is often linked to better CG (Luan and Tang 2007). In particular, a higher presence of independent directors on the board is linked to greater board independence and effectiveness, taking into account that the independent directors have experience in decision-making processes, are less exposed to external influences and have a special interest in defending or building their reputation in the job market (Conyon and He 2008). Independent directors are also expected to be more sensitive to societal needs (Ibrahim et al. 2003), more knowledgeable about the changing demands of stakeholders (Ayuso and Argandoña 2009) and consequently more sensitive to a wide range of stakeholders’ interests. Furthermore, some empirical studies have found that having a higher proportion of independent directors is linked to better decisions on issues such as acquisitions or executive compensation (Cotter et al. 1997; Mayers and Smith 2010).

According to the aforementioned literature, we would expect a positive relationship between board independence and the distribution of value added to shareholders and other primary stakeholders, promoting smaller differences between them.

Hypothesis 3

Independent directors impact the stakeholders’ value distribution (differentiating between shareholders and other primary stakeholders), promoting smaller differences between the shareholders and other primary stakeholders in the value distributed.

2.2.4 Board ownership

The participation of the board of directors in shareholding is also an important mechanism of board effectiveness to protect shareholders interests, mainly, those of minority shareholders. Most studies support the notion that board of directors’ shareholding creates independence and reduces the possibility of opportunistic actions of the directors due to their interests being aligned with those of other shareholders (Shleifer and Vishny 1997). In other words, most studies support a convergence of interest hypothesis between managers and shareholders. However, according to the so called entrenchment hypothesis, this positive effect may become negative if the board’s shareholding is excessive, resulting in the entrenchment of directors who would use their power to carry out decisions aimed at avoiding the cessation of managers in the case of inefficient behaviour. Previous empirical research shows the predominance of one or another hypothesis for different levels of shareholder participation on the part of directors (Morck et al. 1988; Hermalin and Weisbach 1991). In line with the entrenchment hypothesis, higher levels of ownership on the part of directors confers them with enough power to make decisions in their own interest at the expenses of other stakeholders. For example, they could increase their salaries as board members or improve their employment conditions (Rose 2005), direct investment to projects with a personal interest (Lemmon and Lins 2003) or influence funding decisions when directors are representatives of financial investors (Delgado-García et al. 2010).

Taking these arguments into consideration, we expect greater stock ownership by directors to have a negative effect on the interests of the shareholders and other primary stakeholders, generating smaller differences between both parties.

Hypothesis 4

Board ownership impacts the stakeholders’ value distribution (differentiating between shareholders and other primary stakeholders), promoting smaller differences between the shareholders and other primary stakeholders in the value distributed.

3 Research method

3.1 Sample selection and data

The sample in this study comprises all the companies listed in the Spanish Stock Exchange Interconnection System (Sistema de Interconexión Bursátil Español, SIBE) or in the continuous market for the period 2007–2012. Starting from the population of listed companies, whose number varies according to the year of study considered, and after eliminating those for which there were no data for the whole time period, the final study sample is made up of 73 companies (see Table 1, Panel B). This sample provides a broad representation of the various business sectors in Spain (see Table 1, Panel A). Moreover, we have tested the maximum allowable errorFootnote 4 for a finite population. The maximum error is 7.51% with a level of confidence of 95%, corroborating that our sample is representative of the population.

Table 1 Population and sample analysis

We have chosen this period because it is a homogeneous period regarding practice and transparency in terms of financial and corporate governance information. In fact, different reforms in corporate governance transparency and policies were incorporated into the Spanish context after 2012 (Merino and Manzeneque 2017).

The information about financial data (value-added distribution) has been taken from the Annual Accounts, and the corporate governance information (board size, separation of Chairman and CEO, board independence and board ownership) has been taken from the Corporate Governance Annual Report. This information is available on the National Stock Exchange Commission (CNMV, Spain) web page.

We study the Spanish context because it provides an interesting setting in which to exam the role of the board of directors in controlling the conflict between controlling and minority shareholders—principal–principal conflict of interest—and how this can influence the distribution of value added between these and the other stakeholders. The characteristics of Spanish listed companies, among others, are as follows: high concentrated ownership, with almost 46% of the total shares held by controlling shareholders (Manzaneque et al. 2016), and a corporate governance system based on a unitary board structure strongly dominated by the controlling shareholders. For example, in a similar sample of Spanish companies, proprietary directors (who represent the owners of blocks of shares of the company) represent 44% of the members of the board (Pucheta-Martínez 2015). These two characteristics mean that the traditional problem of separation of ownership and control is not as important as it is in other markets with dispersed ownership (i.e., the US) or with a two-tier board system (i.e., Germany, Holland or Japan). Thus, in the context of high concentrated ownership (Spain, Japan or Korea, among others), the role of the board of directors should be crucial in protecting the interests of the different stakeholders.

3.2 Measures

Dependent variable. Value-added distribution The value added was chosen as a proxy of the firms’ value creation. The consideration of the value added as an indicator of a firm’s activity allows a more rational evaluation of efficiency and productivity and provides an overview of the production process as a collaborative process between the actors involved in it. Our objective is to analyse whether an effective CG system under the shareholder CG model is also effective in creating an equitable distribution of wealth creation among the stakeholders. To attain this goal, we measure the net value-added distribution to shareholders (dividends and retained earnings) and other primary stakeholders, including workers (wages and social charges), creditors (interests) and government (taxes). Specifically, the net value added is a measure of wealth that can be extracted from published accounting information, and, particularly, value added creation and distribution is disclosed as follows (see, among others, Askren et al. 1994; Bannister and Riahi-Belkaoui 1991):

$$Net\; value\;added = S - B - DP = Distribution\;of\;net\;value\;added = W + I + DD + T + R$$
(1)

where S = Sales, B = Bought-in materials, DP = Depreciation, W = Wages and social charges, I = Interests, DD = Dividends, T = Taxes, and R = Retained earnings.

In other words, the net value added is the economic surplus generated by the activity of the company during a period of time. Thus, it is a measure of management efficiency since it links the production obtained with the factors that contribute to its generation. Net value added is used in this study in order to avoid the arbitrary and incorrigible effects of depreciation allocation (Riahi-Belkaoui 1999). We construct two variables for the stakeholders’ value-added distribution. Each variable is calculated as a ratio. Specifically, the shareholders’ value-added distribution is measured as the ratio of dividends and retained earnings to total value added distributed by the firm. The other primary stakeholders’ value-added distribution is the ratio of workers costs (wages and social charges), financial expenses (interests) and taxes to total value added distributed by the firm. Additionally, the difference between the added value distributed to shareholders and other primary stakeholders has been calculated in order to measure the difference of interests between them.

Independent variable. Corporate governance To test the effect of board characteristics on value-added distribution, we included in the model the following independent variables: Board size, Separation of Chairman/CEO, Board independence and Board ownership. The first variable, Board size, is measured as the total number of members on the Board of Directors. Separation of Chairman/CEO is a dichotomous variable that takes a value of 1 when different people hold the two positions; otherwise, it takes a value of 0. Board independence is measured as the proportion of independent directors to the total number of members of the board. We consider as independent directors those directors who are in a position to perform their duties without being influenced by any connection with the company, by the shareholders or by the management. Finally, Board ownership measures the level of participation of the board of directors in shareholding as the percentage of shares held by shareholders who are at the same time members of the board.

Control variables We also included seven control variables: Industry, Firm Size, ROA (Return on Assets), Leverage, Firm Age, Ownership concentration and Family management. Industry is a categorical variable with six groups, following the approach established for the companies listed on any Spanish Stock Exchange: (1) oil and energy; (2) basic materials, manufacturing and construction; (3) consumer goods; (4) consumer services; (5) financial services and real estate; and (6) technology and telecommunications. As different industries obtain different levels of performance (Carter et al. 2010), we expect that the industry influences the value-added distribution. Consequently, we create dummy variables for each industry sector and take as reference the first category in the models. Size is measured as the logarithm of total assets. We expect that firms with a greater ROA (measured as ratio of operating income to net assets) also have more added value to distribute to the different stakeholders. Leverage is measured as the book value of the debt to total assets ratio, which is expected to be negatively associated with performance and value creation (Jackling and Johl 2009). We also include Firm Age, measured as the years of the company’s life. Finally, two control variables have been considered in order to capture the effect of ownership structures on the value-added distribution. Because large shareholders may pursue their own interests, causing a negative impact on other shareholders’ wealth (expropriation of minority shareholders’ problem) (Shleifer and Vishny 1997), we control for ownership concentration, which is measured as the percentage of total shares in the hand of block holders (those with 3% or more of the shares). In addition, the previous literature highlights the role of family controllers on the decision making. As family managers’ decisions are often a guide for understanding non-economic goals (see among others, Chrisman et al. 2004) and as family managers are aware of the influence of other non-family members or stakeholders on the success of the firm (Farrington 2009), family managers usually try to attend to their interests to guarantee the future survival of the firm. We measure the family managementFootnote 5 variable as a dummy variable, which takes a value of 1 if one or more members of the family are present on the board of directors and hold shares on the firm (at least 20%); otherwise, it takes a value of 0.

3.3 Model

To test the hypotheses, several different variants of a linear panel data regression model were estimated. Following this methodology, a data sample of 438 observations (73 companies × 6 years) was developed composing a short, linear and strongly balanced panel. Different tests were performed in order to define the econometric specifications of our model. To control possible heterogeneity and endogeneity problems, we use the corrections over panel data proposed by Arellano and Bond (1991) and Blundell and Bond (1998). Hence, a 2 step GMM estimator is computed, using two lagged values. The results of the first and second serial correlation test and the Hansen over-identifying restrictions are reported in Table 4. The results of these tests support the validity of the GMM specifications and the instruments used. In essence, the model estimated is the following:

$$\begin{aligned} Value\;added\;distribution\;to\;stakeholders_{it} = & \alpha + \beta_{1} *Board\;size_{it} + \beta_{2} *Separation\;Chairman/CEO_{it} \\ & + \,\beta_{3} *Board\;independence + \beta_{4} *Board\;ownership_{it} + \beta_{5} *Firm\;size \\ & + \,\beta_{6} *Leverage + \beta_{7} *Firm\;age + \beta_{8} *ROA + \beta_{9} *Ownership\;concentration \\ & + \,\beta_{10} *Family\;management + \sum\limits_{K = 2}^{6} {INDUSTRY\;dummies_{it} + \varepsilon_{it} } \\ \end{aligned}$$

where value-added distribution to stakeholderit is the endogenous variable. According to the hypotheses, we develop different variants of the above model (see Table 4). Those models are described in the next section.

4 Empirical findings

4.1 Descriptive analysis

Table 2 provides descriptive statistics. In panel A, the means of value added for period 2007–2012 are presented. In 2007, the shareholders received 44.6% of value added, and the other primary stakeholders received 55.4%. The distribution of value added to other primary stakeholders has increased in the study period (2007–2012) from a mean of 55.4% in 2007 to a mean of 71.6% in 2012, while for shareholders, the distribution has been reduced (from 44.6% in 2007 to 28.4% in 2012). Undoubtedly, the crisis has mainly affected the distribution of value added to shareholders (through dividends and retained earnings), whose percentage has been reduced since 2007–2012.

Table 2 Descriptive statistics

Table 2 (Panel B) provides descriptive statistics on the dependent, independent and control variables used in the study. As far as continuous variables are concerned, the mean, medium, a 25%, a 75% and the standard deviation are presented. On average, 35.6% of the directors on the board were independent directors. The proportion of stock ownership by directors presents a mean of 22.8%. The mean board size is approximately 12 members. Regarding control variables, the average level of leverage reaches 56.9%; the mean of the ROA is 4.3%; and the mean of ownership concentration is 40.4%. The mean of the firm age is approximately 48 years from firm creation, and 42.24% of the firms in our sample are family managed firms.

With respect to categorical variables, frequency and percentage are presented. Approximately 37% of observations present a separation of CEO and Chairman positions over the sample period.

Finally, the sample presents the following sector classification from highest to lowest percentage: basic materials, manufacturing and construction (24.66%), consumer goods (23.29%), financial services and real estate (20.55%), consumer services (13.70%), oil and energy (10.96%), and technology and telecommunications (6.85%) (see Table 1, Panel B).

Additionally, the Pearson correlations (see Table 3) generally suggest that some CG characteristics are correlated to the value-added distribution to shareholders and other primary stakeholders. In particular, board size and board independence are positively correlated with the value added distributed to shareholders, although negatively correlated to the value added distributed to other primary stakeholders. Contrarily, board ownership and separation of Chairman/CEO appear to be negatively correlated to the shareholders’ value added and positively to other primary stakeholders’ value-added distribution. Regarding the distribution of value added among shareholders and other primary stakeholders, the results show that their interests are in conflict in the sense that value added is restricted, and the increase in the value-added distribution to one reduces the value-added distribution to the other. Despite the obtained results, all binary correlations are lower than 0.7, evincing that this is not a problem in our study.

Table 3 Cross-correlations matrix

To test for multicollinearity, the Variance Inflation Factors (VIF) were calculated for each independent variable (see Table 3). Multicollinearity seems not to be a problem in our study, as all VIF values are low enough to dismiss multicollinearity problems (Belsley et al. 1980; Hair et al. 1998).

4.2 Regression analysis

Table 4 presents the results of the System-GMM in two-step estimations. To test if corporate governance characteristics have an impact on the value-added distribution, in Models 1 and 2, we use as dependent variables the value added distributed to shareholders and other primary stakeholders. In addition, it is possible that corporate governance characteristics could also be shown to cause greater differences in the value-added distribution. According to this reasoning, Model 3 is included, using as a dependent variable the greater value received by shareholders at the expense of other primary stakeholders. Therefore, we calculate the variable “Shareholders greater value-added distribution” as the difference between the value added distributed to shareholders and the value added distributed to other primary stakeholders. We take the value of the previous difference if the result is positive; otherwise, the negative values are replaced by 0.

Table 4 Estimation. System-GMM in Two Steps. Corporate governance and value-added distribution to stakeholders

The results of Models 1 and 2 show that board size has a positive impact on the value-added distribution to shareholders (β = 0.047, ρ < 0.05, Model 1) and a non-significant impact on the value-added distribution to other primary stakeholders (β = −0.026, ρ > 0.10, Model 2). Additionally, Model 3 shows that board size contributes to a greater difference among the value-added distributed to shareholders and other primary stakeholders (β = 0.052, ρ < 0.01, Model 3), which corroborates the previous results. In terms of our hypothesis, as the board size exerts an impact on shareholders but not on other primary stakeholders’ value, this board characteristic is not an effective instrument with regard to attending to the interests of other primary stakeholders. Therefore, H1 is not proved given our results.

Regarding the Separation of Chairman/CEO variable, the results show that this variable has a significant and negative effect on the return to shareholders for the distribution of added value (β = −0.756, ρ < 0.01, Model 1) and a significant and positive effect on the value added achieved by other primary stakeholders (β = 0.557, ρ < 0.05, Model 2). These findings suggest that the Separation of the Chairman/CEO roles is not advantageous for shareholders in terms of return; however, this is supported for an effective impact on other primary stakeholder value added. Further, the results of Model 3 indicate that separation of the Chairman/CEO roles is advantageous to other stakeholders in terms of the existence of a lower difference in the value-added distribution to shareholders and other primary stakeholders (β = −0.694, ρ < 0.01, Model 3). Given the impact of the separation of Chairman/CEO roles on the distribution of value added, both results show that it may contribute to reduce the conflict of interests among shareholders and other groups of primary stakeholders. Therefore, H2 is accepted. Thus, these findings agree with the traditional trend that supports the separation of positions in order to protect the stakeholders’ interests, therefore, imply a true state of effectiveness for the separation of chairman/CEO under the stakeholder perspective.

With respect to the presence of independent directors on the board, the results show that this variable impacts negatively on shareholders’ value added (β = −1.165, ρ < 0.01, Model 1) and positively on the value-added distribution to other primary stakeholders (β = 1.042, ρ < 0.10, Model 2). Although the results are marginal, the signs obtained suggest that having a greater number of independent directors on the board is related to a lower value-added distribution to shareholders and a greater value-added distribution to other primary stakeholders. Moreover, from a stakeholder perspective, Wang and Dewhirst (1992) assert, “outside directors have a very strong stakeholder orientation and recognize that their responsibility encompasses more than shareholders and are very conscious about the needs and expectations of the various constituencies of their firms” (p. 120). Our results are consistent with that point of view, indicating that firms with more independent directors have a greater distribution of value added to primary stakeholders other than the shareholders. In fact, the results of Model 3 confirm that shareholders receive lower value added in relation to other primary stakeholders when the proportion of independent directors is greater (β = −1.011, ρ < 0.05, Model 3). Therefore, H3 is accepted, as board independence contributes to reducing differences in value-added distribution and is in line with the stakeholder corporate governance perspective.

Our final hypothesis suggests that the level of board ownership impacts the value-added distribution to shareholders (H4). As indicated in Table 4, this variable is not significantly associated with shareholders value added (β = 0.493, ρ > 0.10, Model 1). There is, however, a negative relationship between board ownership and the value added received by other primary stakeholders (β = −0.321, ρ < 0.10, Model 2). These results imply that directors’ stock ownership is also a factor that contributes to lower stakeholders’ value and, consequently, it is not an effective factor in responding to the interests of the firm’s primary stakeholders other than the shareholders. Further, the results of Model 3 (β = 0.497, ρ < 0.10, Model 3) reinforce the idea that the directors’ ownership aligns their interests with those of shareholders and has a negative effect with regard to promoting all stakeholders interests in that there is an increase in the difference between the value added distributed to shareholders compared to the value added distributed to other primary stakeholders. In fact, a greater difference in the value-added distribution is shown. Therefore, Hypothesis 4 is not supported.

To summarize, our results reinforce the idea that some board characteristics are not significant to the stakeholders CG system, and new challenges and advances should be made to incorporate this perspective to the board as the main mechanism of internal control of the firm.

4.3 Robustness check

We perform further analyses to check the robustness of our results (see Model 4 and 5 in Table 4). On the one hand, some studies suggest that it may be desirable to set limits regarding board size in order to improve effectiveness (Yermack 1996). In fact, the Spanish Good Governance Code of Listed Companies (CNMV 2015) included a limit recommendation for board size of between 5 and 15 members. Therefore, we add a dummy variable (Board size dummy) to Model 3 to determine if board size is within that limit (Model 4). The results show that if the number of directors is between 5 and 15 members, the board size dummy variable has a significant impact on the difference in value-added distribution among shareholders and other primary stakeholders. However, the board size maintains the positive impact on value-added distribution in this new Model 4, thus rejecting our hypothesis.

On the other hand, in relation to the presence of independent directors, the most common practice in the international context is to recommend that at least one-third of the board members be independent (Spain and Singapore, among others). To test if the limit could be important to the relationship between board independence and value-added distribution, we add a dummy variable to Model 3 that represents those cases in which the number of independent directors are greater than one-third of the total number of directors (Model 5). The results remain constant, so H3 is definitely accepted. All these results taken together indicate that the presence of independent directors affords a board the advantage of incorporating the wider interests of other stakeholders and of contributing to the promotion of the stakeholder theory in future challenges of CG.

5 Discussions and conclusions

This work examines if the structure and composition of the Board of Directors (CEO/Chairman duality, Board independence, Board size and Board ownership) impact the value-added distribution to shareholders and other primary stakeholders and if these factors could contribute to a more equitable distribution of value added between them. In general, we find that in the context of ownership concentration and with a unitary board system, CG is important to balance the interests of different stakeholders. This is particularly relevant in recent times in which meeting the interests of stakeholders has become a goal for all companies.

Specifically, this work makes several contributions to the literature, corporate governance practice and policy formulation, by giving reasons to revise some mechanisms of CG (regarding to structure and composition of the Board of Directors) with the aim that those contribute to integrate all stakeholders’ interests, enabling companies avoid the expropriation of wealth by the shareholders at the expense of other stakeholders.

First, our results show that greater board size contributes to increasing the value added distributed to shareholders. These results are consistent with research by Lipton and Lorsch (1992) and Jensen (1993), which allude to coordination and information problems linked to large boards. Accordingly, it would appear that in those cases in which the board has more members, shareholders control the board of directors’ decisions and impose their own interests regarding value-added distribution over those of other stakeholders. In this regard, our results reinforce the idea of an appropriate board size for achieving an effective corporate governance system, especially from a stakeholder perspective.

Second, with respect to board ownership, the results show that greater board ownership reduces the value added distributed to other stakeholders. This is consistent with the idea that directors align their interest to those of shareholders when they own shares and, consequently, their decisions tend to benefit shareholders interests against those of other stakeholders, exerting externalities on other stakeholders’ value (Tirole 2001).

These results advocate the limiting of board size and board ownership if companies and shareholders wish to reduce the differences between stakeholders in the distribution of value added, especially in contexts of ownership concentration.

Third, the results also indicate that the incorporation of independent directors on the Board is an important factor with regard to dealing with the interests of primary stakeholders in the context of ownership concentration and a unitary board system, as it also reduces the differences between stakeholders in the distribution of value added. This result is consistent with previous literature which asserts that a greater number of this type of director could compensate for the excessive power of executive directors and neutralize possible expropriation problems linked with the principal–principal conflict of interests (large against minority shareholders). In addition, independent directors would appear to be more sensitive to the interests and claims of stakeholders, thereby contributing to lowering the differences in value-added distribution. Therefore, consistent with works that promote board diversity, we show that independent directors are a source of diversity that could protect the interests of the corporation´s stakeholders.

Finally, according to our results, the separation of the roles of Chairman and CEO also reduces the differences in value-added distribution among stakeholders, which contributes to the defence of the interests of the primary stakeholders, reinforcing the criticisms of duality of power. Thus, despite the international controversy over the advisability of their separation or not, our results give reasons to believe that the separation of power is the better option for the interests of all stakeholders.

From the results obtained, we can conclude that the current board system in Spain fails to integrate all stakeholders’ interests through the CG system, as board ownership (with levels near to 23%) and duality (present in 63% of the observations) tend to encourage directors’ decisions that relegate the aspiration of some stakeholders to share in the wealth of the firm. Although García-Torea et al. (2016) found that effective boards of directors (studied as complete units) that promote practices that attend to the interests of shareholders (the shareholders perspective) are also effective in serving the stakeholders’ interests (the stakeholders perspective), our results reinforce the idea that certain board characteristics, particularly board ownership and board size, could in fact contribute to exacerbating the conflict of interests between shareholders and other stakeholders. This suggests then that board practices that affect all stakeholders’ interests should be revised. Future research, therefore, should focus on how CG practices could benefit certain stakeholders’ interests against the interests of others, in order to minimize these externalities and develop CG mechanisms to integrate the shareholders and stakeholders CG approaches into more comprehensive CG systems.

According to the above, this article provides a complementary perspective towards approaching corporate governance from a stakeholder perspective. From a theoretical point of view, the results of this study contribute to the debate concerning the dichotomized approach of CG (shareholders/stakeholders CG approach). Empirical evidence has been found concerning the impact of externalities caused by certain board characteristics on the wealth distribution among shareholders and other stakeholders. In this context, this study reveals that board effectiveness based on those characteristics that contribute to the protection of shareholders rights grants excessive power to shareholders who may abuse their power in pursuit of their own interests to the detriment of other stakeholders’ interests. At this point, we question whether it is necessary to establish other mechanisms to control the decision-making process of the board.

There are, nevertheless, some limitations and unobservable issues to be taken into account. First, due to the focus of our study, we have overlooked some CG internal and external control mechanisms, such as board diversity, board activity, the existence of corporate social responsibility committees, ownership distribution among different types of shareholders (institutional, non-institutional, banks, families, and so on), and the proportion of different types of directors on the board (outsiders, insiders, proprietary), among others. Second, we should highlight the difficulty of disaggregating certain information, as it is not available in the sources consulted and would thus require consulting other documents or databases. Thus, for example, it might be interesting to analyse other secondary stakeholders’ interests. Unfortunately, the public databases are unable to give us this information. In both cases, future research could analyse these issues to gain a better understanding of the factors that influence the distribution of value added among different stakeholders and how that could impact the development of a stakeholders’ CG approach.