The impact of blockholders on firm value during the economic crisis : An event study on the effects of CAR and a causality test among firms in the Netherlands.

(1)

U

NIVERSITY OF

A

MSTERDAM

F

ACULTY OF

E

CONOMICS AND

B

USINESS

The impact of blockholders on firm value during the

economic crisis

An event study on the effects of CAR and a causality test among firms in the Netherlands

M

ASTER

T

HESIS

MS

C.

B

USINESS STUDIES –

G

OVERNANCE AND

V

ALUATION

January 2014 - Version 9.9

Paul van Kekem (5940044)

(2)

“Not dispersed ownership is still the rule rather than the exception throughout the world, and so is family control of even the largest corporations or business groups in most countries”

Thomsen and Pedersen (1996)

Acknowledgements

I would like to express my deepest appreciation to all those who provided me the possibility and trust to complete this thesis. In particular I want to thank my parents, R.G.G. van Kekem en J.J.P. van Kekem – Landa, my study friends Patrick van de Belt and Jurian van der Have, and Michiel Beukers, co-founder ImpacD business consultants, who all kept believing in me and supported me in every way that was

suitable. Best gratitude to Mister Z. Sautner and J.E. Ligterink for their very interesting lectures, ‘Corporate Governace’ and ‘Valuation and Value based management’, which taught me a lot and made me love my study. I would have loved to have had even more finance related courses, but I was fortunate to read many financial papers on my own during my thesis and study times to expand my knowledge further on the base of the one created by the University of Amsterdam. Special appreciation again for Mister J. Ligterink in the support during the process to get my master thesis finished.

During the writing of my thesis, I moved to Berlin and since I could use the databases at the Humboldt University, I want to give my word of thanks via the following recognition:

This research was supported by the German Research Foundation through the Collaborative Research Center 649 "Economic Risk", Humboldt-Universität zu Berlin.

(3)

Prologue

Capitalism has proven to be the best system so far. But with the recent financial crisis, it became clear that also this system has its disadvantages or even limits. It is the task of the scientific world to get more knowledge of the leap holes in the system and to find solutions or test best practices to cope with these dangers. Subsequently it is the task of society to develop their market, their laws, their complete governance system to have capitalism as beneficial as possible and get the disadvantages as good as possible banned. The regulatory power can be considered as one of the most important influencers in monitoring and controlling the dangers, but besides these officials the blockholders can create an appreciated addition.

However, blockholders are not always warmly welcomed. When the entry of a blockholder is announced, the first reaction is commonly a bit reluctant. For instance when the Mexican billionaire, Carlos Slim – estimated as the richest man in the world from 2010 to 2013 (in the Forbes annual list of billionaires), announced that he was interested in KPN, the first reaction of the KPN board was how this could be prevented. But was this anxiety well-founded? Should shareholders normally be scared when a blockholder announces to enter a company? And so, should shareholders be happy when a blockholder exit a company? These questions were one of the bases for this research.

The scientific bases and the results can be read in this thesis, however it is a pity that the thesis itself does not do justice to the amount of effort I put into this study. I emphasize this as the process had its ups and its many downs. It all started from the very beginning where I had to find the perfect thesis topic. Apparently there was not one for me and after four starts, which all went to a ‘cul de sac’ (dead end), I found good ground for this topic. The other four different topics which I ‘ploughed’ were: (i) Do acquisitions by Chinese (or other BRIC-country) firms give higher abnormal returns then regular acquisitions? (ii) Are there during the crisis industries with listed firms that are less affected by the recent crisis? (iii) Is there a relation between country cultures indices and the impact of the recent financial crisis on their stock indices? And (iv) Does stronger corporate governance decrease the volatility of the stock indices?

Secondly, I also had many database problems, what was quite frustrated, but had its advantage that I experienced 3 different databases. I started with Thomson Reuters Datastream. As data was missing and I wasn’t close to the University of Amsterdam I could register to Compustat online (part of the Wharton Research Data Services and for subscribers only) and so my access was not limited by location. This database went offline for master students and so I went to the Humboldt University in Berlin. Here I used Thomson Reuters Datastream and the Bloomberg Professional Database (both for subscribers only).

Thirdly, it is a pity that it is not visible on what kind of excel sheets this thesis is based on, because I have spent a serious amount of time with MS Excel to get the data ready for its statistical tests. Many excel sheets were made, long formulas were defined, VBA-codes were written and so on. The datasheets were one of my highlights during the writing of this thesis. Luckily I enjoy working with Excel.

However I had never problems with learning, understanding, analyzing and so on, thesis writing gave me a big headache. Although I am not proud on the process as a whole and although the famous ‘thesis

mountain’ grew with every setback even to a personal Mount Everest, I am proud that I didn’t chose for the easy way, so didn’t chose to quit. I am happy that I proved persistence and even while I went many steps back times and times again during the process, the sum of all steps forwarded luckily to the finish and results in the following thesis.

(4)

List of tables

2 Review of literature

Table 1. Takeover defences ... 13

Table 2. Ownership structure by types of blockholder ... 13

Table 3. Ownership concentration of voting rights ... 13

Table 4. Ownership concentration and the correlation between the largest blockholders ... 14

Table 5. The relation between blockholders (BH) or block trades (BT) and firm value (FV) in the Netherlands (NL) ... 14

3 Research model Table 6. Timeline for the Event Study ... 16

Table 7. List of empirical variables for the Event Study ... 18

Table 8. List of empirical variables for the Granger causality test ... 19

4 Sample Table 9. Summary statistics event study ... 22

Table 10. Summary statistics Granger causality study ... 23

5 Results Event study Table 11. CAR descriptive for different event windows for sell and buy events ... 25

Table 12. CAR descriptive for different event windows for sell events with and without withdraws ... 25

Table 13 Ordinary least squares (OLS) regression results on cumulative abnormal returns ... 27

Table 14. Quality of the OLS regression variables ... 27

Table 15. Weighted least squares (WLS) regression results on cumulative abnormal returns ... 29

Table 16. Quality of the WLS regression variables ... 29

Granger causality test Table 17. Ordinary least squares (OLS) regression results of Granger causality test ... 31

Table 18. Quality of the OLS regression variables ... 31

Table 19. Granger causality test - Standard Error corrected for heteroskedasticity ... 32

9 Appendices Table 20. Previous studies to the affect of blockholders including the different test methodologies ... 38

Table 21a. Main Corporate Governance Systems and general characteristics ... 39

Table 21b. Dutch Corporate Governance Structures ... 40

Table 22. Event study - Pearson Correlation for sell events ... 42

Table 23. Event study - Pearson Correlation for buy events ... 42

Table 24a. Granger causality test for firm value (Qt) - Pearson Correlation for sell events ... 43

Table 24b. Granger causality test for block ownership (OSt) - Pearson Correlation for sell events ... 43

Table 25a. Granger causality test for firm value (Qt) - Pearson Correlation for buy events ... 44

Table 25b. Granger causality test for block ownership (OSt) - Pearson Correlation for buy events ... 44

Table 26a. Granger causality test for firm value (Qt) - Pearson Correlation for all events ... 45

(6)

The impact of blockholders on firm value during the economic crisis

An event study on the effects of CAR and a causality test among firms in the Netherlands

Abstract

The Dutch market features highly concentrated ownership in firms. The most important agency problem that accompanies with blockholder ownership is tunnelling. The danger that large shareholders extract value out of a company for private wealth is opposite to the main benefit that they will monitor the firm its controlling management for common interests. The relation between a change in blockholder ownership and firm value during the recent financial crisis is researched by the use of an event study and a Granger causality test.

The results of this thesis are that a decrease in blockholder ownership aggravates a negative price reaction and in case a blockholder withdraws completely a little positive effect occurs. Subsequently, in case of blockholder buy events, large trade sizes yield negative price effects. This indicates that information or monitoring effects dominate liquidity explanations. Although, the Granger causality test showed that blockholders are not the cause of a change in firm value during the crisis, but react on the stress situation.

1 Introduction

A blockholder is a large shareholder that owns at least five percent of the firm. Especially for large

(publicly) traded firms, this is a significant value and so a blockholder must have significant powers. Much research is already dedicated to the question what the impact is of blockholder ownership or ownership concentration on firm value (e.g. Short, 1994). However, ‘the question remains whether the presence of large owners does, in fact, improve company performance’ (Holderness et al., 2003).

Berle and Means (1932) indicated the agency problem between ownership and control. And especially when the governance legislation is not in favour of shareholders, blockholders can be added value by monitoring how the firm is managed. Due to a change in ownership structure this conflict of interests between managers and shareholders will change to a new agency problem between dispersed and one or more concentrated shareholders, the blockholders (e.g., La Porta et al., 1999; Barca and Becht, 2001). Because of the conflicts of interests, the market reacts on an entrance or exit of a blockholder into a firm or a significant change in ownership of the large shareholder (e.g. Maury and Pajuste, 2005).

One of the main constrains of ownership concentration is that it reduces the market liquidity (e.g. Coffee, 1991; Bolton and Von Thadden, 1998). And since only liquid markets can absorb large volumes without substantial price changes, blockholders can harm the distribution of information. Bhide (1993) emphasizes that US policies, which promote market liquidity, have two benefits: (i) investors can sell their assets quickly and (ii) can diversify cheaply. However, these policies encourage diffuse shareholding and discourage active investing and ‘this implies a trade off between liquidity and governance’ (De Jong and Van Beusichem, 2006).

Thomsen, Pedersen and Kvist (2006) found a negative association between large shareholders and firm value or accounting returns in the next period and proved via a Granger causality test the predicting value of blockholders in continental Europe. De Jong and Van Beusichem (2006) made their sample more specific and focussed only on the Dutch market. Their event study results showed that for a decrease in blockholder ownership the market showed significant negative effects while for an increase in blockholder ownership insignificant positive returns were measured.

Although there is still much to explore on the topic of blockholders, the impact of a financial crisis, and especially how to react on the impact, is even less covered in literature. The recent financial started in 2007 and the total impact is still incalculable. In times of stress, the market is in a new situation that does not confirm the normal movements. This thesis will explore how the Dutch market reacts on changes in

(7)

block ownership in times of crisis and combines the study of Thomsen et al. (2006) and De Jong and Van Beusichem (2006).

This thesis has six sections. On the base of different headlines in existing literature, the following chapter explains the advantages and disadvantages of blockholders and discusses the relation between block ownership or block changes to firm value or firm performance. This second chapter finishes with a description of the Dutch market in relation to its corporate governance system. The third section describes the research model including all hypotheses, definitions and methodology. The fourth clarifies the sample and the fifth elucidate the results. The final sections finish this thesis with a discussion that also contain the limitations of this research and it is followed by the conclusion.

2 Review of literature

Blockholders facilitate in a different set of advantages, namely: (i) provide in internal and external

financing (e.g. Wruck, 1989; and Winton, 1993), (ii) monitoring the controlling managers (e.g. Shleifer and Vishny, 1986); Admati, Pfleiderer and Zechner, 1994; Burkart, Gromb and Panunzi 1997; and Maug 1998), (iii) being a defence mechanism against hostile takeovers (e.g. Kabir, Cantrijn and Jeunink, 1997), and (iv) help in product markets (e.g. Khanna and Palepu, 2000). But to have beneficial blockholders, one should be aware of the possible disadvantages too. This section will cover this issue by the next topics: (i) agency problems, (ii) blockholders and agency problems, (iii) blockholders and firm value and (iv) block

fluctuations and firm value, after which (v) blockholders in the Netherlands are discussed and this section will finish with a literature overview that the base is for this research.

2.1 Agency problems

This thesis focuses on two types of agency relationships and those are also pointed out by De Jong and Van Beusichem (2006). The first agency relationship is indicated by Berle and Means (1931) and is since its publication heavily discussed in literature. It is between shareholders and management. This gives a direct motive for large shareholders to monitor the management. The blockholders their motive is reinforced as they also have a higher need for sufficient monitoring as they have more incentives to monitor with a larger share (economy of scale) and are more dependent on the value of the firm since their portfolio is less diversified. Because effective monitoring by widely dispersed shareholders is rather difficult, this is considered to be the main benefit of large shareholders (e.g. Shleifer and Vishny, 1986, 1997).

The second agency relationship is the one between large and dispersed shareholders. De Jong and Van Beusichem (2006) emphasize that large shareholders incur agency costs, such as: (i) free riding problem (Grossman and Hart, 1980), and (ii) collective action problems (Shleifer and Vishny, 1986). All these costs give blockholders the incentive to compensate the monitoring costs by gaining profits at the expense of the small shareholder (Demsetz, 1986). Zwiebel (1995) makes the same distinction and

emphasize that blockholders can give themselves these private benefits at the costs of the shared benefits of control. Therefore, De Jong and Van Beusichem (2006) appoint that ‘monitoring will only be effective if interests of both large shareholders and minority shareholders are aligned’.

2.2 Blockholders and Agency problems

The main danger of large shareholders is that they will pursue private goals that differ from profit maximization or just basic business persistence (Shleifer and Vishny, 1997; Burkart et al., 1997). As previously already is indicated, blockholder ownership results in better monitoring. But Demsetz (1986) and Holderness and Sheehan (1988) find that large individual blockholders are more likely to engage in perquisite consumption and insider trading. Especially when blockholders are in the board of directors, ‘blockholders have the opportunity to consume corporate benefits to the exclusion of smaller shareholders’ (Holderness, 2003). This tunnelling phenomenon arises when blockholders undermine the interest of the minority shareholders.

Tunnelling is often an illegal business in which a blockholder (or majority shareholders or high-level company insiders) transfers company its assets and/or profits for personal gains. The common threat for devaluation in the wealth of the minority shareholders can come in two forms. First a controlling shareholder can simply transfer resources from the firm through self-dealing transactions that can include:

(8)

(i) outright theft, (ii) fraud, (iii) asset sales, (iv) contracts such as transfer pricing advantageous to the controlling shareholder, (v) excessive executive compensation, (vi) loan guarantees, (vii) expropriation of corporate opportunities, and so on.

Second, ‘the controlling shareholder can increase his share of the firm without transferring any assets’ (Johnson et al., 2000), through: (i) dilutive share issues,(ii) minority freeze outs, (iii) insider trading, (iv) creeping acquisitions, or (v) financial transactions that discriminate against minorities.

Atanasov (2004) research how much value blockholders can tunnel after privatized companies begin trading on the Bulgarian Stock Exchange and found some obvious (maybe even extreme) results. It reveals ‘that the behaviour of large shareholders closely resembles the predictions of Fama and Jensen (1983)’, namely ’that majority-owned firms cannot persist as publicly traded corporations if the expropriating activities of controlling blockholders are not legally restricted’. His result showed that shares in controlling blocks are priced up to ten times higher than shares in small blocks and controlling shareholders can tunnel as much as 85% of firm value.

Most countries prohibit tunnelling as mentioned above and in assessing conduct, courts generally use two broad principles, which appear in all major legal systems: duty of care1 and duty of loyalty2 (Johnson et al., 2000).

Subsequently, Johnson et al. (2000) conclude that ‘civil law countries may accommodate more tunnelling than courts in common law countries’. They indicate the next reasons: ‘(i) a narrower

application of the duty of loyalty largely to transactions with no business purpose, (ii) a higher standard of proof in conflict of interest situations, (iii) a greater responsiveness to stakeholder interests, and (iv) a greater reliance on statutes rather than fairness to regulate self-dealing transactions’.

When the ownership is much dispersed or at least when minority shareholders fear their investments, a bank Block can provide a solution as it may be willing to behave as a blockholder and monitor the firm plus guide managerial decisions (Bolton and Von Thadden, 1998)3.

2.3 Blockholders and firm value empirical research

2.3.1 Blockholders in general

Maug (1998) appoints that large shareholders make markets less liquid. And because of the low liquidity there is less incentive for them to actively monitor. Although Thomsen, Pedersen and Kvist (2006) did not make any distinction between the different governance systems of continental Europe, they did found significant proof of a negative relation between blockholder ownership and firm value (or accounting returns in the next period). This is in contrast to the two most important countries that have the Anglo-Saxon governance, the US and the UK, where no significant proof was found for such a relation. Further analysis reveals that this association is significant only for companies where the initial level of blockholder ownership is (still) higher than 10 percent.

Holderness and Sheehan (1988) analyzed 114 NYSE- or AMEX-listed corporations and found that: (i) investors have no reason to fear majority shareholders as they do not systematically consume corporate wealth, (ii) having managerial powers as a blockholders is more beneficial then supervisory rights, and (iii) the identity of blockholders is important as firms with an individual majority shareholder experiences less changes in control than firms with a corporate majority shareholder. Subsequently, firms with an individual majority-shareholder underperform, measured by Tobin’s Q, compared to firms with corporate majority shareholder Holderness and Sheehan (1988).

Although Demetz and Lehn (1985) found no relation between ownership concentration and accounting rates of return, Morck, Shleifer, and Vishny (1988) did indeed find a trend for both accounting rates of return and Tobin's Q. Tobin’s Q increased as board ownership increased from 0 to 5 percent, it decreased if ownership increased further from 5 to 25 percent and it increased modestly with a management ownership beyond 25 percent. A similar pattern was visible with accounting rates of return, except that these did not rise as board ownership exceeded 25 percent.

1_{Duty of care is a formalization of the social contract that implicit declares individuals to held responsibilities towards others within society to} avoid tort (irregularities). It is not a requirement that a duty of care is defined by law, though it is often developed through the

jurisprudence of common law. 2

Duty of loyalty is an (business) agreement between corporations and fiduciaries. When there is a conflict of interest, the fiduciaries have to put the corporation’s interests ahead of their own.

3_{‘Because the bank itself is not a major shareholder it may not take the ex post efficient continuation decision’ and it is questionable is the interest} of debt holders is (always) in line with the one of the shareholders (Bolton and Von Thadden, 1998).

(9)

On the base of a sample consisting 136 non-financial Finnish listed companies (during 1993– 2000), Maury and Pajuste (2005) showed that having at least one large shareholder with more than or equal to 10% of the voting power is not beneficial. Their data showed that the firm value, measured by Tobin’s Q, would increase when the voting power was distributed more equally.

2.3.2 Multiple blockholders

Firm value depends on the voting rights and power of shareholders (e.g. Malatesta and Walking, 1988). In traditional theories, it is argued that one large blockholder will have the best effects on firm value, because of its incentive to monitor the firm and to undertake value-enhancing interventions; ‘engaging in voice’ (Edmans and Manso, 2011). Maury and Pajuste (2005) found that the firm value is negatively related to a second shareholder, in particularly when these two blockholders combined have the majority of the ownership or when the two blocks are comparable in size. This collusion incentive can be undone by an extra blockholders. Because of its monitor incentives, a third large shareholder will have a positive effect on the firm value again. Subsequently, Maury and Pajuste (2005) continue that the firm value also is influenced by the interaction between the large shareholders.

Although this, because ‘multiple blockholders cannot co-ordinate to limit their orders and maximize combined trading profits’ (Edmans and Manso, 2011), they trade competitively and impound more information into the prices. By this, the multiple blockholders will strengthen the threat of

disciplinary exit that will increase managerial effort and increase the firm value opportunities. Edmans and Manso (2011) emphasize that ‘the optimal blockholder structure depends on: (i) the relative effectiveness of manager and blockholder effort, (ii) the complementarities in their outputs, (iii) information asymmetry, (iv) liquidity, (v) monitoring costs, and (vi) the manager’s contract’.

Faccio et al. (2001) test the effect of multiple large shareholders on dividends and find that the presence of multiple large shareholders ‘dampens expropriation in Europe (due to monitoring), but exacerbates it in Asia (due to collusion)’. In the US most firms are held by multiple small blockholders, that generates free-rider problems that hinder voice, it also has another governance effect by disciplining the manager power through trading and engaging in exit (Edmans and Manso, 2011).

2.3.3 Family blockholders

Maury and Pajuste (2005) found proof that the relation between multiple blockholders and firm value is significantly affected by the identity of these blockholders. First they found that firms that have family members as managerial or board representatives, will decrease in firm value if the outsider its ability to monitor the insiders is low. Second, if family-controlled firms have a second blockholder that is another member of the family (or more blockholders that are all members of the family) then this is negatively related to the firm value. And third, interestingly, a higher voting stake held by another non-family owner, typically a financial institution, is positively related to firm value in family-controlled firms. These results suggest that ‘the incentives to collude with the largest shareholder or to monitor the largest shareholder are significantly affected by the type of the blockholder’ (Maury and Pajuste (2005).

The findings of Barontini and Capri (2005) lead to the conclusion that ‘family control is positive for firm value and operating performance in Continental European firms.’ They saw that this was true when the founder was still alive as well when the controlling stake was held by descendants that sit on the board as non-executive directors4. Using a sample of 275 German listed companies, Andres (2008) drew the same type of conclusion; family owned firms are more profitable and outperform companies with other types of blockholders. Andres (2008) stated: ‘However, the performance of family businesses is only better in firms in which the founding family is still active either on the executive or the supervisory board’.

While researching if there was a difference in firm performance when there was active or passive family control, Maury (2006) encountered benefits from family control that occurred in non-majority held firms. Although passive family control did not affect profitability, ‘active family control is associated with higher profitability compared to non-family firms’. This conclusion could even be extracted to different legal regimes. Subsequently, active and passive were both related to higher firm valuation, but this result was mainly found in economies with high shareholder protection. These results suggest that ‘family control

4_{‘When a descendant takes the position of CEO, family-controlled companies are not statistically distinguishable from non-family ones in terms of}

valuation and performance’ (Barontini and Capri, 2005).

(10)

lowers the agency problem between owners and managers, but gives rise to conflicts between the family and minority shareholders when shareholder protection is low and control is high’ (Maury, 2006). 2.3.4 Blockholders during the crisis

In general, the two major impacts during a financial crisis are: (i) outside capital is less available, and (ii) the expected returns of existing investments decline (Lins, et al., 2011). Subsequently, a shock, such as a sudden crisis, can also impact the costs of blockholders due to agency conflicts. And since it is proven that the blockholders tradeoffs depend on its return on investment (e.g. Shleifer and Vishny, 1997), it is likely that blockholders extract more private benefits from their investment in times of a crisis.

Lins, Volpin and Wagner (2012) researched if family controlled firms have more risk aversion during liquidity shocks that occur during a crisis. In their research they found proof that during a crisis it is beneficial to have non-family blockholders control. Their results showed that families have a risk averse reaction in times of stress and by this, family controlled firms reduced their capital expenditure and underperformed compared to other firms during the global financial crisis.5

2.4 Block fluctuations and firm value empirical research

There are two different fluctuations in block ownership: (i) block trades where a blockholder changes its ownership position by selling or buying a large number of shares and changes its substantial ownership, and (ii) the extreme version of a selling or buying event that will lead to an exit or entrance of the blockholder and so will (sincerely) change the equity ownership portfolio of the firm.

2.4.1 Block trades

Damodaran (2003) points out two reasons for a price impact when investors trade. First, markets are not completely liquid and so large trades can create an imbalance between buy and sell orders. This will have a price change as result, but will generally be temporary and will be reversed afterwards when liquidity returns to the market. Second, large trades attract the attention of other traders as a block buy usually assumes that the trader is acting because of good news (and a block sell usually assumes that the trader is acting because of bad news) about the corporation. Although investors do not always trade on the right information, the price change this will have as a result will generally not be temporary.

The size of the impact depends on the liquidity of the markets and of the stock itself. Studies of block trades of large corporations on the New York Stock exchange (NYSE), so which is very liquid, concluded that prices adjusted within a few minutes after block trades (e.g. Dann, Mayers and Rabb, 1978). Studies that looked at less liquid stocks (also at the NYSE) found that markets did not have the capacity to adjust quickly and the price imbalances took more time (e.g. Hasbrouck, 1991).

Holthausen, Lefwich and Mayers (1990) found proof that it is likely that prices bounce back after a block sell, but they also suggest that‘when prices go up after a block buy, they are more likely to stay up’ in markets that are not very liquid. Spierdijk, Nijman and Van Soest (2002) saw that the less liquid the market is, the longer it takes to revert back to a normal level.

2.4.2 Entrance and exit moments

De Jong and Van Beusichem (2006) 6 did an event study to the relation of entrances or exits of

blockholders to firm value. They did not find any positive returns on purchasing situations, 0.0% at trade and 0.5% at disclosure, but they did find a significant negative reaction of the market response at the sales moment of block trades -3.2 percent (p=0.00) at trade and -1.7 percent (p=0.05) at disclosure. Less recent event studies found the next effects of block trades on firm value: (i) 1.2 percent (p=0.00) increase for sold firms and 2.9 percent (p=0.00) increase for purchased firms (Mikkelson and Ruback, 1985), (ii) 0.4 percent for block sells by regular blockholders (Holderness and Sheehan, 1985), (iii) 7.3 percent (p=0.00) increase after the initial announcements (Holderness and Sheehan, 1988), (iv) 5.1 percent (p=0.00) abnormal returns for all announcements; respectively 2.1 percent for firms staying independent and 9.8 percent for acquired firms (Barclay and Holderness, 1991), (v) 1.94 percent (p=0.01) abnormal returns on initial

announcements, (vi) 20.4 percent is the average premium for block purchases and block purchases with

5_{Family firms with financial slack underperform most, while constrained family firms do not underperform.} 6_{These results are not yet published (and were requested not to use for quotes in official papers).}

(11)

premiums show a significant positive abnormal returns of 2.7 percent (p = 0.02). Subsequently, block trades show a significant positive relation between trade size and block premium (p=0.00) (Barclay and Holderness, 1989), and (vii) a significant positive relationship between percentage holding(s) of the block purchaser(s), and a positive effect of non-financial traders and of financial traders on abnormal returns. (Shome and Singh, 1995)7. De Jong and Van Beusichem (2006) emphasize, in line with Demsetz (1986) and Holderness and Sheehan (1988), that this (joint) result show that ‘large individual blockholders are more likely to engage in perquisite consumption and insider trading, at the expense of minority

shareholders'.

2.5 Blockholders in the Netherlands

This paragraph gives a broad overview of the findings about blockholders and their ‘environment’ in The Netherlands.

2.5.1 Corporate Governance

Many researchers place the corporate governance system in The Netherlands as one that lies between the Anglo-Saxon and the German systems (La Porta et al., 1999). One could even argue that it is in the middle of three systems and add Japan to the other two corporate governance systems. An overview of the Dutch, Anglo-Saxon, German and Japanese corporate governance system can be seen in the table 21a in the appendix.

The Dutch corporate governance can be summed as: (i) it has a strong financial and corporate governance related legislation, but the protection of the individual shareholders is weak, (ii) the ownership is (moderately) concentrated, (iii) it has a high degree of takeover defences that makes hostile takeovers uncommon, and (iv) because of various defence mechanisms the role of the shareholder is very limited. Subsequently, the banks are an important group of shareholders but commonly just fulfilling the role of debtors with limited powers, especially compared to Germany. Most banks are financial conglomerates that are active as commercial banks, investment banks and insurance companies. When they are shareholders they are often represented on the supervisory boards of firms.

Van Ees et al. (2002) emphasizes that shareholders and investors can exercise control through voting at the annual meeting (‘Algemene Vergadering van Aandeelhouders’), and blockholders can take place on the supervisory board. De Jong, Mertens and Roosenboom (2004) partly contradict the effect of the annual meeting as they did an analysis of the minutes of 245 general meetings in the period 1998-2002 and revealed that about 30% of the shareholders is present at the meeting. This is low in comparison with shareholder turn-out in Anglo-Saxon countries and did draw the conclusions that minority shareholders in the Netherlands have hardly any influence on the sitting management.

This suggests that the supervisory board can be an essential monitor in the Netherlands; it is strictly separated from the management board and the composition is formally an internal matter and is not an election by the shareholders nor by the employees. Van Ees et al. (2003) describes the three primary tasks of the supervisory board in a full structured regime: (i) ‘to appoint, monitor, suspend, and dismiss members of the management board, (ii) to draft the annual financial statement for presentation at the annual

shareholders meeting, and (iii) to monitor and ratify major business decisions proposed by the management board concerning, for example, expansions, acquisitions, restructuring, or financing’.

In the Netherlands the supervisory board has a system of co-optation, which gives it the power to elect the supervisory board members itself and is automatically adapted by the structured regimes (De Jong et al., 2005). If a supervisory board is mandatory depends on which legal regime structure is required and this is covered in the next paragraph.

Besides a supervisory board, Dutch firms can implement three other types of defence mechanisms that severely restrict shareholders their power and Kabir et al. (1997) found that: (i) 40 percent of the firms issued certificates, (ii) 45 percent issued priority shares, and (iii) 59 percent issued preference shares (e.g. Kabir et al, 1997). Limited shareholder power protects the firm against hostile takeover bid and

subsequently ‘leaves much room for managers to exercise discretion in their acquisition decisions’ (De Jong, Van der Poel and Wolfswinkel, 2007). Kabir et al. (1997) found that the majority of Dutch firms

7_{De Jong and Beusichem (2006) highlighted that: ‘the positive effect of financial traders is consistent with efficient monitoring hypothesis and with} the evidence of Agrawel and Mandelker (1990), Bircley, Lease and Smith (1988, 1994), and Jarrell and Poulsen (1987)’.

(12)

protect themselves with at least two of the takeover defences and thus making blockholders less important for this purpose (table 2). Chirinko et al. (1999) sums the next findings on defence mechanisms in their sample of 94 Dutch non-financial listed firms: (i) 38 percent of the firms issued certificates of deposits through an administrative office (certificates), (ii) 22 percent issued priority shares, (iii) 65 percent issued preference shares held by a continuity foundation, and (iv) 16 percent issued finance preferences. Although the role of blockholders is considered to be minimal during hostile takeovers in the Netherlands, proof was found that firms with large block ownership use less defensive preference shares and hereby blockholders are also considered at least as a boundary for hostile takeovers (Chirinko et al., 1999).

Table 1. Takeover defences

The number of firms and its percentage per number of takeover defence adaption in The Netherlands in a sample of 177 Dutch AEX-listed firms in 1992.

No. of measures No. of Firms Percentage

0 16 9.1 1 52 29.4 2 62 35.0 3 39 22.0 4 8 4.5 Total 177 100

Source: Kabir, Cantrijn and Jeunink (1997)

In addition to external control by shareholders, Morck et al. (1988) also address the issue of internal control. The first one is the organisation of control within the firm and can also be tackled by a supervisory board or blockholder monitoring. And the second one is ownership of shares within the firm, but (in contrast to e.g. the US), this is less relevant in the Netherlands as De Jong, Kabir and Roëll (1998) found only 18 percent (25of 137) of the firms had insider ownership and only 6 of these 25 firms had majority insider ownership (five within the board of directors and one within the board of supervisors).

2.5.2 Legal Regime structures

In contrast to the Anglo-Saxon corporate governance where shareholders may control management decision making through the board and through the market for corporate control, in the Continental-European system, where the legal investor protection and the market for corporate control are missing, the shareholder control can only take place through company boards (Van Ees et al., 2003). Previous research has demonstrated that’ firm value is adversely affected by constraints placed on shareholders’ voting rights either permanently or by management‘s attempt to mitigate the market for corporate control’ (Malatesta and Walking, 1988). Even though, De Jong, Mertens and Wasley (2005) points out that the structured regime is used to directly limit shareholders their influence as shareholders lose their ability to directly monitor the supervisory and management boards. Although this, The Netherlands is well-known for its corporate governance system which encourages self-regulation and since the shareholders vote on mergers and acquisitions the market for corporate control does function (De Jong et al., 2005).

The Dutch legislation describes in their Civil Code Book (II) three regimes structures for corporations and the directives to be exempted: (i) common regime, (ii) full structured regime, (iii) mitigated structured regime. Once a company attains a certain size, it has to adapt one of the structured regimes. Before this state, the firm can chose voluntary to adapt its corporate governance to one of the structures. By this, it can experience for (expected near) future times when the firm would increase to a size that requires a structured regime or for any other reason. In the Netherlands most multinationals chose to have a supervisory board in favour of their shareholders their trust (De Jong et al., 2007).Van Ees et al. (2003) emphasize that ‘the opportunity to adopt the structural regime voluntarily can be seen as a major device to protect management from investor pressure’ and find in their sample that about 25 percent of the

(13)

firms have voluntarily adopted a structured regime. An overview of the different regimes is listed in appendix II.

2.5.3 Legislation and register notifications of substantial holdings

Since 1988 (December 12) the member states of the former European Commission agreed on the

Transparency Directive (Large Holdings Directive 88/627/EEC), that required large shareholders of listed companies to notify the relevant authorities and the corporation itself within seven days whenever their voting rights cross the threshold of 10%, 10%, 20%, 1/3 (or 25%), 50% and 2/3 (or 75%). The Netherlands has integrated this disclosure in their legislation not before 1992 (February 1) and since then it has been reformulated in 1996, 2002 and 2006.

Originally, the Dutch disclosure of major holdings in listed companies (WMZ; ‘wet melding zeggenschap’) required to notify an upwards or downwards change in block ownership that started at of 5%, and thereafter 10%, 25%, 50% and 75%. This was mandatory for all legal and natural persons and was registerd by the Securities Board of The Netherlands (STE; Stichting Toezicht Effectenverkeer). Below the 5 percent ownership there were and still are no requirements.

Several years after this agreement, (between 1st of February 1992 and 1st of June 1997) the act was revised into the ‘disclosure of major holdings in listed companies 1996’ (WMZ 1996; ‘Wet Melding Zeggenschap in ter beurze genoteerde vennotschappen’) and blockholders had to notify the authority when their ownership crossed the 5%, 10%, 25%, 50 and 66%. In 2002 the disclosure became also obligatory for the management and supervisory board.

The 2002 act was replaced in 2006 (October 1) by the ‘Disclosure of major holdings and capital interests in securities-issuing institutions’ (WMZ 2006; ‘Wet Melding Zeggenschap en kapitaal belang in effectenuitgevende instellingen’). ‘The aim of the WMZ 2006 is to increase transparency regarding major holdings and capital interests in securities-issuing institutions and to simplify the disclosure process for the parties that have a duty to disclose’ (De Jong and Van Beusichem, 2006). The thresholds of this disclosure are still today its standards and are: 5%, 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 75% and 95%. (And official description of the WMZ 2006 can be find in appendix III.)

The Securities Board of The Netherlands was succeeded by the Dutch Authority for the Financial Markets (AFM; ‘Autoriteit Financiële Markten’) and since March 2002 they register all blockholders. The AFM is responsible for supervising all operations in behalf of: investments, savings, loans and insurances and pension funds. The Dutch Monitoring Commission of Corporate Governance (‘Commissie Corporate Governance’) has the task to monitor the actuality and enhance the utility of Dutch corporate governance instruments with the goal to limit the influence of the individual shareholder in the Netherlands. It is created in 2003 by the former Commission Tabaksblat and the act is revised on 2009 (July 2).

2.5.4 The qualification and quantification of blockholders

As shareholders are not well protected in the Netherlands, the Dutch market is known for its high

ownership concentration. Especially, when the equity concentration is compared to the block ownership in the US, UK and other countries that have more in common with the Anglo-Saxon governance systems (Kabir et al., 1997; Becht and Roëll, 1999). De Jong, et al. (1998) researched the ownership structure in the Netherlands and draw the next conclusions: (i) the average ownership stakes of the largest and the three largest shareholders are 27% and 41%, respectively, (ii) the average ownership stakes of banks, insurance companies and other financial institutions are relatively low, (iii) the voting rights are more concentrated than ownership rights, (iv) the use of a supervisory board representing interests of different stakeholders is ubiquitous, and (vi) shares are largely held by foreigners (approximately 50 percent).

According to Kabir et al. (1997) have blockholders (based on average shares) more than half of all shares in Dutch companies (51%) and the average blockholder has an ownership of 31 percent in The Netherlands. Cantrijn et al. (1993) found that only a few firms were majority share owned, which makes active monitoring more rare and creates free-rider problems. Besides these findings, Kabir et al. (1997) also stated that the most important blockholders in the Netherlands are companies (20%), management and family members (8%), and individual blockholders (5%). In a more recent study of Lins, Volpin and Wagner (2011), found in a sample of 83 Dutch listed companies that: 58 percent was widely held, 17

(14)

percent had multiple blockholders, 14 percent was non-family controlled, and 11 percent was family owned8.

The Tables 2, 3 and 4 show the ownership by types of blockholders and the ownership

concentration of the Dutch market in previous research. De Jong et al. (1998) (table 3) find, with a sample of 137 AEX-listed firm in 1996, that blockholders have nearly half (48 percent) of the total ownership and that the financial institutions are the most important investors (33 percent) on the Dutch market.

Subsequently, De Jong et al. (1998) (table 4) find that the average voting right of the largest shareholder is 43 percent9 and the three largest shareholders already have a voting right that exceeds 58 percent.

Further, Kabir et al. (1997) (table 5) find that that the ownership concentration varies quite a bit over time as they find that the average blockholder (C1) is just 30.8 percent. Although the largest and three

largest individual investors are positively correlated, the statistics between the largest shareholders and the institutional investors show that there cannot be a correlation determined.

Although the Dutch stock market can be generally considered as one with a high liquidity,

Chirinko et al. (1999) argue that several institutional features also enforce managers to act in the interest of the firm. Chirinko et al. (1999) researched the performance of 94 Dutch non-financial firms in the period 1992-1996 and found a positive nonlinear relationship between firm performance and ownership by banks.

Table 2. Ownership structure by types of blockholder

The percentage and frequency of different Blockholders in The Netherlands in a sample of 137 Dutch AEX-listed firms in 1996.

Financial institutional investors Individual investors State Insurance Other Financial Pension Venture Individual Individual Dutch Banks companies institutions Funds Capitalists Firms Capital Governmnt

Mean (μ) 7.44 10.44 14.66 0.60 0.30 2.41 10.80 1.08 Median 4.14 0 8.52 0 0 0 0 0 Minimum 0 0 0 0 0 0 0 0 Maximum 58.28 93.17 85.61 19.06 18.09 27.01 97.05 50.00 Frequency 78 44 87 10 3 38 48 4 Total mean = 47.73 33.14 13.53 1.08

Source: De Jong, Kabir and Roëll (1998)

Table 3. Ownership concentration of voting rights

The percentage of voting concentration of the largest shareholder (C1), the three largest shareholders (C3), the five largest shareholders (C5) and all blockholders (Call) in The Netherlands in a sample of 137 Dutch AEX-listed firms in 1996.

25th 75th Standard

Mean Median percentile percentile Minimuma Maximumb Deviation

C1 42.78 43.46 11.21 60.79 0 99.99 32.10

C3 58.13 61.08 27.60 92.24 0 148.73 37.13

C5 61.05 67.51 34.90 92.59 0 148.73 36.73

Call 65.52 69.84 35.81 93.72 0 148.73 36.73

Source: De Jong, Kabir and Roëll (1998)

a_{Minimum concentration of zero (0%) represents a concentration of voting rights of 5% or less.}

b_{‘Maximum concentration of more than 100% is due to multiple disclosures of same shareholdings by different owners’} (occurred in case of at least eight firms) (De Jong Kabir and Roëll, 1998).

8

Lins, Volpin and Wagner (2011) defined this categorization by a 25 percent threshold. So this means that the family blockholder, or the non-family blockholder, or all multiple blockholders together owned and controlled at least 25 percent of the company.

9_{Sweden is a country that is also known for its shareholder concentration and cccording to Bergstrom and Rydqvist (1990), the average ownership} concentration of the largest shareholder in Sweden was in their samples 43 percent as well.

(15)

Table 4. Ownership concentration and the correlation between the largest blockholders

The different ownership concentration variables (mean, median, standard deviation and correlations) of the largest shareholder (C1), the three largest shareholders (C3), the five largest shareholders (C5), all blockholders (Call) and the institutional blockholders (Cinst.) in The Netherlands in a sample of 177 Dutch AEX-listed firms in 1992.

Standard

Mean Median Deviation C1 C3 C5 Call Cinst.

C1 30.8 25.5 1.68 1

C3 45.1 42.5 24.9 0.89 1

C5 49.2 49.8 25.8 0.81 0.97 1

Call 50.9 55.1 26.5 0.75 0.93 0.98 1

Cinst. 9.9 6.0 12.5 -0.16 -0.02 0.06 0.08 1

Source: Kabir, Cantrijn and Jeunink (1997)

2.6 Literature overview

Blockholders are studied in the recent financial crisis, but these studies are far less presented in literature. Research is only done before the recent crisis and in the recent crisis only during ownership. The variables blockholders and firm value are not tested towards each other for the Dutch market specifically. Hence, this research focuses on this gap and combines the researches done by De Jong and Van Beusichem (2006) and Thomsen, Pedersen and Kvist (2005). Table 5 below gives a literature overview of the relation between blockholders and firm value, researched before and during the crisis on the moment of entry, exit as well during the ownership. (Appendix I gives an overview of studies and the different methodologies of how the relation between blockholders and firm value is tested).

Table 5. The relation between blockholders (BH) or block trades (BT) and firm value (FV) in the Netherlands (NL)

BH/BT Before the crisis During the Crisis

Entrance No proof was found between BT purchased and FV in NLa No proof was found in current literature

Ownership Neg. relation when BH(s) has/have voting powerb Positive relation when there is/are non-family BH(s)c Exit Neg. relation during BT sales and FV in NLa No proof was found in current literature

Researches Research period Research area Source: a _{De Jong and Van Beusichem (2006)} _2000-2004 _{The Netherlands}

b

Thomsen, Pedersen and Kvist (2005) 1990-1998 US, UK and Continental Europe

c _{Lins, Volpin and Wagner (2012)} _{mid Aug’08-mid March’09} _{35 (Western) countries}

3 Research Model

This section will start with defining the hypotheses that are based on the literature overview. The second part elucidates the definitions used and the final paragraph will explain the research methodologies.

3.1 Hypotheses

Based on the literature overview drawn in the previous sections, this research has two hypotheses that contain nine tests in total. The first hypothesis tests if the market will also show a negative reaction on block sales, just like De Jong and Van Beusichem (2006) found, but then in times of crisis. The second hypothesis tests if the market will react positively on block purchases. De Jong and Van Beusichem (2006) did not find a market reaction (0.0 percent) for these trades, so the market indicated to be aware of the transaction, but this outcome was not significant. This thesis will test if the outcomes will be different in turbulent times, so during the recent global financial crisis. For all hypotheses a subscript one stands for sell events and a subscript two stands for buy events.

(16)

H1a1 A blockholder has a negative effect on the firm value when the blockholder decreases its ownership of publicly traded firms in the Netherlands during the global financial crisis. H1a2 A blockholder has a positive effect on the firm value when the blockholder increases its

ownership of publicly traded firms in the Netherlands during the global financial crisis. Just like De Jong and Van Beusichem (2006) this thesis will also test if there is a difference in blockholder sells, when there is just a decrease in ownership or a complete withdraw. Therefore the following

hypothesis is formulated.

H1b A blockholder has a positive effect on the firm value when the blockholder completely withdraws from publicly traded firms in the Netherlands during the global financial crisis. The final hypothesis will test, just like Thomson et al. (2006) did in their study, if the change in firm value is actually caused by the change in block ownership or if it could be the case that a change in firm value is predicting a change in block ownership. To test this, the following hypotheses are formulated.

H2a1 Prior firm value is causal related to a decrease in blockholder ownership of publicly traded firms in the Netherlands during the global financial crisis.

H2a2 Prior firm value is causal related to an increase in blockholder ownership of publicly traded firms in the Netherlands during the global financial crisis.

H2a3 Prior firm value is causal related to a change in blockholder ownership of publicly traded firms in the Netherlands during the global financial crisis.

H2b1 Subsequent firm value is causal related to a decrease in blockholder ownership of publicly traded firms in the Netherlands during the global financial crisis.

H2b2 Subsequent firm value is causal related to an increase in blockholder ownership of publicly traded firms in the Netherlands during the global financial crisis.

H2b3 Subsequent firm value is causal related to a change in blockholder ownership of publicly traded firms in the Netherlands during the global financial crisis.

3.2 Definitions

The definition for blockholders used by Thomsen, Pedersen and Kvist (2005) can be used for the Dutch market as well and defines a blockholder as: ‘Blockholder ownership is measured by the fraction of closely held shares, which includes: (i) shares held by owners who hold more than 5%, (ii) shares held by officers, directors and their families, and (iii) shares held in trust.’ This research does not make any distinction between the type of blockholders and can thus be: family, (institutional) investors, or affiliates.

A change in ownership, increase or decrease, is only considered worth testing if this change is substantial (De Jong and Van Beusichem, 2006). This means that: (i) the change in ownership means a new registration by the AFM (see appendix III), and (ii) the sell or buy events has at least an ownership change that respectively decreases or increases the block ownership with five percent. By this a change from 9.5 to 10.1 percent ownership is not taken into account for this research as it does cope with the first criterion, but not with the second. All definitions of all variables used in the calculations of this research are explained in the next section, methodology, directly after they are mentioned in the formula.

3.3 Methodology

In this research, two types of tests methodologies will be used: (i) an event study and (ii) a Granger causality test. MacKinlay (1997) states that by the use of ‘financial market data, an event study measures the impact of a specific event on the value of a firm’. Although direct productivity related measures may

(17)

require many months or even years of observation and given that the marketplace is rational, an event study can measure the economic impact (by the use of security prices) over a relatively short time (MacKinlay, 1997). In this research the event can be determined as the actual sales or the purchases of block ownerships. Thus, hypotheses 1 to 3 will be tested by the methodology of an event study.

Although this methodology seems to be perfect for this research, an event study has its limits. Thomsen, Pedersen and Kvist (2006) state that a Granger causality test has advantage above an event study as it is ‘able to detect an indicator of causality without the use of instrumental variables and it can cover a wider range of changes than those detected in event studies’. This can explain a high share of the variance that can produce a more precise estimation of the direction of causality. The disadvantage of the Granger causality test is it requires time series information, which can be hard to find in corporate governance studies. In this study we will circumvent this problem as we test only one event and use only one moment before this event that is the first that has all the lagged values available.

Like De Jong and Van Beusichem (2006), the WMZ register, that is accessible at the AFM website, is used to identify block trades. Since the register does not report trade sizes, neither whether the change in block ownership was caused by a buy or sell, the WMZ register data are compared to the register data of the previous disclosure. The difference is considered to be the trade size. Thus a positive change, an increase, was a buy event and a negative change, a decrease, was a sell event. If previous disclosures are not available this can mean the shareholder is completely new and so enters the firm or the previous substantial change in ownership was before March 200210. Because of this uncertainty, these substantial block changes are excluded from this sample.

3.3.1 Event Study

An event study, will be used, as described by MacKinlay (1997), to find proof for the relation between the effects of a decrease or increase of a blockholder and the value of a firm. And subsequently to find proof if the exit moments of blockholders create a firm value change that is significant different compared to its market. Thus the event study is used to test hypothesis 1 (H1a1, H1a2 and H1b).

De Jong and Van Beusichem (2006) also use in their study the standard event study methodology that is described by Brown and Warner (1985) and Mackinlay (1997). The event date is pointed as day 0 and the estimation period has a length of 100 days and ranges from day – 115 to -16. Three different event window ranges are used: (i) -2 to +2 days, (ii) -4 to +4 days, and (iii) -15 to + 15 days. Below, the timeline of this event study is shown in the figure and the necessary calculations, using the event study methodology of MacKinlay (1997) which is also used by De Jong and Van Beusichem (2006), are compiled step by step.

Table 6. Timeline for the event study

τ

L1 L2 L3

Event moment/date = τ = day 0

Estimation window = L1 = T1-T0 = tτ-16 - tτ-115, = 100 days

Event window = L2 = T2-T1 = tτ-x, tτ+x = 5, 9 or 31 days (x = 2, 4 or 9 days) (De Jong and Van Beusichem, 2006)

So, | T1 - τ-1t | = | T2- τ+1t | = 0 days

Post-event = L3 = T3-T2 = tτ+115 - tτ+16 = 100 days

Source: MacKinlay (1997)

10_{May 2002 was the date since when the AFM is registering all substantial changes in blockholder ownership (see paragraph 3.3).}

(18)

The event study starts with calculating the intercept (αi) and the slope (ßi) over the estimation period

(L1 = T1 – T0) by running by the use of an OLS regression of the stock returns and the return on the market:

1. Return on security = Ri,t = Rit - αi + βiRm,t + εi,t.

Where, Ri,t is the (price) return on security for firm (asset) ‘i’ at day ‘t’, defined as ln(Pi,t/Pi,t-1), and Rm,t is

the return on the market index, in this case the AEX.

The abnormal returns are calculated via the expected returns, which are calculated using the intercept and slope of the 100 days estimation period.

2. Expected Return = ERi = αi + ßi * ACPRmarket

3. Abnormal Return = ARi = ACPRi – (E)Ri

Where, ACPR = Adjusted Closing Price Return (for the share and for the market) ACPR = (ACPt-ACPt-1) / ACPt-1

4. Cumulative Abnormal Return = CARi = ART1 + ART1t + ART2t +… + ARTDt

CAR(τ1, τ2)= � ARiτ

τ2

τ=τ1

Where ‘D’ and (τ1, τ2) are both 5, 9 or 31 days.

To test if the event has significant Abnormal Returns, the Z-statisticAR,t is used and here for the Sum of

Standardized Abnormal Returns (SAR) need to be calculated.

5. SAR = ARit / SEi

Where, SE is the standard error which is calculated per firm over the 100 day estimation period before the event.

This SAR is implemented in the following formula, to test if the event has significant abnormal returns,:

6. Z-statisticAR,t = ∑ SARt �∑ Di− 2 Di− 4 N i=1

De Jong and Van Beusichem (2006) rewrite this test statistic to test the significant effect on CAR with more events that have a change in block ownership. For this, they define for the Cumulative Abnormal Returns the following Z-statistic:

7. Cumulative Z-statisticCAR,t =

� 1 √N�

∑ SARTT21 it

�(T2− T1+ 1) �D_Di_i− 2_{− 4�}

Where, SAR is the standardized Abnormal Returns at day ‘t’.

Di is the number of observed trading day returns for firm i over the estimation period.

(19)

3.3.2 Regression analysis

The regression formula for the cumulative abnormal returns in this research is the same as De Jong and Van Beusichem (2006) has used, except that the independent variables ‘news on Mergers or acquisitions’ and ‘listed in the US/UK’ are not taken into account. This results in the next formula:

8. CAR = αi + ß1 AT + ß2 Lev + ß3 ROA + ß4 Q + ß5 Size_traded ( + ß4 Dbh_withdraw ) + εit

The shareholder withdraw is a dummy variable and takes into account if the blockholder sell means a complete withdraw out of the firm (value 1), or only a decrease in blockholder ownership (value 0). There is no dummy variable taken into account for the buy events therefore this variable is written in between brackets.

The standard error (nonsystematic risk: σi) of the returns will be different for different firms, as a

result there should be taken into account that abnormal returns will not be distributed identically (Roon and Veld, 1998). Because of this, besides the ordinary least squared (OLS) regression, also a weighted least squared (WLS) regression is conducted just like De Jong and Van Beusichem (2006) did. The WLS regression tackles the problem of heteroskedasticity and solves this variance dispersion over the sample that influences the level of significance. Both the dependent and independent variables are weighed with the inverse of the estimate of the nonsystematic risk (σi) from the estimation period (L1, -115 to -15 days)

in the event study.

Table 7. List of empirical variables for the event study

Code Description Definition

CAR Cumulative Abnormal Return Sum of all abnormal return for the given event window.

AT Firm size Total assets

Lev Leverage Total debt / Total assets

ROA Profitability Return on assets = Net. Income / Total assets

Q Firm value as Tobin’s Q the market value of the total assets divided by its replacement value, thus: (Market Capitalization + Total Liabilities) / (Total Assets) = (Total Equity Market Value + Total Liabilities) / (Total Equity Book Value + Total Liabilities)

Size_traded Substantial change in blockholder ownership

Dbh_withdraw Dummy variable Value 1 when the blockholder completely withdraws with a sell event

Value 0 when the blockholder does not completely withdraw with a sell event

Source: MacKinlay (1997)

3.3.3 Granger causality test

A Granger causality test (Granger, 1969) is also used byThomsen, Pedersen and Kvist (2006) to find proof if there is any relation between blockholder ownership and the value of the firm. Thus the event study is used to test hypothesis 2 (H2a1, H2a2, H2b1 and H2b2).

Thomsen et al. (2006) measured the firm value (Q) by ‘dividing the sum of the equity market value and the total debt book value by the book value of the total assets’.11 Thomsen et al. (2006) emphasize that the (positive or negative) effect of blockholder ownership on firm value (Q) can be directly or indirectly. So it should be tested if Q follows or precede an increase (or decrease) in blockholder ownership. If the market appreciates blockholders in The Netherlands, the firm value will follow the change in blockholder

11

Thomsen et al. (2006) noticed that: ‘The Tobin’s Q measure of equity at replacement costs was not available. Therefore, an approximation was

used (denoted the simple Q by Loderer and Martin, 1997). Chung and Pruitt (1994) found that the correlation between the simple Q and a measure of Q that attempts to use market values throughout is as high as 0.97. To correct for a right-skewed distribution of the firm value variable, log values are used.’

(20)

ownership; so an increase in blockholders will react in an increase in Q or the other way around with a negative effect. When the market does not appreciate blockholders in The Netherlands the reaction should be contrary; so an increase in blockholders will decrease the firm value or the other way around when the blockholder ownership will decrease.

These are the broad scenarios in the situation that Q is following the change in blockholders. However, as mentioned above, a change in Q can also precede a change in blockholders. This can be the scenario when there are large inside blockholders or blockholders that are not inside but have at least access to inside information12. When they know positive or negative figures that will infect the (near) future firm value, they can change their ownership by an increase respectively decrease of their share.

This all results in the next two formulas:

9. Qit = α1 + ß1OSi(t-1) + ß2Qi(t-1) + μ1,it

10. OSit = α2 + ß3OSi(t-1) + ß4Qi(t-1) + μ 2,it

The α’s and ß’s are parameters of the models, and μ1,it plus μ2,it are uncorrelated error processes. The

dependent variables that should be taken into account are: (i) sales change, (ii) sales/asset ratio change, and (iii) Equity/assets ratio change (Thomsen et al., 2006). A change in sales reflects mergers and acquisitions and captures large changes in strategy and structure. Subsequently, ‘higher sales growth could also have a positive influence on firm value’ (Thomsen et al. 2006). A change in sales/assets ratio change indicates a change in capital intensity that could also influence the need for external finance and thereby could change the ownership structure. A change in equity/assets ratio indicates a change in capital structure and thereby influencing the ownership structure (Short, 1994). Next to this, ‘a change in capital structure can put more pressure on managers to maximize profits to meet debt payments’ (e.g. Jensen, 1986) that can affect the firm value positively. See table 8 for a list of all empirical variables used.

Table 8. List of empirical variables for the Granger causality test

Code Description Definition

Qit Firm value as Tobin’s Q the market value of the total assets divided by its replacement value, thus:

(Market Capitalization + Total Liabilities) / (Total Assets) = (Total Equity Market Value + Total Liabilities) / (Total Equity Book Value + Total Liabilities).

Qi(t-1) Lagged value of Qit (one lag).

OSit Block ownership Transformation of the fraction of closely held shares (CHS). CHS is

considered as a block ownership when it has at least 5% of the voting power.

OSi(t-1) Lagged value of OSit (one lag).

Schange Changes in sales form (current value of sales – lagged value of sales) /

time ‘t-1’ to ‘t’ lagged value of sales.

(S/AT)change Changes in sales/assets from Current value of sales / assets – lagged value of sales / assets time ‘t-1’ to ‘t’

(Eb/AT)change Changes in equity/assets from Current value of equity (book value) / assets – Lagged value of time ‘t-1’ to ‘t’ equity (book value) / assets.

Source: Thomson Reuters Datastream

12

Although this is illegal in most countries, Thomsen et al (2006) emphasize that this cannot be excluded and this study should be aware of such effects.