Board Characteristics and Corporate Performance in Tunisia

Abstract

Drawing upon prior empirical research on the potential endogeneity of both ownership structure and firm performance in developed markets, this study examines the reverse causations that can exist between corporate performance and ownership structure in Tunisian listed companies. The study was extended to include another governance mechanism - board characteristics - as the principal internal control mechanism for monitoring managers and an assessment of its potential effects on firm performance and ownership structure. Our findings proved the existence of endogeneity and a two-way causality between ownership variables and MTB performance. However, our findings also revealed that corporate governance in Tunisian firms needed to be more strengthened based on board characteristics.

Keywords: corporate governance, ownership structure, board characteristics, corporate performance, endogeneity, simultaneous equations models.

Introduction

The relation between ownership structure and performance has been a matter for an important and ongoing debate in the corporate finance literature. The debate goes back to Berle and Means' thesis (1932). Since then, researchers have been interested in the effects of a separation between ownership and control of corporate enterprises. This separation creates agency costs because owners (principals) and managers (agents) have different objective functions (Jensen and Meckling, 1976). There are different monitoring mechanisms that, if implemented, should improve corporate governance. Furthermore, there are internal and external mechanisms. The external ones (e.g. market for corporate control) are particularly important in the Anglo-Saxon systems that experience dispersed ownership structures. The internal mechanisms (e.g. ownership concentration and board of directors) are, somewhat, predominant in emergent markets, including Tunisia. We investigate the Tunisian context that is completely different from the Anglo-Saxon one, where shareholders are mainly concerned with the value of their portfolios. This is true in most emergent countries, where ownership structures are very concentrated, institutional investor shareholding is very low and the board of directors represents the main organ of governance.

Our study presents a new evidence of the corporate ownership structure as a mechanism of corporate governance in Tunisia. First, we examine the influence of ownership structure (ownership concentration and insider ownership) on firm performance with particular attention to the non-linear relationship between managerial ownership and firm performance. We note that equity ownership by managers and monitoring by large blockholders are two ways that can potentially reduce the severity of the agency problem between managers and outside shareholders. Second, we analyze the impact of board characteristics on the performance of the firm. Finally, we evaluate the possible endogeneity between firm performance and ownership structure (concentrated ownership and managerial ownership). Thus, we consider ownership structure as endogenous and we try to identify its determinants. This is a useful contribution to the literature, as we are not aware of any other study, in the Tunisian context, of the endogeneity of ownership structure. In addition, endogeneity is important in practice but difficult and underexplored in empirical corporate governance research. Our findings proved the existence of endogeneity and a two-way causality between ownership variables and MTB performance. In addition, the monitoring mechanism (board of directors) has, above all, an insignificant influence on firm value. The last finding confirms the weakness of this mechanism in Tunisian listed firms.

The rest of the paper is organized as follows. Section 2 presents the institutional environment in Tunisia. In section 3, we briefly review the theoretical and empirical arguments defining the relationships that exist between governance aspects (ownership structure and board characteristics) and firm performance. We present the data and specify the variables used in our study in section 4. Our empirical results are given in section 5. Finally, our conclusion is presented in section 6.

Institutional environment in Tunisia

The Tunisian case presents many interesting features that make its study relevant in terms of policy recommendations for this country and for others in the Middle East and North Africa (MENA). We first present the principal characteristics of MENA region and then we present the main features of the Tunisian context.

2.1 Corporate governance in MENA regionA 

In most MENA countries, the financial markets are thin and rigidly regulated, government ownership is prevalent and the market forces have a limited role. Corporate governance in MENA economies present four main characteristics: concentration of ownership (due to large family dominated companies), family ownership and control (the market for corporate control is not active throughout the region), bank-based corporate finance (bank loans are the most important form of external finance) and underdeveloped capital markets (capital markets lag behind the banking sector and foreign participation is limited). Corporate governance in MENA countries is also characterised by a major role of the board of directors, the absence of call for a separation between the chair and the CEO, limited protection of the shareholders' rights and the absence of board independence. Corporate governance in MENA region needs to be recognised as a public policy concern of rapidly increasing importance in the region. However, the number of large companies contributing to growth is still limited.

2.2 Corporate governance in Tunisian listed firms

Corporate governance in Tunisia has generally improved in recent years. In fact, a tentative of corporate governance best practices has lastly (in June 2008) been implemented in Tunisia by the Arab Institute of Business Managers. However, the firm's commercial code has constituted up to now, the main reference concerning corporate governance. The fundamental principles of this code are the protection of the shareholders' rights, the equitable treatment of the shareholders (including the minorities) and the transparency and diffusion of information. Indeed, the board of directors (or the supervisory board and the executive one) constitute the main organ of governance in Tunisian firms. For both structures, the Law specifies the size, composition, responsibilities and tenure of its members. The board of directors (BOD) is composed of at least three members and of twelve members at most. A shareholder is not required to sit on the board, indeed an employee of the firm can be nominated as a member. Directors' tenure is limited to three years that can be renewed. The BOD elects the CEO who must be a shareholder. The BOD is vested with the most extended powers to act in all circumstances in the name of the firm. Besides, they cannot deliberate any valid decision without at least half of their members being present or represented.

The second main feature of Tunisian corporate governance is ownership concentration. In fact, the Tunisian system is characterised by the presence of strong blockholders (often including families) which leads to agency problems between blockholders and minority shareholders. In fact, 80% of total shares are held by the five largest shareholders (Omri, 2003). In addition to families, the state and banks hold an important part in the shares of the listed firms. However, the participation of the state is becoming more limited due to privatization. Moreover, foreign investors can participate within the limits of 50 percent of the offering of a company. Nevertheless, they hold less than 10 percent of all stocks. Finally, the participation of institutional investors is still very limited.

2.3 Tunisian stock market

The Tunisian financial market consists of the semi-privatized Tunis Stock Exchange (TSE). The government intended the TSE, opened in 1990, to serve as an additional source of funding through the mobilization of domestic and foreign investments. The TSE is regulated by a state-run watchdog, the Financial Market Council. The stock market was reformed according to international standards in 1994. Later, the TSE was privatized in 1995 and became the central part in the Tunisian financial market. The Tunisian financial system is dominated by banks with domestic bonds. However, equity markets are playing a limited role in savings mobilization. In Tunisia, the government bonds represent the majority of tradable securities, principally due to the absence of companies that are able to make public offering securities.

On the other hand, the most widely used and modern company structures are the limited liability companies and joint stock companies. Only the latter ones may issue shares to the public and thus be listed on the TSE. Once primarily based on agriculture, the Tunisian economy has become more varied with important manufacturing and tourism sectors. The number of firms listed on the TSE increased from 13 to 45 between 1991 and 2003. At the end of 2007, only 50 companies were listed. Hence, the market capitalization increased from TND 610 million in 1991 to TND 2.98 billion (approximately US $ 2.34 billion) in 2003. At the end of 2006, it reached TND 5.49 billion (approximately US $ 4.32 billion). Although growing in importance, the TSE remains small and relatively illiquid. For these reasons and in order to increase the market activity, many reforms have been implemented. Among these reforms, the government implements fiscal advantages to encourage companies to list their stock on TSE, grants financial incentives to make the stock market more attractive and allows foreigners to buy up to 50 percent of a firm's shares on TSE. Despite these important incentives, there has been no inherent interest for firms in being listed on the stock market.

Literature review

Ownership structure is a central and distinguishing theme in the corporate governance literature. We consider two issues of shareholding that are addressed in research: ownership concentration and insider ownership. In addition, the relation between board characteristics and firm performance remains a fundamental issue in the corporate governance literature.

3.1 Ownership concentration and firm performance

Ownership concentration is considered as a key governance mechanism. Empirical research has examined the importance of block or large shareholdings in controlling managers, and hence reducing agency costs. The results of these researches are mixed. The expected performance effect of ownership concentration is unclear, as it reflects the net impact of benefits and costs. Following Berle and Means (1932) and until the eighties, the literature had focused on the advantages of ownership concentration. The main concern was the cost of the separation between ownership and control, or the agency costs (Jensen and Meckling, 1976). The idea is that dispersed ownership in large firms increases the principal-agent problem due to asymmetric information and uncertainty. A review of the studies on the relationship between ownership and performance indicates, in general, a higher profitability in owner-controlled firms as compared to manager-controlled firms (Short, 1994). There is a widespread consensus that a higher degree of control by an external shareholder enhances productivity performance. Chen et al. (2005) report a weak relationship between ownership concentration and firm performance. Shleifer and Vishny (1986) show that large shareholders have the incentive to monitor firm management, and that the presence of large shareholders enhances firm performance. Thomsen and Pedersen (2000) find a positive relation between ownership concentration and firm performance. However, the relation is nonlinear indicating that concentration has adverse effects on performance after a certain level. Cho (1998) does not detect any significant link between firm value and shares held by large shareholders. Denis and Denis (1994) analyze samples of U.S. firms with low and high ownership and find that there is no significant difference in the firm value between them. Other similar international studies (e.g. Minguez-Vera and Martin-Ugedo (2007) for Spain) report insignificant relationships between concentrated ownership and performance.

Large shareholders are assumed to play a monitoring role that raises the value of the shares for the minority. However, concentrated ownership also gives the insiders (owners and managers) the opportunity to expropriate efficiently (La Porta et al., 2000). Consequently, monitoring managers is not the main problem of corporate governance and the real concern is the risk of expropriation of minority shareholders. Ownership concentration may lead to the extraction of private benefits by controlling blockholders at the expense of minority.

3.2 Managerial ownership and firm performance

The conflict of interest between managers and outside shareholders is another version of agency problem that is less likely to be observed in a highly concentrated equity ownership structure. The extent of managerial shareholding affects the congruence between managers and shareholders (Jensen and Meckling, 1976). Whereas the primary governance function of outside owners is to monitor management, a larger insider stake reduces the need for such control. Managerial ownership is one way to align the objective functions of the owners and of the managers. The convergence of interest hypotheses predicts that insider holdings and economic performance are positively related. From this incentive effect, a positive association between managerial ownership and firm performance is expected. In the same vein, Krivogorsky (2006) finds a positive relationship between managerial ownership and firm performance. However, this relationship remains insignificant. Cornett et al. (2008) detect a positive and significant association between managerial ownership and performance. On the other hand, if managers hold large shares of the equity, it becomes more difficult for outside owners to exercise control. This entrenchment effect is especially important for high shares of managerial ownership. In this case, managers will not probably maximize firm value and a negative relationship between managerial ownership and firm performance is possible. Switzer (2007) finds a negative and insignificant relation between CEO ownership and Tobin's Q using 3SLS regression.

Moreover, Morck et al. (1988), among others, show that there is a complex relationship between agency costs and managerial shareholdings. Taking the incentive hypothesis and the entrenchment one into account, a non-linear relationship between management's ownership share and firm performance is expected. At low levels of ownership, the incentive effect is probably dominant, that is, a positive effect is expected. However, at very high levels of ownership, the entrenchment effect might be more important and the effect of ownership on performance could be negative.

3.3 Board characteristics and firm performance

Fama (1980) argues that the board of directors is the central internal control mechanism for monitoring managers that help control agency problems. The board plays a major role in the corporate governance framework and is responsible for monitoring managerial performance and preventing conflicts of interests.

3.3.1 Board size

Agency theory predicts that because groups communicate less effectively beyond a certain size, there is pressure from self-serving managers or entrenched owners to expand board size beyond its value maximizing level. The implication is an inverse relationship between board size and performance. Empirical research also reports conflicting results concerning the association between board size and performance. For instance, Yermack (1996), Eisenberg et al. (1998) and Cornett et al. (2008) report a negative relation. Cheng (2008) shows that larger boards are associated with lower performance. This association is consistent with the view that both coordination/communication problems and agency problems become more severe as a board grows larger. Conversely, Bhagat and Black (2002), Chen et al. (2005) and Black et al. (2006) do not find a statistically significant association. Kiel and Nicholson (2003) and Adams and Mehran (2005) find positive board size effects on firm performance. Accordingly, a larger board size brings more resources to firms and, therefore, might improve their performance.

3.3.2 Board composition

In order to effectively fulfil their monitoring role, boards must have some degree of independence from management. Indeed, outside directors can play an active role in arbitrating in disagreements between internal managers and help reduce agency problems between managers and residual claimants. Several studies test for the effect of outside directors' representation on the board on performance and the results are mixed. Rosenstein and Wyatt (1990) find that stock markets react positively to the appointment of outside directors. Using data from New Zealand, Hossain et al. (2000) also find a positive relationship between higher levels of board independence and firm performance. Chung et al. (2003) find that board independence affects performance positively through the ability of outside directors to provide effective management-monitor activities. However, Bhagat and Black (2002) find a negative association between the proportion of outside directors and firm value. On the other hand, Adams and Mehran (2005), Prevost et al. (2002) and Connelly and Limpaphayom (2004) do not find a statistically significant relationship.

3.3.3 Board leadership structure

When the CEO is also the chairperson, the capacity of the board to monitor the CEO is weaker (Jensen, 1993). Gul and Leung (2004) suggest that CEOs who also serve as board chairpersons could reduce the board's ability to exercise effective control over management and thereby negatively affect performance. Brickley et al. (1997) argue that there are also costs associated with having two persons holding the CEO and chairperson titles. They find no evidence that firms with separate persons holding the CEO and chairperson titles perform better than those with the same person holding both titles. In contrast, Pi and Timme (1993) find that firms with one person holding both titles have less cost efficiency and performance than those with two persons holding the two titles. Krivogrsky (2006) and Cornett et al. (2008) detect a negative relation between CEO duality and firm performance. Chen et al. (2005) find the same relation when they use Market-to-Book. However, they find insignificant relation when they use ROA and ROE as measures of performance.

Data and methodology

4.1 Sample and data

Our sample includes an unbalanced panel of 23 non-financial companies listed on the TSE during the period 1998-2006. Companies from the banking and financial sectors are removed from the sample. The financial services sector includes several trusts, which have distinctive different corporate governance structures from other firms. In addition, financial firms are subject to a regulatory framework which may affect the ownership-performance relationship. Our sample size includes 181 firm-year observations with a predominance of industrial firms (61%) in comparison to firms from the service sector (39%). Moreover, the service sector includes firms in communication, transportation, tourism, etc. Company annual reports, obtained from the Financial Market Council, are used as the source for the shareholding structure and for board characteristics. Financial and accounting data come from the TSE.

4.2 Definitions of the variables

Our specification of the variables used in this study draws on the current literature in empirical corporate finance, though we are somewhat constrained by data availability for Tunisian firms.

4.2.1 Performance variable

We use the natural logarithm of the Market-to-Book ratio (MTB) in order to measure firm performance. MTB ratio is defined as the market value of equity divided by the book value of equity. The market value of equity uses closing prices for shares on the last trading day of the year. This ratio is also an indicator of growth opportunities. It is important to note that the MTB ratio is a permanent valuation indicator of choices of the firm, of management and of strategic perspectives. It also depends on the anticipations of the investors.

4.2.2 Ownership structure

It is represented by ownership concentration (CO) and managerial ownership (MO). The former is measured by the Herfindhal index calculated by summing the squared percentages of shares held by the three largest shareholders. The latter is defined as the percentage of ordinary shares held directly by the CEO and executive directors.

Based on the discussion presented above, we would expect to find either positive or negative effects of ownership variables (ownership concentration and managerial ownership) on firm performance.

4.2.3 Board characteristics

They are represented by board size, board composition and board structure.

Board size (BS) is measured as the total number of directors on board. Lower board size is associated with greater firm performance. Hence, we expect a negative relationship between these two variables.

Board composition (BC) is defined as the number of outside directors on the board divided by the total number of directors. Greater board independence (outside directors) is associated with greater firm performance. This leads to predict that board composition is positively related to firm performance.

Board leadership structure (CEO) is a dummy variable that equals 1 if the CEO and the chairman are different persons (i.e. separation of functions) and 0 otherwise. In fact, the value of 1 strengthens the level of independence between the board and the leadership. We expect board leadership structure to be positively associated with firm performance.

4.2.4 The control variables

Firm size (LTA) is measured by the natural logarithm of total assets. Firm size may affect both firm performance and ownership structure of listed companies. Large firms benefit from scale economies and have better access to financial resources. As a result, firm performance will be improved. Therefore, we expect firm size to be positively related to firm performance. On the other hand, larger firms require more investment from an owner of a given proportion of shares. Therefore, the amount that has to be invested for a given share should be larger. Hence, previous studies, such as Demsetz and Villalonga (2001), usually argue that ownership concentration and managerial ownership decrease with the size of the firm, which we expect to be negatively associated, with the ownership variables.

Leverage (LEV) is defined as the ratio of total debt to total assets. There is some ambiguity in the expected sign of debt. The signalling hypothesis proposed by Ross (1977) and the free cash flow theory of Jensen (1986) support a positive relationship. Alternatively, the pecking order theory proposed by Myers and Majluf (1984) predicts a negative association between debt and firm performance. Indeed, as far as the Tunisian firms are concerned, we could argue a negative effect of leverage on firm performance. Tunisian listed companies are heavily in debt. Otherwise, in the ownership equations, leverage serves to reflect the creditor's monitoring of management. In this case, debt can discourage managers to entrench themselves in large shareholdings. In addition, larger value of debt is associated with lower percentages of shares owned by the largest shareholders. Based on these arguments, we predict that leverage is negatively related to firm performance and to the ownership variables.

Liquidity (LIQ) is defined as net working capital as a fraction of the book value of assets. We use this variable in the ownership equation as a control variable. Indeed, greater liquidity is associated with greater agency costs (Gonenc, 2006). This means that management is not monitored effectively. Consequently, if the management and shareholders have aligned interests, we should observe a negative relationship between liquidity and ownership concentration (or managerial ownership). Nevertheless, if the interests diverge, the relationship must be positive.

Market risk is assessed by the beta coefficient (BET) of the stock. Market risk measures the risks inherent in stock ownership; it is estimated by the market model through a regression of the monthly return on a stock. Therefore, variation in risk causes variation on ownership structure. In fact, higher market risk signifies higher return for shareholders. When shareholders (concentrated owners or managers) can effectively make predictions about future economic conditions, the market risk makes higher ownership less costly (Demsetz and Villalonga, 2001). Accordingly, the estimated sign of BET should be positive. However, when concentrated owners and managers are risk averse, they tend to reduce their participation. In this case, the estimated sign of BET must be negative.

4.3 Methodology

Our primary objective is to evaluate the relationship existing between firm performance and ownership structure. To do this, we developed two different models to check the relation between our different variables. In the first model, we used a linear regression to examine the simple links. We developed two sets of equations. The first set had firm performance as the dependent variable. The ownership concentration (managerial ownership) was the independent variable in the first (second) equation. The second set included ownership variables as the endogenous variables in two different equations in which the firm performance was the exogenous variable. Moreover, several previous studies had proved that the association between managerial ownership and managerial ownership was not linear. For this reason, we included the squared value of managerial ownership in the performance equation in order to check the curvilinear relation between these two variables.

The first model (Model 1) came as follow:

MTBit = β0 + β1 COit + β2 BSit + β3 BCit + β4 CEOit + β5 LTAit + β6 LEVit + εit (1)

MTBit = β0 + β1 MOit + β2 (MO)2it + β3 BSit + β4 BCit + β5 CEOit + β6 LTAit + β7 LEVit + εit (2)

COit = β0 + β1 MTBit + β2 BSit + β3 BCit + β4 CEOit + β5 LTAit + β6 LEVit + εit (3)

MOit = β0 + β1 MTBit + β2 BSit + β3 BCit + β4 CEOit + β5 LTAit + β6 LEVit + εit (4)

In the second model, we considered ownership structure as an endogenous variable that could be influenced, among other observed factors, by firm performance. Consequently, the relationship between ownership structure and firm performance should be detected in the two ways. Hence, the consideration of endogeneity requires working with simultaneous equations models. For this reason, we developed a system of three simultaneous equations. To estimate this system empirically, we employed the two-stage least squares (2SLS) approach. The first equation relates to firm performance as a dependent variable while the second and the third ones relate to concentrated ownership and managerial ownership as dependent variables. To test our hypotheses, the following model (Model 2) was used:

MTBit = β0 + β1 COit+ β2 MOit + β3 (MO)2it + β4 BSit + β5 BCit + β6 CEOit + β7 LTAit + β8 LEVit + εit (1)

COit = β0 + β1 MTBit + β2 LIQit + β3 BETit + β4 LTAit + β5 LEVit + εit (2)

MOit = β0 + β1 MTBit + β2 LIQit + β3 BETit + β4 LTAit + β5 LEVit + εit (3)

Empirical results

5.1 Univariate tests

Summary statistics relating to our different variables are represented in Table 1 that reports the mean values, standard deviations, minima and maxima of variables for the 181 observations.

Table 1: Descriptive statistics

Variable

N

Mean

Standard Deviation

Minimum

Maximum

MTB

CO

MO

BS

BC

CEO

LTA

LEV

LIQ

BET

181

181

181

181

181

181

181

181

181

181

0.386

0.422

0.026

9.215

3.685

0.254

17.784

0.561

0.184

0.664

0.597

0.179

0.062

1.782

1.195

0.436

0.917

0.486

0.275

0.623

-1.049

0.113

0

5

1

0

16.148

0.046

-0.687

-0.283

1.999

1

0.462

12

5

1

21.102

3.564

0.746

2.641

Mean values of ownership concentration and managerial ownership were 42.2 % and 2.5 %, respectively. The Market-to-Book ratio had a mean value of 0.386. We note that the boards of directors of Tunisian firms had, 9.2 directors, on average. In addition, there were only 3.68 independent non-executive directors on our sample boards representing, 40 % of board membership, on average. It is important to note that this number probably overstates the fraction of really independent directors given that the corporate sector in Tunisia is small and it may be difficult to hire true independent outsiders. Moreover, this increases doubts about the independence and robustness of boards and about the possible positive impact of these directors on firm performance in case they actually have any effective monitoring functions. Furthermore, in only 25.4 % of all firms in our sample, there was no CEO duality, which means that the CEO was also the chairman of the board in 74.6 % of Tunisian firms.

Table 2 shows the correlation coefficients of the independent variables. The correlations that should be noted are the (-0.66) correlations between leverage (LEV) and liquidity (LIQ) and the 0.56 correlations between board size (BS) and board composition (BC). None of the remaining variables is correlated to a level that is worth noting.

Table 2A : Correlation coefficients of independent variables

CO MO BS BC CEO LTA LEV LIQ BET

CO 1

MO -0.199** 1

BS 0.257** 0.074 1

BC 0.194** -0.195** 0.561** 1

CEO -0.117 0.43** 0.143* -0.037 1

LTA 0.147 * -0.168* 0.367** 0.259** -0.262** 1

LEV 0.321** -0.078 0.133* 0.311** -0.129* 0.148* 1

LIQ -0.029 -0.031 0.030 -0.016 0.117 -0.318** -0.669** 1

BET -0.076 -0.028 0.085 -0.178* -0.217* 0.319** -0.249** 0.182* 1

**, * denotes significance levels at 1% and 10%, respectively.

To test for multicollinearity, we computed the variable inflation factor (VIF) for each variable. The results showed that there were no variables included in the tests with VIF > 1.95. Thus, multicollinearity did not to seem to be a problem for the results.

5.2 Multivariate analysis

5.2.1 Linear regression results

Linear regression results are reported in Table 3. By focusing on firm performance equations, we remark that MTB ratio decreases with ownership concentration and increases with managerial ownership. These findings reflect the disadvantages (or costs) of ownership concentration and the benefits of managerial ownership. On the one hand, ownership concentration deteriorates the firm market performance and, thus, does not reduce the conflicting interests between large and minority shareholders. Actually, it is argued that ownership concentration increases the majority-minority conflicts (Shleifer and Vishny, 1997) and reduces the market liquidity.

Table 3A : Linear regression

MTB MTB CO MO

MTB -0.02 0.005

(-0.20) (0.99)

CO -2.66

(-2.06)**

MO 0.43

(2.52)**

MOA² -0.833

(-0.27)

BS -0.349 -0.34 0.005 -0.005

(-1.83)* (-1.77)*** (0.94) (-2.16)**

BC 0.22 0.87 0.005 0.006

(1.33) (1.32) (0.41) (1.17)

CEO -0.18 -0.328 -0.07 0.06

(-0.19) (-1.49) (-1.69)* (2.05)**

LTA 0.129 0.149 -0.03 -0.003

(1.69)* (1.89)* (-1.77 * (-1.22)

LEV -0.448 -0.424 -0.02 -0.001

(-3.29)*** (-3.03)*** (-0.74) (-0.21)

Const -2.178 -2.451 0.47 -0.08

(-1.58) (-1.71) (1.43) (-0.06)

R-sq 0.295 0.263 0.129 0.258

The t-statistics are in parentheses below estimated coefficients.

***, **, * denotes significance levels at 1%, 5% and 10%, respectively.

On the other hand, MO decreases conflict of interest between managers and outside shareholders. This result is consistent with the theory and with previous studies stating that agency problem between managers and outside shareholders is less observed in highly concentrated equity ownership structure (Davies et al. 2005). It is important to note that the majority of Tunisian firms are family-owned companies. In addition, the manager (CEO) is, in general, a family member; if not, he is well controlled by the family members represented on the board. For this reason, large shareholders and/or managers worry about their own interests which are strongly related to the firm performance. Furthermore, we test for a curvilinear relationship between firm performance and managerial ownership. Thus, our estimation shows that MOA² affects negatively but insignificantly the MTB ratio. This finding refutes the hypothesis of curvilinear relationship between managerial ownership and firm performance.

With respect to the board variables, BS and CEO separation are negatively related to MTB ratio, and BC presents a positive coefficient. However, only BS is statistically significant. This last result casts doubt on the effectiveness of independent non-executive directors as monitoring mechanism in Tunisia and may be due to the difficulty in recruiting truly independent board members in a small market because there are generally interrelationships (family, friendship...) between internal and external board members. Chen et al. (2005) reported a negative coefficient for CEO and positive coefficients related to BS and BC.

Considering the third and fourth equations where firm performance is the independent variable in addition to board variables. MTB ratio was found to be related negatively to ownership concentration and positively to MO, but these associations are insignificant. The first finding proved that Tunisian shareholders were less concerned with the market performance. In fact, Tunisian shareholders are not well informed about the advantages of market performance. Hence, they are not able to accept it. Although these relationships are not significant, they show that there is a reverse causation between ownership structure and firm performance. To check this possibility, we use the simultaneous equations analysis.

On the other hand, BS and BC affect positively and insignificantly ownership concentration. In addition, we found that board leadership structure (CEO) appeared to be negatively and significantly related to concentrated ownership. The negative association between CEO and CO means that when there is a separation of functions of management and control, the interests of the largest shareholders are not maintained. The explanation that can be suggested by this result is that the board as a monitoring mechanism takes into account the interests of all shareholders including the minority ones, and this may prevent the opportunity to expropriate. Furthermore, we remark that only BS and CEO had a positive significant relationship with managerial ownership (MO). This corroborates the finding in Mak and Li (2001).

Finally, the reported results show that firm size (LTA) has a positive impact on MTB ratio. This result is consistent with previous studies stating that greater performance should be associated with larger firms (e.g. Davies et al. 2005). In the MO and CO equations, LTA presents again the predicted sign. In fact, the larger a firm is, the larger the capital sum investors need to hold will be. Thus, it would be logical that managerial ownership and concentrated ownership decrease with firm size. Concerning leverage, it was found to be negative in all regressions. However, it is only significant in performance equations.

5.2.2 2SLS regression results

2SLS estimation results are reported in Table 4. By focusing on the first equation, ownership concentration is shown to be negatively related to MTB ratio. This finding reflects the fact that even if large shareholders perform efficiently their monitoring functions on management in order to improve firm performance, they are not still interested in market performance and this is explained by the culture of the Tunisian investor who is not yet ready to accept the advantages of market performance. However, this association is not significant.

Table 4: 2SLS regression

MTB CO MO

MTB -0.05 0.008

(-2.04)** (3.52)***

CO -9.82

(-0.45)

MO 137.61

(3.34)***

MOA² -23.16

(-0.64)

BS 0.08

(2.80)***

BC -0.09

(-1.75)*

CEO -3.69

(-1.27)

LTA 0.07 -0.12 0.06

(0.08) (-2.22)** (1.94)***

LEV -3.94 0.09 -0.02

(-2.86)** (2.25)** (-1.16)

LIQ 0.30 -0.05

(2.50)** (-2.41)**

BET -0.09 -0.02

(-3.30)*** (-1.11)

Const -3.32 0.37 0.17

(-0.14) (0.80) (2.00)

R-sq 0.362 0.213 0.193

The t-statistics are in parentheses below estimated coefficients.

***, **, * denotes significance levels at 1%, 5% and 10%, respectively.

On the other hand, BS was found to have a positive impact on firm market performance. On the contrary, CEO and BC affected the market performance negatively. These results show, on the one hand, that the separation of functions is associated with a lower level of market performance. On the other hand, and once again, there are some doubts on the effectiveness of independent non-executive directors as a monitoring mechanism in Tunisian firms. However, only CEO's coefficient is insignificant. Our results show that corporate governance mechanisms (BC and CEO) had proved to be weak in Tunisian companies. Actually, larger boards and the absence of truly independent non-executive directors may result in ineffective corporate governance. MO is positively and strongly associated with MTB ratio. Alternatively, the positive coefficient may be explained by the alignment hypothesis. In addition, the insignificant impact of the squared value of MO on MTB ratio disproved the nonlinearity relation between these two variables.

By focusing on the ownership concentration regression, we note that all coefficients of different variables are significant. We show that MTB ratio has a negative impact on ownership concentration. This result confirms the one found in the firm performance regression. Market risk (BET) affects CO negatively. This result is consistent with the fact that when the largest shareholders do not effectively make predictions about future economic conditions, the market risk makes higher ownership more costly. The last regression relating to managerial ownership indicates that MTB ratio is found to be positively related to the MO. This result implies that firm market performance increases with the importance of shares held by managers. BET is shown to have a negative and insignificant effect on managerial ownership. This finding implies that managers are risk averse and they would like to reduce their participation as long as the market risk increases.

Finally, for the 2SLS estimation of the ownership equations, we pay attention to the role of liquidity in explaining variations in ownership structure. In the CO regression, LIQ presents a positive coefficient. This result suggests that the largest shareholders and minority shareholders have divergent interests. In the MO regression, LIQ presents a negative sign. This result suggests that management and shareholders have aligned interests. Our finding is inconsistent with the one found by Davies et al. (2005) who reported a strong positive association between firm level liquidity and managerial ownership.

Overall, our results indicate that although ownership variables (concentrated ownership and managerial ownership) are determined by firm performance, firm performance itself is partly determined by these variables. Those findings show that there is a reverse causation between firm performance and ownership structure in which these variables can be influenced, among other observed factors, by each other. Therefore, our findings prove the endogeneity of ownership structure.

Conclusion

Our study examines the roles played by two aspects of ownership structure that is ownership concentration and managerial ownership. In addition, we included board characteristics as another mechanism from corporate governance theory. We analysed a sample of 23 listed Tunisian firms over the period 1998-2006 and we tried to answer our core question: does ownership structure affect firm performance and is it as well affected by it?

Our empirical analysis under linear regression shows a negative (positive) and significant association between ownership concentration (managerial ownership) and MTB performance. However, we find that MTB has a negative (positive) but insignificant impact on ownership concentration (managerial ownership). Accordingly, these results led partly to the conclusion that there was a reverse causation between ownership structure and firm performance. Hence, our 2SLS regression proved the existence of endogeneity and of a two-way causality between ownership variables and firm performance. Concerning board characteristics, board leadership structure (CEO) is negatively associated with ownership concentration. This finding revealed that CEO duality was much more likely in family-owned and controlled firms whereby the ownership was concentrated. In addition, we also found that CEO had a negative impact on MTB. Our findings showed that board composition presented, in all linear regressions, a positive and insignificant coefficient. This result casts doubt on the effectiveness of independent non-executive directors as a monitoring mechanism in Tunisia. It may be due to the difficulty recruiting truly independent board members in a small market because there are generally interrelationships between internal and external board members. Concerning board size, we found that it was positively and insignificantly related to ownership concentration. For the other three regressions, it presented a negative and significant sign. Therefore, our findings concerning board characteristics provide evidence that the board, as the principal monitoring mechanism in Tunisian firms, appears weak. These conclusions arose from linear regressions relating to board characteristics were confirmed under 2SLS regression.

This paper provides two major contributions. First, we provide the evidence that endogeneity and the two-way causality between ownership variables and firm performance already exist in Tunisian listed firms. This means that variation in firm performance affects variation in ownership structure (ownership concentration and managerial ownership) and is also affected by it. Second, our findings disproved the hypothesis of curvilinear relationship between managerial ownership and firm performance. Indeed, our analysis extends previous studies in this area and failed to capture this nonlinear association. This is may be due to the insignificant level of shares hold by managers in Tunisian listed firms. Despite these findings, we failed to fully explore alternative specifications of board characteristics in relation to corporate ownership and corporate performance. One implication that can be derived from this analysis is that corporate governance in Tunisian firms needs to be more strengthened. In addition, more effort is needed in order to ensure the true independence of non-executive directors to make sure that they are able to perform an adequate monitoring function. Another potential explanation may be in the endogenous relationship between firm performance and board characteristics. Future studies in this area may focus more on these points and should detect the endogenous role that can be played by board characteristics.

Cite this page