This paper studies the ownership - liquidity relation in the context of the Tunisian Stock Exchange. In particular, we examine two empirical relationships: The relationship between ownership concentration and stock liquidity and the relationship between the separation of ownership from control and market liquidity .Our empirical findings verify that the structure of ownership remains concentrated in the majority of the Tunisian firms. We find that stock liquidity decreases significantly with concentrated ownership. Different devices used to gain control and hence a significant separation of ownership from control affect liquidity in different ways. Our results indicate that pyramidal structures have a significant negative impact on liquidity for all controlled firms. However, for family firms, nonvoting shares increase liquidity for minority shareholders by reducing the probability of informed trading. Overall, our study reports that nonvoting shares may be a liquidity enhancing device for family - firms.
Key words: ownership, liquidity, controlling shareholders, pyramidal structure, voting rights, bid - ask spread.
Introduction
The relationship between market microstructure and corporate finance has received recently considerable attention in financial literature, focusing on the problem of how corporate governance could be associated with market liquidity. In particular, corporate governance, market liquidity and their effects on the firm's value are usually examined separately. However the informational and operational characteristics of liquidity have a much influence on shareholder value, especially when they interfere with corporate governance mechanisms. Prior research shows that high market liquidity results in lower cost of capital and therefore higher firm values. Liquidity generally refers to the ability to engage in rapidly trading a large number of securities at a low cost with little impact on market prices.
Market microstructure theory predicts that informational benefits are reflected in market liquidity through higher trading costs; a large fraction of shareholders exploit information that others would not have when investor interests are poorly protected; this access to private information increases the adverse selection component of spread, a wider bid -ask spread and a lower quoted depth. The focus of most empirical studies on the ownership-liquidity relation is on the effect of ownership structure on market liquidity, while there is a little empirical work shedding light on the effect of separation of ownership from control on liquidity.
Considering this relation, researchers focus on two hypotheses: the adverse selection hypothesis and the trading or free-float hypothesis. The adverse selection hypothesis posits that when informed shareholders possess superior information compared to outside shareholders, information asymmetry arises, reducing market liquidity (e.g., Grossman and Stiglitz, 1980; Glosten and Milgrom, 1985; Kyle, 1985; Easley and O'Hara, 1987). In practice, it is difficult to classify market participants as informed and which as uninformed. According to the Securities and Exchange Commission SEC controlling shareholders at the 10 % threshold are considered as insiders. Institutions, while not having the same access to private information as insiders ,can create an informational advantage by exploiting economies of scale in information acquisition and possessing. Following the SEC definition of insiders, this study focuses on controlling shareholder ownership for the Tunisian firms.
The trading hypothesis, on the other hand, supposes that large stakes by blockholders reduce the availability of floating shares; therefore the monitoring role of insiders'blockholders has a high cost in terms of market liquidity. Specifically, when a firm's ownership is concentrated, there are fewer trade, therefore the free-float is limited and liquidity is reduced (Demsetz, 1968). Furthermore, ownership structure may affect liquidity through the production of information. Moreover, ownership concentration decreases the benefits of monitoring the firm by stock market participants, thereby reducing the amount of public information available about the firm (Holmström and Tirole, 1993).
Given these two hypotheses, this paper analyses the effect of ownership structure and corporate governance on market liquidity .Using a sample of Tunisian firms ,we examine on the one hand the relation between concentrated ownership and stock liquidity .On the other hand ,we investigate the impact of separation between control and ownership on adverse selection and therefore on liquidity ; specifically , we investigate the impact on liquidity of how the firm chooses to enhance control .The two main ways to achieve separation of voting rights from cash flow rights are pyramidal structures and different classes of shares that provide different voting rights for given cash flow rights .
We think that it would be of great interest to conduct this study for many reasons. First it contributes to extent existing empirical work on emerging markets by examining a new database given by the case of the Tunisian Stock Market. This lead us to identify if the empirical results concerning other market hold for the Tunisian Stock Exchange .Second , the focus of most empirical studies on the ownership-liquidity relation provide evidence on quote-driven market . It would be interesting to provide a contribution to the literature by releasing new ultimate ownership data for a sample of firms listed on pure agency market such as the Tunisian Stock Market , and documenting the nature of controlling owners and the separation of ownership from control . Finally, the findings may be helpful for market participants to understand the influence of trading practices on stock price and for stock exchange authorities to adopt optimal regulatory policies and choose efficient trading systems in response to information asymmetry.
Consistent with the trading hypothesis, we find a positive impact of free float on liquidity. Stocks with a greater deviation between control and ownership, have a positive impact on spreads .However, this impact on spreads of ownership-control deviation depends on the control devices used by the firm. Pyramidal structures are positively related to spreads as well as on the adverse selection component of the bid-ask spread and therefore have a negative impact on liquidity. On the contrary , nonvoting shares decrease spreads. In accordance to the adverse selection hypothesis, pyramidal structures enhance information asymmetry component of the bid -ask spread and decrease liquidity for all pyramiding firms ,whereas nonvoting shares prevent insiders to trade on their private information ,they reduce information asymmetries and therefore the bid - ask spread ; this positive effect is limited to small family -firms .
The remainder of the paper proceeds as follows. In Section 1, we start with the literature review. In section 2, we describe the institutional and legal environment of the Tunisian firms' ownership, and present the structure of the TSE. In section 3, we describe data. In section 4, we report liquidity measures and methodology. In section 5, we analyze empirical results. Section 6 concludes.
I. Literature review
Berle and Means (1932) argue that professional managers are effectively in control of widely - held firms at the expense of shareholders. However, when one shareholder takes control of management by holding a large block of shares, he can exploit other shareholders (Shleifer and Vishny (1997), La Porta and al. (1998, 1999 and 2000)). To increase the chance of executing his plans, the large shareholder would minimize and delay the disclosure of information so that other shareholders can not intervene, or must base their decisions on inadequate information. Poor disclosure worsens the information asymmetry problem, and the large shareholder may even trade on his insider information to extract the private benefits of control. Keeping other factors constant, these will result in a wider bid - ask spread and lower stock liquidity.
Prior empirical studies, have attached great importance to the relation between stock liquidity and the fractional ownership of insiders and institutions. Glosten and Milgrom (1985) document that one cause of illiquidity is the presence of privately informed traders. Bhide (1993) further reports that active stockholders who reduce agency costs by monitoring managers may also reduce stock liquidity by increasing informational asymmetries. Most of these studies, lead to ambiguous results, partly because they do not distinguish between institutional and insider blockholdings. Sarin, Shastri and Shastri (2000) report that higher institutional and insider ownership are both associated with wider spreads and smaller quoted depth. Heflin and Shaw (2000) find a positive association between spreads and block ownership (including insiders and institutional blocks).Along a similar vein, Dennis and Weston (2001) show a negative relation between spread and both insider and institutional ownership , whereas Kini and Mian (1995) document no support for a significant relationship between spreads and blockhodings. Rubin (2007) finds a positive relation between spread and institutional blockholdings, whereas spreads is negatively related to both institutional and insiders 'holding. In Australia, Comerton-Forde and Rydge (2006) report that insiders 'holdings greater than ten percent of issued capital reduce liquidity, whereas insiders' holdings of less than ten percent enhance liquidity. Institutional blocks have no impact on liquidity. For Norway, Naes (2005) reports a positive relation between spreads and block ownership, but institutional ownership concentration has no effect on spreads.
To examine the nature of ultimate controlling owners and the means used to enhance control, La Porta, Lopez-de-Silanes and Shleifer (1999a) show that ownership and control can be separated through deviation of one-share-one-vote, pyramiding and cross-holdings to the benefit of the large shareholders. Claessens, Djankov, and Lang (2000) confirm a significant separation of ultimate ownership and control, and report the overwhelming control of wealth by a small number of families. In the same way, Faccio and Lang (2002) argue that families are the most pronounced type of controlling shareholders in Western European countries and find a significant concentration of wealth within a small number of families.
A discrepancy between ownership and control implies higher agency costs, as those in control exploit their superior information. Claessens and al. (2002) document that deviation of control from ownership leads to agency costs that decrease firm value. Furthermore, using a Canadian sample of publicly - traded firms, Attig , Fong , Gadhoum and Lang (2006) report that the deviation between ultimate control and ownership widens the bid-ask spread .Control is often enhanced via devices like corporate pyramids , cross - holdings , multiple class shares ,etc .A pyramid structure allows the ultimate shareholder to own and to control indirectly some shares of a firm ,but does not control directly the intermediary firms. Cross-ownership allows separation but increases the difficulty to evaluate the integrated ownership and control of a company. Both devices allow separation. However , dual-class shares, allows insiders to hold superior voting rights shares ,holding a lower fraction of shares .This implies, that they may not sell shares with inferior or no voting right. As a consequence, the probability to trade against insiders may be reduced and therefore the float will be more important and liquidity increased (Becht, 1999).
Ownership structure and market liquidity relationship can be examined with two potential explanations. First the trading hypothesis posits that stock liquidity is increasing in the market capitalization of a firm. Average transaction costs depend on the number of shareholders participating in trade .The number of market participants willing to invest in information acquisition in a firm is larger , the larger the anticipated gains from the trade , which are directly related to the stock's free float .When the ownership structure is concentrated , only the shares included in the free float are likely to be traded and therefore the free float is a better measure of the market size for the firm's share .A large shareholder controlling a listed company takes into account private control benefits and therefore trades differently than small shareholders. When a firm's ownership is concentrated, the availability of shares is limited, there are fewer shareholders who can participate in the trading of the stock and the frequency of trades is reduced and therefore liquidity is decreased .
Second, the adverse selection hypothesis supposes that ownership concentration is potentially costly, because blockholders may have private information about the firm value. In response to a higher probability of informed trading, liquidity providers may widen spreads. While the focus of most empirical studies on the ownership-liquidity relation is on the adverse selection hypothesis (e.g. Kini and Mian, 1995; Heflin and Shaw, 2000; Sarin, Shastri and Shastri, 2000) .There is no consensus in the literature about which entities constitute the informed group of investors. Some studies focus on insiders, some focus on institutions, and some focus on blockholders. In view with the informed trading, large shareholders are more likely to be informed; their presence should increase the adverse selection discount associated with the trading of the stock, which implies a reduced liquidity. The ultimate owner may adopt a poor disclosure policy, to be able to trade on his private information. In the presence of large private benefits, outside investors may not buy the stock, reducing market liquidity.
Our study concentrates on the impact of controlling owners on market liquidity .We take into account institutional shareholders if they control the firm. We suppose that controlling shareholders are informed and therefore the spread is larger and depth is lower reflecting the high probability of informed trading.
Second, we examine the effect of deviation of ownership from control on liquidity. Third, we investigate how the method shareholders use to separate ownership and control affects liquidity. Pyramids have typically a negative impact on liquidity. On the other hand, nonvoting shares generate a transparent divergence between cash-flows rights and voting. Nonvoting shares spread is smaller because insiders hold shares with high voting rights.
Our hypotheses are thus sum up as follow :
H1. Market liquidity is positively related to the float: The liquidity of a closely-held firm stock is lower because the float is smaller.
H2. Concentrated ownership is negatively related to market liquidity: The liquidity of a closely-held firm is lower because of a higher probability of informed trading.
H3. In the presence of family shareholding, the impact of the separation of control from ownership depends on the device used by the firm. Pyramidal structure should reduce liquidity whereas shares with no voting right, which deter informed trade, increase liquidity.
II. Corporate governance characteristics and TSE market structure
II.1. Corporate governance characteristics in Tunisia
Compared with other Anglo-Saxon economies, ownership structure is highly concentrated in Tunisia. On average the five largest shareholders holds more than 80 % of the capital .Several studies show that only 8% of the Tunisian firms are widely held ,whereas in European firms this percentage increases to 36.93 % ( Faccio and Lang, 2002). Family control and pyramidal holdings have long predominated the Tunisian corporate sector. Further, this concentration of ownership in Tunisia differs appreciably from that prevailing in the US. Heflin and Shaw (2000) report that the average total blockholding is 12.3 % of shares outstanding. In, Dennis and Weston (2001) insiders ownership is on average 9.97 % of the US firm's stock while institutions own 31.06 % of the firm's stock.
According to the Tunisian Commercial Companies Code, there are two types of general meetings. Decisions by ordinary meetings, which approve the accounts, decide bond issue, appoint and dismiss directors, require a majority of 50 % of voting right to be adopted. Decisions by extraordinary meetings pertain to all decisions amending the charter and issuing shares and require a two - thirds majority of voting rights. Following many prior studies as Claessens and al (2000) and Faccio and Lang (2002),we use 20% of the voting rights cut-off to classify firms into two groups widely-held versus closely held .Accordingly we consider firms with more than 20% of voting rights as closely-held .Furthermore, we take into account pyramidal structure to determine ultimate control right. We measure ultimate control and ownership in terms of voting rights and cash flow using the weakest link along the control chain .For example, if a firm X owns 30% of the firms A which owns 20% of the firm B ,then this firm X owns 6% of the cash flow rights of firm B and controls 20% of the firm B.
Ultimate ownership can deviate from ultimate control through the use of nonvoting shares which are another common device to enhance control in Tunisia. The firm's charter can authorize priority dividend share, which provide a higher priority in dividend distribution to the shareholder, and limit his voting rights in general shareholders' meetings. Preferred shares are also shares without voting rights and provide only a fixed dividend. Voting certificate provide other share's rights. Shares with no voting rights structure has not been taken into account in previous studies examining the case of Tunisia.
II.2. TSE market structure
The Tunisian Stock Exchange is a pure order driven market without market makers. Orders are submitted by brokers on the behalf of investors and executed through an automated trading system, using a computerized limit-order book, known as SUPERCA.
Trading is carried out from 9:00 to 11:00 from Monday to Friday. It starts by a pre-opening session (from 9:00 a.m. to 10:00 a.m.) during which investors can place, modify or cancel orders but no trades are permitted. A theoretical opening price is displayed in real time to show the market tendency. The TSE operates as continuous market for the more actively traded stocks and a call auction "fixing" for the less liquid stocks. A call auction determines the opening and closing price in the continuous market. Orders are executed using time priority at a given price and price priority across prices. Investors can choose between market and limit orders, so as liquidity is only provided by limit order traders. A limit buy (sell) order specify the maximum (minimum) price at which the investor will accept to exchange. However, a market order is executed immediately against the best quote on the opposite side, but the quantity that cannot be executed at that price, will remain in the order book as a limit order at the transaction price.
TSE offers substantial transparency. The five best bid / ask limits (price and quantity) in the order book are publicly released although the identity of the broker is not shown. Orders and trades' information are updated and disseminated in a real time .However; the electronic limit order book is fully visible to brokers and regulatory authorities.
III. Data
This study uses intraday data for 40 listed firms on the TSE. Our dataset contains a time-stamped record to the nearest second of all transactions and orders submitted to the market from January 2001 to December 2005. The transaction data include the date and time of transaction, transaction prices and the number of shares traded. The order data display the date and time of order entry and execution, the price, the quantity and the best limits of the order book (bid and ask prices, and bid and ask size). These data include as well as market capitalization.
There is no electronic database on Tunisian firm ownership, including information on o control rights and corporate governance characteristics. These information were collected manually from the annual reports of listed firms for three years 2001, 2003, 2005.The ownership stakes are those available on the 31December of each year. Our sample includes all Tunisian firms, for which the annual reports provide information on the ownership structure of all major shareholders. Our definition of ownership from control relies on cash flow and voting rights. Table 1 reports characteristics for 220 firm-observations of our database.
[Insert Table 1 here]
These data reports that Tunisian firms are mainly closely-held .On average; the largest shareholder holds directly 49.56% of the capital and 51.48% of the voting rights. The second largest shareholder owns on average 15.25 % of the capital and 16.42% of the voting rights. The float, which corresponds to the fraction of listed share capital that is freely traded on the market, amounts on average to 31.65 % for the total sample.
We further identified the ultimate controlling shareholders in the presence of pyramidal structure , for a threshold of 20 % of the voting rights ,according to Faccio and Lang (2002).On average , the largest shareholder has 39.56% of the capital and 41.34 % of the voting rights .We also take into account the identity of the controlling shareholders . We find that 39% of the companies are controlled by a family, which own on average 53.44 % of the capital and 55.57 % of the voting right. The Tunisian listed firms are therefore largely controlled by families. Apart from family control, 28% of controlling shareholders are financial institutions and 21% of listed companies are controlled by the Tunisian state. However, for the total sample, only 7% of the firms are widely-held.
We then investigate the devices used by the Tunisian firms to separate ownership and control (Table 2). Pyramidal structures are the most frequent method used by 15.55% of the firms. Cross-holdings are used by 11.11% of the firms, Priority dividend shares by 4.5 % of the firms, and preferred shares and voting certificate by 2.5%. However 63.48 % of firms don't adopt any devices and they are in the "one share -one vote" setting.
Controlling shareholders identity seems to affect how firms ensure separation between ownership and control rights .Families frequently prefer pyramiding, the other controlling shareholders (financial institution, State and another widely held firm) use typically cross-holding and shares with no voting rights.
[Insert Table 2 here]
IV. Liquidity measures and methodology
In this section , we present liquidity measures used for intraday variables and the adopted methodology .
IV.1. Liquidity measures
Liquidity measures used in this study are divided into trade-based measures and order-based measures .Trade-based measures include the turnover and the number of trades. Order-based measures include effective relative spread and depth. Depth is measured by the number of shares to purchase or to sell , respectively at the quoted bid and ask prices. The daily average relative effective spread, is calculated using full day data records. The effective relative bid-ask spread takes into account the possibility of a transaction within the spread .The effective relative spread is related to the midpoint of the spread :
Effective Relative Spread = , Where
Midpoint =
The variables considered are measured at the end of each 15-min interval from the limit order book. We average all liquidity variables over 12 months around the date of ownership structure for each year.
IV.2. Methodology
We adopt in our analysis panel data in order to explore ownership-liquidity relationship, using ordinary least square (OLS). We consider the empirical model describes as follow:
+ (1)
Liquidity (LIQ) is the dependent variable represented by either relative effective spread or depth. As independent variables ,we use a set of corporate governance variables and a set of control variables . Among the corporate governance variables, we include:
AMAJ is the percentage of equity owned by the main controlling shareholder.
ASEC is the percentage of capital owned by the second largest shareholder.
SEPAR measures the discrepancy between cash flow and control rights for the main shareholders .It is defined as the ratio of voting rights minus cash flow rights to cash flow rights.
CFAM is a dummy variable that takes the value 1 if the controlling shareholder is a family at 20% threshold, 0 otherwise.
CNFAM is a dummy variable, equal to 1 when the controlling shareholder at 20% threshold is not a family, 0 otherwise.
PYRM ,CROIS ,SNVR are dummies capting the presence of pyramids , cross-holding and shares with no voting rights ,equal to 1 if the provision is present ,0 otherwise .
In order to take into consideration the combined effect of the control variables suggested by previous studies of the determinants of the bid-ask spread, we include:
CAPB is a control variable that measures the size of the firm. It is defined as the log of the market capitalization of the firm. It is calculated as the logarithm of the daily market value of the firm, averaged over 12 months around the year end.
VOLT is a control variable measuring volatility. It is defined as the average of the daily standard deviation of rate of return over one year round the date of ownership structure.
YEAR is a dummy variable that takes the value 1 if the liquidity variables (spread, depth) refer to the year 2003 or 2005, 0 otherwise.
V. Empirical results
The effect of controlling shareholders on liquidity
In this section, we examine the effect of controlling shareholders on stock liquidity. We expect liquidity measures to be higher for widely-held firms than for firms with controlling shareholders. Table 3 reports liquidity measures and market characteristics by subsamples of family-controlled firms ,non-family-controlled firms and widely-held firms. The last columns provide parametric and non parametric tests to compare family-controlled firms , non-family-controlled firms and widely-held firms.
[Insert Table 3 here]
Our results confirm that widely-held firms are the largest firms in the market value, are involved in the most transactions ,are more volatile and are more liquid than family-controlled firms. The average market capitalization of widely-held firms is 2622 MD, compared to 1580 MD (1842 MD) for family-firms (non-family-firms) .Their turnover and their number of transactions are significantly larger than the other controlled firms ones, which are larger than the turnover and the number of transaction of family-controlled firms .The average relative spread is 5.83% for family firms, 5.35 % for non-family controlled firms and 5.22% for widely-held firms. The average depth is 125 shares for family-firms, 163.67 shares for non-family controlled firms and 174.25 shares for widely-held firms. The breakdown of the sample indicates that the volatility and the size of listed companies on the same market segment show only limited differences according to the existence of a controlling shareholder. Indeed, family-controlled firms are still characterized by wider spreads and lower depth. Generally, the descriptive statistics show that family firms are the smaller and less liquid companies. Moreover widely-held firms are the largest and the most liquid.
We use further several regression models to complement the univariate analysis using ordinary least square (OLS) .We use either the relative effective spread or the depth as the dependent variable to explain liquidity . Table 4 presents the estimation results . Regardless of measures, liquidity increases with market capitalization and decreases with volatility. This result confirm the previous findings on the determinants of the bid-ask spread .
[Insert Table 4 here]
We regress in the first column, the spread on the free float, and we find a negative relationship (-0.0009) .This result is consistent with the trading hypothesis. Information production and trades frequency decrease when the market size of the firm's shares is reduced and therefore liquidity is decreased. we include in the second column two dummies for controlled firms (CFAM ,CNFAM ) to test more directly the role of controlling shareholders on liquidity .Spreads of family-controlled firms (other controlled firms) shares are 13.05% (14%) larger on average than widely-held shares spreads .We then include the percentage of capital owned by the first and the second shareholders .We find that the percentage of the capital held by both the controlling and the second shareholder ,increases the spread (0.0010) and depth decreases (-2.6017) confirming the free-float hypothesis and also the adverse selection hypothesis if we consider controlling shareholders as informed traders .
We also examine the impact of separation of the ultimate ownership and control on liquidity. We find that the discrepancy between cash flow rights and voting rights have a positive impact on spread (0.0200) but the impact on depth is negative and insignificant. This result can be explained for several reasons . First, when minority shareholders suspect private benefits to become a large part of the firm's value, they may renounce to buy the shares and prefer selling them. Second, the controlling shareholder may choose a poor disclosure policy, to prevent outside investors to trade on his private information.
Our results remain robust when we consider subsamples by control type ( family control , non-family control ) .The presence of a controlling shareholder at the 20% threshold increases the bid-ask spread , for all controlled firms . The separation between ownership and control reduces liquidity ,whether he is a family firm or not. One explanation is that in the existence of a large deviation between ownership and control , most institutional investors do not invest .Therefore , transactions are mainly initiated by individual investors with smaller order size. Indeed, as a fraction of these investors are informed, liquidity decreases. We therefore find evidence for the adverse selection hypothesis which suppose that liquidity is reduced in the presence of controlling shareholders.
B. Market liquidity and enhancing control devices
The main devices used by Tunisian firms are pyramids, cross-holding and non-voting shares. We examine how these mechanisms affect liquidity. Table 5 (panel A) presents the characteristics of firm-year observations using pyramidal structures by control type (family control / non-family control).We find that pyramids have not a significant impact on non-family controlled firms characteristics .On the contrary, for family-firms, the free-float is significantly higher (35.41% on average, compared to 22.55% for non-family controlled firms) .They are more liquid than other controlled firms. The average relative spread is 1.78% for family firms with pyramids (1.96% for other firms without pyramids). The median depth is 140 shares (125 shares for other controlled firms). This finding confirms that families want to keep the control of their firms anyway, and the use of pyramids allow them to enhance control for a small cash flow stake.
[Insert Table 5 here]
Table 5, Panel B compares controlled firms using cross-holdings to controlled firms without cross-holdings. Cross-holdings do not affect significantly the characteristics of non-family controlled firms. However , family-firms controlled by cross-holdings are larger , have a higher turnover and a larger free-float .They are more liquid : the average spread is 1.64% for cross-holding compared to 1.95% for other firms and the median depth is 149.5 shares compared to 97 shares for firms without cross-holding.
We then run regressions, to examine the impact of the different enhancing control means on market liquidity. Table 6 displays the regression results for the total sample.
[Insert Table 6 here]
In the first set of regressions, we examine the impact of direct ownership variables:
( AMAJ,ASEC,SEPAR) respectively the percentage of capital owned by the main and the second shareholders, deviation between cash flow rights and control rights .The effect of the two direct ownership variables( AMAJ,ASEC) is the same as those of the ultimate ownership variables in Table 4. On the contrary, the direct ownership/control deviation has a significant negative impact on spread; however, in Table 4, the ultimate exhibit a positive impact. Nonvoting shares are the main origin of direct deviation. Ultimate deviation is the consequence of both pyramidal structure and cross-holdings. This finding confirms that the different means of enhancing control have different impact on liquidity .
To examine this hypothesis, we include in the next regressions dummies (PYRM,CROIS,SNVR) equal to one if the firm has used one of the following control enhancing means : pyramidal structure, nonvoting shares and cross-holding .The regression results report that pyramids have a positive impact on spreads ,however nonvoting shares lead to increase spreads . These results confirm that family-firms preferably use nonvoting shares , and other controlling shareholders use frequently pyramids and cross-holdings .To verify these findings ,we run regressions for subsamples by control type ( table 7) ,to check whether this result is invariant whatever the identity of the controlling shareholders.
[Insert Table 7 here]
For family-firms, we find that direct deviation between ownership and control is negatively related to spreads. However, pyramidal structures increase spreads and nonvoting shares decrease spreads. Whereas, for non-family controlled firms, neither the impact of direct ownership/control deviation nor the nonvoting rights dummy impact is significant. Indeed, spreads increase with pyramidal structure .These results confirm that non-voting shares may have a positive effect for family-controlled firms. In fact, the free-float is greater for family-firms with non-voting shares . Furthermore, non-voting shares allow family shareholders to control their firms by holding a lower stake of shares; as non-voting shares provide a higher priority in dividend distribution to the shareholder, and limit his voting rights.
C. Adverse selection spread component and ownership structure
We further examine the effect of ownership structure on the adverse-selection component of spread . Particularly, we check if there is a variation in the adverse selection spread component in the presence of a large shareholder or when the deviation between ownership and control is higher .
To measure the adverse selection spread component ,we use the Huang and Stoll (1997) model.
(2)
We denote:
P : the price of the transaction at time t.
Q : the buy-sell trade indicator variable for the transaction price, P. It equals +1 if the transaction is buyer initiated and - 1 if the transaction is seller initiated .As the TSE is an electronic order driven market , the direction of the transactions can be identified in a precise way , by comparing the price and the immediately preceding spread .
α and β: cannot be estimated separately, and represent the adverse selection component and the cost of inventory component of the spread.
S: the traded spread is estimated with the Huang and Stoll's model.
The adverse selection component of the spread is estimated as follow:
Adverse selection component =
We average the adverse selection component of spread for each share and each of our three periods (2001, 2003, 2005). These variables are estimated using GMM method. Table 8 compares the characteristics of firm-years for which the adverse spread component are calculated and the other observations.
[Insert Table 8 here]
We find that the firms for which the adverse selection component of the spread is calculated are characterized by a larger average market capitalization (1262 MD compared to 76 MD for the other firms). The free float is significantly greater (32.85% compared to 17.75%), the main shareholder holds 40.76% of the capital (44.07% for other firms), and the average spread is 1.08% compared to 1.20% for the other firms. In general, firms for which the adverse component of the spread is calculated are larger and more liquid than the other firms.
We next run regressions to examine the relation between the adverse selection component of spread and ownership structure .More precisely we test the hypothesis that information asymmetry increases with both large shareholders and ownership/control deviation.
[Insert Table 9 here]
Our results confirm that the direct and ultimate percentage of capital held by the main and the second shareholders are associated with a higher adverse selection component of the spread. This finding is consistent with the fact that controlling shareholders are typically informed traders .Furthermore the deviation between ownership and control increases the information asymmetry. In particular ,by adopting a poor disclosure policy ,the controlling shareholder extract private benefits from his private information and therefore information asymmetry arises .However the cross-holding and nonvoting shares dummies have an insignificant impact on the adverse selection component of the spread .
VI. Conclusion
This paper provides an empirical examination of the ownership-liquidity relation in the Tunisian stock exchange. The major objective of this study is to investigate the impact of ownership concentration and the separation of ownership from control on market liquidity. To reach this objective, we have used a sample of 40 firms, selecting the actively traded stocks on the TSE over the period 2001-2005. Liquidity is represented by either effective relative spread or depth, measured at the end of each 15-min interval from the limit order book. Our definition of ownership from control relies on cash flow and voting rights.
Our results indicate that Tunisian companies are characterized by concentrated ownership. More precisely, we find that liquidity is significantly reduced for closely-held firms. This result is robust when we consider the direct/ultimate ownership of the controlling shareholder and taking into account the controlling shareholder identity. For all controlled firms ,the separation of ownership from ultimate control increases the spread and its adverse selection , but the effect is more pronounced for family-firms.
The regression results indicate that pyramidal structures and cross-holding have a positive impact on spreads for all controlled firms. Whereas, nonvoting shares permit a greater float for family-firms that want to keep control and deter informed shareholders to trade on private information. Overall, our study confirms that nonvoting shares enhance small family-firms control ,deter information asymmetry and thus may improve stock liquidity for minority shareholders .
Finally, these results suggest that firms may mitigate information-based trading and enhance stock market liquidity by adopting corporate governance standards that prevent insider trading. The results raise questions for future research. First, recent availability of intraday data from the TSE allows for a multitude of new research questions and can deepen the understanding of the order book as well as enhance the preciseness of practical applications. Furthermore ,it would be interesting to extent this analysis and explore other emerging stock markets .Second, the non availability of electronic database on Tunisian firm ownership, including relevant information on voting rights and corporate governance characteristics explains the rarity of studies on governance attributes.
