Do shareholder rights add value to M&As? : evidence from the European Takeover Directive

(1)

1

“Do shareholder rights add value to M&As? Evidence from the European

Takeover Directive”

University of Amsterdam Msc. FIN: Corporate Finance

Master Thesis

Stefan Klink July 2018

(2)

2 Statement of Originality

This document is written by Stefan Mateo Klink who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3 Acknowledgement

I would like to thank Dr. R. Ribas Perez, my thesis supervisor, for his guidance on writing this paper and providing the necessary feedback and comments.

(4)

4

Does Country Level Takeover Law Improve Returns of Acquiring Companies?

Abstract

The level of corporate governance is incremental for company’s corporate valuation and allocation of assets. High standards of corporate governance are related to higher corporate valuations. I research the effect of the improvement of country level corporate governance on takeover efficiency. By exploiting the harmonizing effect of the European Takeover Directive as a natural experiment. This study employs a difference-in-difference approach, corrected for sample-selection bias to capture the effect of better laws regarding protecting minority shareholders on takeover efficiency. I document a weak significant positive effect supporting the evidence of a direct link between country level legal shareholder rights and takeover efficiency.

(5)

5

1. Introduction

Following one of the European Commission’s objectives to integrate European capital markets and harmonize takeover laws amongst all member states, the EU promulgated the European Takeover Directive (ETD) in the beginning of 2004. The aim of the ETD was to harmonize country level legal shareholder rights amid their member states. While, the European Commission states that their strategy and policies have a significant effect on its member states and the ETD has led to improvements. Economic researchers state the contrary, in their opinion the implementation has led to increased vagueness around shareholder laws and made it costlier to acquire a company (Humphery-Jenner, 2012; Lysandrou & Pra, 2010). This study uses the implementation of the new regulations imposed by the European Commission as an experiment, to test whether country level shareholder rights have an influence on acquisition efficiency. The results indicate a weak link between the improvement of the country level legal shareholder rights and acquirer’s returns.

A series of recent studies by (La Porta et al., 2000; Agrawal, 2013) examine the importance of corporate governance laws and corporate valuation. They document a positive relationship between the legal shareholder rights and the corporate valuation and performance of companies. Findings by Masulis (2007) reported that companies with a weak level of corporate governance are more disposed to making value diminishing acquisitions. Further, Harford (2012) showed the sources of value destruction involved with takeovers. Harford argues that the source of value destruction is imposed by agency costs arising due to the separation of ownership and control. The costs of expropriation by the controlling shareholders and management related to corporate acquisitions include empire building (Jensen, 1986), entrenchment of the shareholder’s stake in the company (Harford et al., 2012) and self-dealing behaviour. These issues are more directly resolved by firm-level corporate governance provisions. The country-level corporate governance states the legal minimum to diminish these expenses. However, it remains unclear whether the legal minimum of country level shareholder rights directly affects acquirer’s profitability after the announcement of an acquisition.

In addition, the takeover market is one of the the major and readily observable forms of corporate investment. The acquisition of a company by firm’s managements also tend to intensify the inherent conflicts of interest induced by the separation of ownership and control, between managers and shareholders in large public corporations (Jensen & Meckling, 1976). It is well recognized that controlling shareholders and management extract private benefits of the company at costs of the minority shareholders. Hence, takeover law is important to reallocate both assets and control rights in society. In imperfect markets stronger corporate governance laws are associated with more

(7)

7 efficient resource allocation (La Porta et al., 1998). Better corporate governance laws have a positive effect on solving the agency problem between shareholders and management.

According to the European Commission the introduction of the ETD resulted in enhancement of the protection of minority shareholder. Hence, this suggests that the implementation resulted in a reduction of agency costs. Therefore, this study uses the implementation of the ETD as an instrument to test whether country level legal shareholder rights causally determine acquisition efficiency, or in reverse if a low level of country level legal shareholder rights leads to inadequate allocation of assets. So, this leads to the following hypothesis: The improvement of country level legal shareholder rights

has a positive effect on acquirer’s returns.

This study exploits the introduction of the European Takeover Directive (ETD) as a natural experiment. The EU propagated the ETD in 2004 and it was enacted from 20th_{May 2006 improving}

shareholder rights in several member states. The ETD improves country level corporate governance in countries with initial low country-level shareholder rights. The ETD has the effect that firms increase their wealth allocation efficiency after a takeover and diminish expropriation of the corporate insiders. The ETD harmonized the country-level corporate governance across EU member states, since Europe’s takeover market experienced substantial differences in corporate governance (McCahery & Renneboog, 2003; Humphery-Jenner, 2012). But, it did not entail significant changes in countries who already had the core provisions of the ETD in place. The countries where the ETD improved the legal shareholder rights are signified as treated, leaving other countries in the control group.

To test the hypothesis, I use a sample of 659 intra-European acquisitions of publicly traded firms during the period 2001-2017. Further, the introduction of the ETD had a substantial effect in not all the member states. The ETD enhanced in eleven countries their legal shareholder rights, while it had no substantial effect in seven countries. By performing a difference-in-difference approach, the model separates the causal effect by differencing out confounding factors (Wooldridge, 2013). The difference-in-difference estimator shows a positive weak significant relationship between the ETD-inferred improvement of shareholder rights and acquirer’s returns.

However, this research model has its limitations. Since, the decision where to incorporate or invest can cause self-selection of companies to be assigned to the treatment group by unobserved determinants. Therefore, the classifications of the member states into the treatment- and control group is prone to self-selection bias. This study adopts a derivation of Heckman’s correction model (Heckman, 1976) to correct for this self-selection bias. Where, the index of the level of anti-director constructed by (La Porta et al., 1998) serves as instrument of the unobserved determinants.

(8)

8 This study contributes to the existing literature regarding corporate governance and takeovers. This study weakly supports the evidence for a causal relationship between legal shareholder rights and acquirer’s returns. Prior research focused on the influence of firm level corporate governance and firm performance. In addition, most literature emphases almost completely on the effect of country level legal shareholder rights and external finance. In this study, I tried to establish a more direct connection between country level legal shareholder rights and acquisition efficiency.

The rest of this study is proceeds as follows, section 2 introduces the European takeover directive, proposes the hypothesis. The third section expresses the data, methodology, covers the sample-selection bias of the research and reports a summary of the descriptive statistics. Section 4 reports the main results of this study followed by implications of the research. This includes univariate estimates of the average abnormal returns of acquiring companies before and after employment of the ETD, estimates of a multivariate model based on a difference-in-difference approach, an analysis based on Maddala’s correction model of self-selection bias, and robustness tests. The fifth section is a conclusion of this paper.

2. The Introduction of the European Takeover Directive

2.1. Investor rights and bidder’s returns

Earlier research around the topic of investor rights and corporate valuation by La Porta et al. (2000) examines the influence of the level of shareholder protection on valuation of corporate assets. They argue that large publicly traded firms generally have controlling shareholders (La Porta, Lopez-de-Silanes, & Shleifer, 1999), which causes entrenchment for corporate management. The controlling shareholders have the power to expropriate minority shareholders, within the constraints imposed by the law. Better investor rights diminish the possibilities to expropriate minority shareholders and lead to a reduction of these agency costs. The authors find consistent with theory that better investor rights are empirically associated with higher valuation of corporates assets.

Other prior work of La Porta et al. (1998) scrutinizes the importance of country level corporate governance and the functioning of financial markets, since it gives investors the right to claim their fair return on investment by enforcing legal rights. Agency costs are a result of the separation of ownership and control, because managers or controlling shareholders have control rights (discretion) and shareholders cannot write complete contracts to align their interest with the managers’. Managers and controlling shareholders can pursue value destroying actions such as self-dealing and transferring of resources out of the company for personal benefits instead of maximizing shareholder value. The market of takeovers is a fitting environment to examine this argument, as agency costs are

(9)

9 readily observable. Prior research on this topic showed a negative effect of low-level of shareholder rights on firm value. For example, self-dealing behaviour such as engaging in empire building (Jensen, 1986) or picking targets by entrenched managers (Harford, 2012).

There is a recognized literature regarding firm-specific corporate governance and the returns of takeovers. Companies intend to govern their firms as good as possible, to monitor their management and with that reducing the costs induced with the separation of ownership and control. Therefore, a standard of a country’s level of legal shareholder rights plays a significant role in minimizing the agency costs. Further, the takeover market is a suitable option to examine, since a considerable wealth of assets and control rights are reallocated. While corporate insiders exert their power to redirect wealth to benefit themselves (Harford, 2012; Moeller, Schlingemann & Stulz, 2005; Jensen, 1986). For that reason, it is of great importance to understand the process and the influence of legal shareholder rights on allocation of wealth with takeovers.

In addition, Masulis, Wang and Xie (2007) examine the effect of firm level corporate governance on the profitability of firm’s acquisition. The authors find that acquirers with more anti-takeover provisions experience significantly lower abnormal stock returns. This supports the hypothesis that managers at firms protected by more anti-takeover provisions are less subject to the disciplinary power of the market for corporate control and thus are more likely to indulge in value destroying acquisitions.

Also, country level legal shareholder rights are important for acquiring companies when selecting a target. Since, in some cases public firms do not improve their corporate governance beyond the country’s default level due the fact that changes in corporate governance are costly (Bergman & Nicolaievsky, 2007). Moreover, Doidge, Andrewkarolyi, & Stulz (2007) argue that in countries with less developed capital markets, country-level corporate governance better explains distinctions in governance ratings than firm-level corporate governance. In addition, Rossi & Volpin (2004) state that an increase in shareholder protection is related to an increase in takeover activity. Further, they argue that with cross-border deals, targets are typically from countries with poorer investor protection than their acquirers’ countries. This indicates that with cross-border transactions corporate governance plays a role by improving the degree of investor protection within target firms.

Preceding work by Renneboog, Martynova, & Goergen (2005) examine the influence of reforming takeover law reforms in Europe prior to the introduction of the ETD. They find that takeover regulation has a significant impact on a firm’s corporate governance system. They argue that the adoption of a unified takeover code may result in dispersed ownership. With that, dispersed

(10)

10 ownership leads to an enhancement of the separation of ownership and control and decrease in the expropriation of minority shareholders.

Other work regarding investor protection laws and firm performance by (Agrawal, 2013; McLean, Zhang, & Xhao, 2012) find that investor protection is associated with a significant impact on firm’s corporate policies, performance and investment decisions. To be more specific, Agrawal examines the effect of the introduction of new state regulations established as safeguards for investors against securities fraud. He finds strongly supportive evidence that the introduction of these investor protection laws had a positive significant impact on firms performance and corporate policy.

To summarize, country’s legal institutions have an important role in reducing the discretion of insiders and therefore reducing agency costs between firm’s management & minority shareholder and controlling shareholders.

2.2 Institutional Framework of ETD

Prior to the implementation of the European Takeover Directive (ETD) the rules relating to the conduct of takeover offers varied widely throughout the European Union. As a result, takeover offers could not be undertaken with the same expectation of success across different EU member states and shareholders did not have the same protection and opportunities. Against this background, the EU’s general objective with the ETD is to harmonise shareholder protection laws in the community and to promote integration of European capital markets. The ETD’s more specific goals include: i) legal certainty on the takeover bid process and common clarity and transparency with respect to takeover bid throughout all EU member states, ii) protection of interest of (minority) shareholders, employees and other stakeholders, when subject to a takeover bid. The ETD comprises several rules that constitute to changes in legal shareholder rights. These regulations are the adoption of the mandatory bid rule, board neutrality rule, and the squeeze-out and sell-out right, which will be later elaborated in section 2.3. After the European Commission announced the transition in 2004, these regulations had to be enacted into national law by 21st_{May 2006.}

The new imposed regulation has a different impact on the member states, since the level of quantified change depends on three factors. The content of the present legal framework in 2006, its application in practice by supervisory authorities and its application and perception by interested parties. In this research approach, I produce a mapping of the EU member states into a treatment group and control group based on the significance of change imposed by the ETD. Countries that significantly had to improve their legal shareholder rights are assigned to the treatment group, while countries who had not to significantly improve their shareholder rights are assigned to the control group, this is based on the research of Marccus Partners (2013). The aim of the research is to measure

(11)

11 the difference in acquirer returns affected by significant changes in legal shareholder rights pre-post the implementation of the ETD, in comparison to acquirer returns from the control group where the ETD made no significant change in legal shareholder rights. Table 1 of the appendix presents a detailed summary of the proposed changes of the ETD provision by country. The observations are assigned based on the country of the bidder. The rationale behind this is that when a bidder fully acquirers a target, then the target adopts the nationality of the bidder and the legislation of the bidder’s country applies. Furthermore, according to Martynova & Renneboog (2008) there are significant spill-over effects caused by corporate governance manners when a merger is cross-border. The authors argue that when an acquirer is from a country with strong shareholder orientation relatively to the target, that the part of the total synergy value of the takeover may result from the improvement in the governance of the target assets.

2.3 The ETD and bidder’s returns

Prior work more directly relating to the introduction of the ETD and acquirer returns found varied results. Humphery-Jenner (2012) argues that the implementation of the ETD lead to a negative relationship with acquirer returns. The author states that the ETD may have increased value destroying takeovers. Because, the introduction of the ETD lead to legal uncertainty caused by ambiguity of some ETD rules, which consecutively lead to entrechment of managers. Furthermore, the increased costs of the ETD provisions have a negative impact on bidder returns. As supported by the report of Marccus Partners (2013), the disadvantage of the mandatory bid rule (equitable prices to all shareholders) is that it increases the costs of an acquisition. However, Humpherey Jenner’s study is different from this research approach, as he uses non-European takeovers for the composition of the control group.

Despite these contradicting evidence and reasons, it is hard to state whether the implementation has a positive or negative effect when comparing European and non-European acquisitions. According to McCahery and Renneboog’s (2003), the ETD has a positive value enhancing effect on intra-European mergers. In their exhausive research of the effect of the ETD provisions, the authors conclude that there could be value achieved by creating an active cross-border acquisitions market that protect minority shareholder and promote higher disclosing standards.

To further specify, the ETD aims to harmonize the rules relating to: the conduct of takeover offers and the protection of (minority) shareholders of target companies in the European Union. The ETD restricts corporate insiders’ discretion in several forms. For example, it addresses conflicts of interest between block holders and minority shareholders by giving both sides the option to sell their stake at a fair price with the combination of the out and sell-out right. The right of squeeze-out is combined with a sell-squeeze-out right enabling minority shareholders to require the majority

(12)

12 shareholder to buy their securities following a takeover offer at a fair consideration if the bidder meets the requirements for the takeover-related squeeze-out.

Further the ETD focuses on resolving these conflicts of interest between both management and controlling shareholder and (minority) shareholders. For instance, the mandatory bid rule enhances the protection of minority shareholders, particularly in concentrated ownerships. Because, the new ruling prohibits partial bids and obliges the bidder to extent a binding bid to all shareholder for an even price. In this manner, bidders can no longer extract private benefits of minority shareholder by cashing in takeover premia (McCahery & Renneboog, 2003).

In addition, according to the ETD, EU member states shall ensure that companies subject to the ETD publish detailed information about the bid offer. This contains information necessary to enable the shareholders of the target company to reach a properly informed decision on the bid. Further, the new restrictions imply that the board presents an explanatory report to the annual general meeting of shareholders on subjects such as: capital structure; significant direct and indirect shareholdings (including pyramid structures and cross-shareholdings); restrictions on voting rights; agreements between the company and board members for compensation.

The information which must be provided according to the ETD makes it much easier for a bidder to conduct a meaningful due diligence of the target company with publicly available information, something that is particularly important in hostile situations. Therefore, these stricter disclosure requirements improve access to information, resulting in more efficient markets (La Porta, Lopez-De Silanes and Schleifer, 2006).

Finally, by introducing the board neutrality rule the ETD addresses the conflict between different stakeholders. The new article in the law proposes to restrict the options of anti-takeover provisions (ATP) for the target companies in the EU member states. The rule partly eliminates the classical agency conflict between managers and shareholders by prohibiting the adoption of antitakeover provisions during bids. Since ATP’s weaken shareholder rights (Bebchuk and Cohen, 2005) and managers who are bad bidders are entrenched will engage in “empire-building’’ behaviour (Humphery-Jenner and Powell, 2011).

Built on the abovementioned arguments, I believe that the acquirer’s returns increase if country’s legal shareholder rights improve. Thus, the introduction of the ETD has a positive impact on resolving the conflict of interest around the agency problem, better transparacy due to new disclosing standards. Taking all this together, I assume that the improvement of legal shareholder rights caused by the implementation of the ETD will have a beneficial effect on bidder returns.

(13)

13

3. Sample Description

The data comprises 659 intra-European acquisitions announcements of completed takeovers from January 2001 until December 2017. The information comes from the Thomson Reuters M&A database and complies with the criteria that both the acquirer and target company are incorporated in an EU member state. Where the takeover complied to the following criteria: the transaction resulted in a change of control. A change of control is defined as: the acquiring company owns a controlling interest of at least 50 % of shares or more of the target company after the acquisition. Furthermore, the deal value disclosed from the Thomson Reuters database is at least $ 1 million dollar. In addition, the acquirer has annual financial statement information and stock return data (110 trading days prior to acquisition announcements) available from Datastream database.

3.1 Variable Construction

In the next subsections, the measurement of the dependant variable, acquirer returns, key explanatory variable and bidder and acquirer characteristics as control variables are discussed.

3.1.1 Acquirer returns

This study measures the effects of bidder announcements by market model adjusted stock returns around acquisition announcements. The announcement dates are obtained from Thomson Reuters’s M&A database. The acquirer’s returns are calculated by exploiting the OLS market model conform the standard event study methodology proposed by MacKinlay (1997). This study makes use of an estimation window of [-100,-10] and event window of [-5,5] trading days, where day zero is the announcement date. The abnormal returns are calculated by subtracting firms expected normal returns from the firm’s actual return. The expected returns are the market returns if the transaction had not occurred, with the Euro Stoxx 600 as benchmark index.

3.1.2 Improvement of shareholder rights

The key explanatory variable that indicates the improvement of legal shareholder rights is the interaction between the two dummy variables treatment and period after the implementation of the ETD. The periods before and after the implementation date, 21st_{May 2006, of the ETD signify the}

before and after treatment period. The sample’s treatment and control group are stipulated by reports of (Marccus Partners, 2013) and the (European Commission, 2012). Countries are assigned to the treatment group when the country’s legal shareholder rights are significantly improved by the implementation of the ETD, otherwise to the control group. Specifically, the countries are assigned to the treatment group if the countries legal system adopted at least a significant change in one of the proposed rules of the ETD.1_{The treatment group is comprised of: Estonia, Belgium, Germany, Greece,}

(14)

14 Finland, Hungary, Italy, Luxembourg, Netherlands, Poland and Spain. While the control group consists of: Austria, Denmark, France, Ireland, Portugal, Sweden, UK.

3.1.3 Acquirer-, Deal- and Target characteristics

Besides the improvement of legal shareholder rights, there are other variables such as deal and acquirer characteristics potentially determining the level of acquirer returns after an announcement of an acquisition. This study control for acquirer- and deal characteristics to capture the effect of these variables on bidder’s returns. Furthermore, this study controls for target characteristics in the endogenous treatment model.

The bidder characteristics that are controlled for are company size, Tobin’s Q and leverage ratio. All these variables are measured at the fiscal-year end prior to the announcement date. Earlier work on the relationship between these control variables and bidder returns find that, acquirers who have a higher level of Tobin’s Q realize higher acquirer’s returns (Servaes, 1991). On the other hand, recent findings by (Harford, 2012; Masulis,2007), report no significant relationship between Tobin’s Q and bidder’s returns. This study defines Tobin’s Q as the ratio between the bidder’s market value of the firm and the book value of total assets. Where the market value of the firm is obtained from Datastream at the announcement date of the acquisition.

Moreover, preceding work on the effects of bidder’s size and acquirer returns show that the size of the acquirer has a negative relationship with returns. Due to managerial hubris (Roll, 1986) large firms tend to overpay the value of a target and make more value destroying acquisitions (Harford, 2012). However, Moeller et al.(2004) argue that the size of a publicly traded bidder has a positive effect on acquirers returns. This study expresses the size of the firm as the log transformation of total assets at fiscal-year end prior to the announcement date.

Furthermore, this study controls for financial leverage, as prior research by Jensen (1986) shows that financial leverage is an important governance-related factor. Since higher debt levels have a negative effect on future cash flows, therefore limiting managerial discretion and agency costs. Additionally, there is also evidence that firm’s leverage provides incentive for managers to improve firm performance, since managers have to give up control and might lose their job when the firm is in financial distress. Therefore, I expect that the firm’s leverage has a postive effect on acquirer’s returns. Firm’s leverage is defined as the ratio between the book value of long and short term debt over the book value of total assets.

The deal specific variables that are controlled for are the covariates: deal size, deal destination, deal attitude, and type of payment. Prior research regarding deal size and acquisition premia show contradicting evidence as Loderer (1990) state that acquirers experience greater losses

(15)

15 when buying large targets because bidders are more likely to pay too much. On the other hand, Alexandridis et al. (2013) find a negative relation between offer premia and target size, indicating that acquirers tend to pay less for large firms, not more. These variables are defined as i) the log of the deal value in millions, ii) a dummy variable taking the value of one if the deal was hostile, zero otherwise and iii) a dummy variable taking the value of one if the deal was fully paid in cash, zero otherwise.

The variables that control for target characteristics in the endogenous treatment model are target’s Tobin’s Q and the target’s initial anti-director rights. The targets Tobin’s Q is a proxy for the profitability of the firm and is a determinant of selection criteria for the acquiring firm. The target’s Tobin’s Q is defined as the ratio between the market value of the targets equity over the book value of assets. To control for the target’s initial anti-director rights, I make use of the index constructed by La Porta et al. (1998). It gives an indication of target’s initial anti-director rights on country-level. The authors, assigned for each country an anti-director rights score which is based on their rights regarding voting rights on directors (abbreviated as LLSV). The index scores the countries anti-director rights through a rating from 0-5, zero being the lowest amount of anti-director rights and 5 the highest. A detailed mapping of the country’s corporate governance scores can be found in table. The rationale behind determining the level of anti-director as explanatory variable in the selection model is, is that the anti-director rights of the targets country of residence influences the choice of acquisition. A more detailed of the country’s index scores can be found in table 2 of the appendix.

3.2 Descriptive Statistics

Figure 1 depicts the number of acquisition announcements of intra-European acquisitions between publicly traded firms. The figure shows that the number of announcements decreased over time from 2001 until 2018. The red vertical line indicates the date of enactment of the ETD on 21st_{of May 2006.}

The acquisition announcements increase from 2002 onwards until 2007 due to improvement of the aftermath of the dot-com crisis. In 2006, the year of the implementation of the ETD the number of acquisitions slightly increase peaking in 2007. Thereafter, the number of acquisitions decrease significantly in the following period. The substantial decrease in the number of acquisition announcements is caused by the recent economic meltdown. This study controls for this pattern by including year fixed effects in the research design.

(16)

16 Figure 1

Figure 1 presents the number of completed intra-European acquisitions, where both acquirer and target are publicly traded firms during the sample period 2001-2017.

Table 3 describes the sample’s summary statistics. The table shows a summary of deal characteristics, acquirer characteristic and general deal info. Over the whole sample, the acquisitions are most often deals between two companies of the same country. Further, almost half of the deals are financed wholly by cash. Examining the deal and acquirer characteristics from the treatment and control group, I find that the acquisition and the size of the bidders are larger in the treatment group relatively to the control group. Further, acquisitions from countries from the treatment group are relatively more often cross-border deals, which might indicate that investors tend to seek deals in countries with different litigation regarding shareholder rights. Comparing the deal characteristics to US takeovers, European company’s deal size and acquirer’s assets to deal size ratio are relatively smaller comparing to firms of the United States. However, comparing the sample to US bidder characteristics, European bidders are larger, are more leveraged, and have a lower Tobin’s Q (Masulis, 2007).

(17)

17 Table 3

This table presents descriptive statistics of 659 intra-European mergers between the period January 2001 and 31st December 2017 in European countries. Description of the variables is included in the appendix. Figures are in million dollars. The sample is composed by deal size, acquirer characteristics and general deal info.

Deal Size

Acquirer characteristics

General deal info Median Deal Size Median Total Assets Median Tobin's Q Median Leverage Number of deals Domestic Percentage All-Cash Payment Control Group Austria 114.0 5,640 0.44 0.29% 7 42.3% 57.1% Denmark 24.1 10,300 0.40 30.4% 19 57.9% 42.1% France 78.1 14,600 0.47 27.2% 145 71.7% 51.0% Ireland 27.7 979 0.89 34.0% 4 50.0% 75.0% Portugal 81.7 4,700 0.22 47.4% 8 75.0% 87.5% Sweden 38.9 5,840 0.61 23.4% 55 69.1% 34.6% UK 52.1 213 0.69 18.0% 209 88.5% 41.1% Subtotal 53.9 2,010 0.57 23.5% 450 78.2% 44.89% Treatment Group Belgium 157.0 10,800 0.56 27.3% 19 26.3% 57.9% Finland 181.0 1,500 0.51 30.3% 10 40.0% 60.0% Germany 76.1 3,570 0.28 25.6% 55 67.3% 54.5% Greece 136.0 4,700 0.53 41.8% 9 100% 44.4% Italy 347.0 6,000 0.33 33.4% 28 50.0% 50.0% Luxembourg 45.7 355 0.79 58.4% 2 0% 50.0% Netherlands 195.0 2,500 0.73 23.3% 27 48.2% 55.6% Poland 20.9 4,800 0.47 15.5% 19 100% 31.5% Spain 420.0 10,600 0.42 39.5% 32 50.0% 59.4% Subtotal 162.0 5,230 0.44 29.5% 203 57.6% 53.2% Total 73.5 2,091 0.53 25.2% 659 71.8% 47.5%

3.3 Empirical model

To test the effect of the improvement shareholder rights on acquirers returns, I will conduct a difference-in-difference approach. Where the cumulative abnormal returns (hereafter CAR’s) upon announcement date of an acquisition is used as dependent variable, whereas the difference-in-difference variable estimates the effect of the improvement of shareholder rights on the CAR’s.

The main variable of interest of the right-hand side of the equation is the interaction between the dummies treatment (improvement shareholder rights) and period after the implementation of the ETD. A difference-in-difference model is employed to test average treatment effect of the implementation of the ETD in countries with weak initial shareholder rights on the bidder’s returns.

(18)

18 The difference-in-difference model is arranged as follows:

[(∆𝑌₁𝐼𝑚𝑝𝑟𝑜𝑣𝑒𝑚𝑒𝑛𝑡|𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡 𝐺𝑟𝑜𝑢𝑝) − (∆𝑌₁𝑁𝑜 𝑖𝑚𝑝𝑟𝑜𝑣𝑒𝑚𝑒𝑛𝑡|𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡 𝐺𝑟𝑜𝑢𝑝)] - [1] [(∆𝑌₁𝐼𝑚𝑝𝑟𝑜𝑣𝑒𝑚𝑒𝑛𝑡|𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝐺𝑟𝑜𝑢𝑝) − (∆𝑌₁𝑁𝑜 𝑖𝑚𝑝𝑟𝑜𝑣𝑒𝑚𝑒𝑛𝑡|𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝐺𝑟𝑜𝑢𝑝)]

The improvement indicates the period after the implementation of the ETD, treatment group signifies the acquisitions that had been taken place in countries with initial weak shareholder rights and control group otherwise.

Central for the internal validity of the difference-in-difference approach is the parallel trend assumption, which implies that without the treatment the treatment and control group follow the same trend.The treatment, improved shareholder rights, must be exogenous to the shareholder laws already in place to estimate its precise effects on the acquirer’s returns. The implementation of the ETD by the European Commission was an exogenous law change imposed to promote European capital markets and harmonize legal shareholder rights (Marccus Partners, 2013).

3.4 Sample selection issues

3.4.1 Self-selection bias

Besides threats of internal validity due to the violation of the parallel trend assumption, the research design is prone to selection-based endogeneity. Since the sample could be non-representative of a true population and thus threaten internal validity. Based on earlier work of Heckman (1979) the sample of the control and treatment group could be prone to self-selection bias. Self-selection bias is of a different nature than sample-selection bias, as in these empirical contexts there are no issues regarding to the dependent variable not being observed for relevant subsamples of the population. Instead, the self-selection concern arises when the dependent variables are observed for different sub-samples, yet a non-randomness is involved with the manifestation of these dependent constructs. This process is related to omitted variable bias since the fundamental source of endogeneity is that factors that cannot be observed by the research have an influence on the selection process and the potential outcome (Clougherty, 2016). For example, a firm’ decision of the location of incorporation could be determined by hidden covariates such as tax system, existing legal shareholder rights and legislature. These variables may both determine the selection into the treatment group as the outcome variable, acquirer’s returns.

This research design could be prone to threats of self-selection, since firm’s management could influence the decision of the location of incorporation of a company. Same as illustrated in the example above, the firm’s decision where to locate is determined by several variables. In the case of

(19)

19 the introduction of the ETD, firms may decide to incorporate or to invest more in countries where the introduction of the ETD has a stronger effect on the improvement of shareholder rights.

3.4.2 Endogenous treatment model

To resolve the threat of internal validity around self-selection bias problem, I use Maddala’s (1983) endogenous dummy-variable model with the maximum likelihood estimator. The endogenous dummy-variable model is a derivation of the Heckman correction model (Heckman, 1976). The model estimates an average treatment effect and other parameters of a linear regression model that also includes an endogenous binary-treatment variable.

Following the same methodology as (Clougherty, 2016) for self-selection bias, I adopt a latent variable approach (Maddala, 1983) that is similar to the Heckman procedure for treating sample-selection issues. The endogenous- treatment issue is a classical endogeneity-problem that can be resolved with an instrumental variable (IV) approach. However, in this case the potentially endogenous variable is a discrete dummy variable instead of a continuous variable. In this case the discrete dummy variable of interest is the interaction term between treatment and post-ETD.

As stated by (Clougherty, 2016), several disadvantages exist with regard to employing an IV approach for sample-selection bias. For instance, the coefficient estimates are sensitive to the observation range of the employed data and could over- or underestimate coefficients (Wooldridge, 2013). Hence, the most appropriate approach to deal with the self-selection issue is to endorse a latent variable approach that is akin to Heckman’s procedure for treating sample-selection issues. Where the latent variable z is modelled as follows:

𝑍𝑖 = {_{0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒} 1 𝑖𝑓𝑍_𝑖∗>0

, where [2]

𝑍𝑖∗= 𝛽0+ 𝛽1𝑇𝑎𝑟𝑔𝑒𝑡′𝑠 𝑄𝑖+ 𝛽2𝑇𝑎𝑟𝑔𝑒𝑡′𝑠 𝐿𝐿𝑆𝑉 𝑖+ 𝛽3𝐷𝑒𝑎𝑙 𝐶ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠𝑖+ 𝜔𝑖 [3]

𝐶𝐴𝑅𝑖 = 𝛿𝑍𝑖∗+ 𝛽1𝐴𝑐𝑞𝑢𝑖𝑟𝑒𝑟 𝐶ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠𝑖+ 𝛽2𝐷𝑒𝑎𝑙 𝐶ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠𝑖+ 𝜀𝑖 [4]

The latent variable Z is determined by the partition of the sample in two exclusive sub-samples. First, if the merger takes place after the implementation of the ETD and in a country of the treatment group, then (𝑍𝑖∗=1). Second, if the merger takes place in the period before the implementation of the ETD

and in a country of the control group, then (𝑍𝑖∗=0).

The control covariates, target characteristics and deal- specific characteristics, are added to better identify the selection model, since these characteristics should explain the selection mechanism into the treatment group but not directly determine acquirer’s CAR’s. The variables 𝑇𝑖 and 𝐷𝑖 are

(20)

20 characteristics that are controlled for are the variables target’s Tobin’s Q and the target’s level of anti-director rights (abbreviated as LLSV) modelled by the index of (La Porta et al.,1998). The anti-anti-director rights of the target’s country does not directly influence the returns of acquirer’s upon acquisition announcement since the acquiring company abides to its own country level of anti-director rights. Further, the deal characteristics that are controlled for are deal size, deal attitude, whether the deal is cross-border and the type op payment.

The acquirer characteristics that are controlled for are firm size, firms’s Tobin’s Q and leverage ratio. Since, as mentioned in the variable construction section, these variable influence acquirers CAR’s.

The error terms of the both the selection and main model, 𝜔𝑖and 𝜀𝑖 respectively, are normally

distributed with mean zero and the variance - covariance matrix is characterized by a correlation coefficient rho (𝜌), with 𝜎𝜔 = 1. The likelihood-ratio test tests if the correlation between the error terms is not zero, 𝜌 ≠ 0, then the interaction variable (difference-in-difference) estimator is endogenous to acquirer’s returns.

[(𝜎

2_𝜌𝜎

𝜌𝜎 1)]

Like the previous estimated models, the dependent variable of the endogenous treatment model is the eleven-day CAR’s of the acquiring firm. As mentioned above, the main explanatory variable Z is the variable where the observed agents could choose to be part of the treated group, in this case the interaction term between the dummies post-ETD and treatment group. In addition, the control variables 𝐴𝑖and 𝐷𝑖 are acquirer characteristics and deal-specific characteristic, respectively. With the

normally distributed error term the model results in a probit model where 𝐶𝐴𝑅0is the outcome that

firm i obtains if firm i selects treatment 0, and 𝐶𝐴𝑅1is the outcome that firm i obtains if firm i selects

treatment 1.

3.4.3 Further Limitations

Besides the potential endogeneity issues, one other weakness arises regarding the causation of the empirical model with the difference-in-difference estimation. Because, in case of a full acquisition, the target company becomes subject to the acquirer’s country’s legal system. However, the change of control of the target leads to several changes of corporate governance at firm-level such as anti-takeover provisions, managerial equity ownership and size of the board. Briefly, the changes of shareholder rights on country-level go together with changes of corporate governance on firm-level. Companies have different views on interpreting and implementing country-level shareholder rights

(21)

21 together with firm-specific corporate governance. The omission of the firm-specific corporate governance practices leads to a limitation of the research approach.

Because, this has the implication that when the improvement of legal shareholder rights reports a significant effect on the returns of acquirers, this could (partly) be credited to the adoption of the acquirer’s firm-specific corporate governance practices by the target company. The magnitude of the effect improvement of country-level legal shareholder rights could be reduced. Nonetheless, the sign of the causal relationship remains the same.

Moreover, the goal of the European Commission with the ETD is to harmonize legal shareholder rights across all member states. Countries with weak initial shareholder rights improve their country- level corporate governance after the implementation of the ETD. As years pass after the enactment of the ETD, the countries with initial weak shareholders move towards the level of shareholder rights of countries of the control group. Therefore, the effect of the treatment may not be constant over time.

4. Empirical Results

4.1 Results multivariate model

This study uses the methodology proposed by the lecture notes of (de Jong, 2011) to calculate the test-statistic of the average of the CAR’s. The test statistic is calculated in the following manner:

𝐻0: 𝐸(𝐶𝐴𝑅𝑖) = 0 and TS = √N ∗𝐶𝐴𝐴𝑅 𝑠 [5] where, 𝑠 = √ 1 𝑁−1 (𝐶𝐴𝑅−𝐶𝐴𝐴𝑅) 2 [6]

The amounted average cumulative abnormal returns are depicted in table 4 below. The CAR’s are positive and significant over the whole sample period and over both treatment group and control group. The average CAR’s amount to 1,18 % percent and this finding indicates that firms in the EU-15 make value enhancing acquisitions during the period 2001-2017. When further analysing the CAR’s, I find that acquisitions before the implementation of the ETD led to negative returns, average CAR’s of (-3,66%) while transactions after the implementation of the ETD led to positive returns, average CAR’s of (4,27%). These results are in line with the theoretical framework, since they indicate that the introduction of the ETD had a positive effect on acquirers returns after acquisition of a public company.

(22)

22 Table 4

This table reports the results of the average cumulative abnormal returns by assignment of control or treatment group and before and after the period of implementation of the ETD. The test statistic of the CAR’s is calculated as follows:

𝐻0: 𝐸(𝐶𝐴𝑅𝑖) = 0 and TS = √N ∗𝐶𝐴𝐴𝑅_𝑠 where, 𝑠 = √ 1

𝑁−1 (𝐶𝐴𝑅−𝐶𝐴𝐴𝑅)2 . *** indicates significance at 1% level, ** at 5% level and *

at 10% level based on two-sided tests.

Average CAR: Results for OLS Market Model

CAR TS Overall 1,18% -5.4844*** Treatment group 8,12% -2.4709*** Control group -4,22% -0.4094 Post-ETD 4,27% -1.1954 Before-ETD -3,66% 0.2163

Treatment group & Post-ETD 1,20% -2.0326**

Treatment group & Before-ETD 0,18% -1.6272

Examining, figure 2 more closely, I find a somewhat disentangling trend between the treatment and control group’s average CAR over the sample period indicating that the effect of the ETD is stronger for companies incorporated in the treatment group. However, the results are driven by a sharp increase in CAR’s in 2012, where the average CAR’s for the treatment group were almost 8 percent in comparison to approximately -1 percent for the control group. The development of the CAR’s show that the returns of the control group remain stable over the sample period, while the CAR’s of countries of the treatment group exceed the control group’s CAR’s after 2005 in general. The red vertical line shows the date of the enactment of the ETD. The rationale behind the fact that the CAR’s of the treatment group are higher even before the implementation date of the ETD, could be caused due to countries anticipating on the changes of the ETD.

(23)

23 Figure 2

Development of CAR’s

Figure 2 presents the progress of the cumulative abnormal returns of both the treatment and control group over the sample period 1st_{January 2001 until 31}st_{December 2017. I have conducted the OLS market model to estimate the} cumulative abnormal returns as proposed by (MacKinlay, 1997) with [-100; -10] and [-5;5] as estimation window and event window respectively relating to announcement date of the acquisition. To calculate the normal returns, I have used the Europe Stoxx 600 as a benchmark index for market returns. The treatment group is comprised of: Estonia, Germany, Greece, Hungary, Luxembourg, Netherlands, Poland and Spain. While the control group consists of: Austria, Belgium, Denmark, Finland, France, Ireland, Italy, Portugal, Sweden, UK.

The results from the univariate tests on the CAR’s, depicted in figure 2 indicate that the acquirer’s returns increased in countries of the treatment group after the implementation of the ETD, while the CAR’s of the control group remained constant over time. These results suggest that the ETD have a value -increasing impact on the affected countries. To farther research this hypothesis, I employ a difference-in-difference test where variable DiD is the interaction variable of the dummies treatment group and Post-ETD. The two dummy variables capture the effects of the improvement of takeover law (treatment group) after the implementation of the ETD (Post-ETD). The two dummies control for time-invariant differences and common trends between the treatment and control group. For that reason, if the difference captured by the DiD variable is positive and significant, it will support the assumption that improvement of takeover law causes an increase in acquirer returns.

Table 5 shows the regression results of the difference-in-difference approach. The first two models are not controlled for time-fixed effects in contrast to the last two models. In model 1 and 3,

(24)

24 the dependent variable 11-days CAR’s are regressed on the independent variables without the inclusion of deal and firms specific control variables. The estimates in table 5 report consistent results over all four models, the difference-in-difference estimator remains positive for all models and significant for model 2 and 4. The difference -in-difference estimator DiD is significant and positive for model 2 and 4. In for example model 4, the DiD is significant on 5 percent level and is 0,00839. This indicates that on average, acquiring firms from countries of the treatment group earn 0,84% higher returns in comparison to firms incorporated in countries of the control group. Based on the median of the acquirer’s market value, the improvement of takeover law translates to an economic significant increase in returns of approximately on average $11,5 million per deal. Juxtaposing the difference-in-difference estimator Post X Treatment with the dummy variable post-ETD leads to the finding that the improvement of takeover law exceeds the decreasing trend in acquirer’s returns after the implementation of the ETD. To sum up, the results from table 5 suggest that there is a weak causal link between improvement of takeover law and acquirer’s returns.

For the control variables of acquirer and deal characteristics I find consistent estimates over model 2 and 4. Most estimates correspond to earlier findings regarding acquirer and deal characteristics influence on returns. As argued by Moeller et al. (2004) the size of public acquiring firms has a positive significant influence on acquirer returns. In line with findings of Masulis (2007) the results show that Tobin’s Q has no significant relationship with acquirer returns. Further, coherent with earlier findings of Yook (2003), I find that cash offers have a slight positive impact on acquirer’s returns after announcement of a merger. As shown by (Moeller, Schlingemann, & Stulz, 2005), the results from table 5 show that a hostile deal attitude has a detrimental effect on acquirer’s returns after announcement of a merger. Furthermore, Alexandridis et al. (2013) state that a larger deal would lead to higher acquisition premia and smaller returns on deal announcement, consistent with this statement, I find a negative significant effect of deal size on acquirer’s returns.

(25)

25 Table 5

Regression results of Difference-in-Difference test

This table presents the regression results of the difference in difference analysis of the cumulative abnormal returns of 662 European acquisitions during the period 2001-2017. The dependent variable are the abnormal returns in the 11-days OLS market model CAR for an announcement of an acquisition. The independent variables are defined below. The variable of interest is the difference-in-difference estimator, DiD, which is an interaction term of the dummy variables POSTETD and treatment group. In contrast to models 3 and 4, the first two models do not include year fixed effects. All standard errors are adjusted for

heteroskedasticity and are reported in the parentheses. The ***, ** and * indicate statistical significance at 1; 5 and 10 percent, respectively.

Dependent variable: OLS Market Model 11-days CAR

(1) (2) (3) (4) POST-ETD 0.0084*** 0.0058*** -0.0098* -0.0135** (0.0018) (0.0020) (0.0057) (0.0060) Treatment group 0.0118*** 0.0083*** 0.0114*** 0.0077** (0.0032) (0.0032) (0.0034) (0.0032) Post X Treatment 0.0014 0.0094** -0.0017 0.0084** (0.0039) (0.0041) (0.0039) (0.0042) Acquirer characteristics (ln)Assets 0.00119*** 0.0010** (0.00462) (0.0066) Tobin’s Q 0.000465 0.00053 (0.00128) (0.0012) Leverage ratio -0.0184*** -0.0194*** (0.00475) (0.0050) Deal characteristics (ln)Deal value -0.00199*** -0.0021*** (0.000397) (0.00042) Domestic -0.0101*** -0.0116*** (0.00202) (0.0020) Hostile -0.0127*** -0.0157** (0.0046) (0.0066) All-Cash Deal 0.00553*** 0.0067*** (0.00166) (0.0018) Constant -0.00946*** 0.0116 0.000761 0.0284*** (0.00140) (0.00905) (0.00321) (0.0095) Observations 8,295 7,4722 _8,295 _7,4722 R-squared 0.010 0.023 0.004 0.020

Year FE YES YES

4.2 Results Endogenous Treatment Model

This section represents the paper’s central results. Table 6 presents the empirical results of the endogenous treatment model estimated with. All models are estimated with maximum likelihood estimators which should be more appropriate than the previous estimated difference-in-difference model. Because, they correctly treat the self-selection process within a non-linear probability (probit) model. The first model estimates the influence of the improvement of shareholder rights on acquirer’s

(26)

26 returns controlled for self-selection bias without control variables. Whereas, the second and the third model contain deal-specific, acquirer and target characteristics as control variables. In addition, in the third model is controlled for year and country fixed effects. All estimations include the dummy variable that captures the effect of the improvement of shareholder rights (Z= Post-ETD X Treatment) as the principal right-hand-side variable of interest. The dummy variable capturing the anti-director rights of the target’s country is repressed from the main estimations since it is identified as the main independent variable influencing the latent variable Z.

When considering the results from table 6, we observe throughout all three models a positive and significant relationship between the improvement of shareholder rights and the acquirer’s returns. Furthermore, the estimations of the interaction variable of the difference-in-difference model are not significant nor positive throughout all estimated models. While in the maximum likelihood model the estimations of the influence of the implementation of the ETD in countries of the treatment group has a positive significant effect in all models. The interaction variable of the dummies treatment and the period after the implementation of the ETD has a stronger effect (5.02%) when estimated with maximum likelihood in comparison to previously estimated difference-in-difference model (0.84%). Taking the median of acquirer’s market value into account we can give this result a more interpretable characterization of economic significance. The increase of 5.02 % on acquirer’s returns after merger announcement translates to an on average increase of approximately $ 69 million acquirer’s market value.

Besides, table 6 shows that the estimates of the acquirer’s characteristics are insignificant for all three models. This is contradicting with earlier findings and previous difference-in-difference and OLS estimates of the control variables. Earlier findings by (Moeller et al., 2004) and estimations of the difference-in-difference model, suggest a positive relationship between acquirer’s size and bidder returns. Also, table 6 shows that the control variables leverage ratio and acquirer’s Tobin’s Q have no significant effect on acquirers returns. This is supported by (Masulis, 2007), who finds no significant evidence of the influence of acquirer’s leverage ratio and Tobin’s Q on bidder’s returns. Further, the estimations of both target and deal control variables of the selection model are all insignificant in the third model. However, when controlling for country fixed effects in the third model, only the target’s anti-director score has a signifanct small negative effect on whether a firm is incorporated in a country of the treatment group after the implementation of the ETD.

(27)

27 Table 6

Regression results endogenous treatment model

This table present the results of the maximum likelihood estimation of bidders returns. A derivation of Heckman’s correction model, Maddala’s endogenous treatment model, is used to correct for self-selection bias. In the model two stages are estimated. The latent variable Z (Post X Treatment) and the influence of Z on the acquirer’s CAR’s. The data comprises the 11-day CARS’s of 659 acquisition announcements over the sample period 2001-2017. In all model’s standard errors are controlled for heteroskedasticity and year fixed effects, while Model 3 is controlled for country fixed effects. The ***, ** and * indicate statistical significance at 1; 5 and 10 percent, respectively.

(1) ML (2) ML (3) ML Model

Dependent Variable: CAR Z=1 CAR Z=1 CAR Z=1

Post-ETD -0.0126 0.0059*** -0.0274 (0.0213) (0.0020) (0.0208) Treatment group 0.0203*** 0.0274*** 0.0212*** (0.0061) (0.0025) (0.0074) (ln) Assets 0.0001 -0.0006 (0.0003) (0.0010) Tobin’s Q -0.0008 -0.0007 (0.0012) (0.0033) Leverage ratio -0.0239*** -0.0205 (0.0052) (0.0167) Z -0.0462*** 0.0574** 0.0502** (0.1512) (0.0224) (0.0169) LLSV - 0.3525** -0.3404*** -0.3597** (0.0021) (0.0493) (0.1543) Target Tobin’s Q -0.0075 (0.0050) (ln) Deal value 0.0236 (0.0373) Domestic 0.1450 (0.204) Hostile -0.4370 (0.5347) Lambda (λ) -0.0338*** -0.0042 -0.0320** (0.0112) (0.0037) (0.0118) Rho (ρ) -0.4576*** -0.5346*** -0.4307*** (0.0292) (0.0414) (0.1314) Constant -0.0830*** -0.6170 -0.0204** -0.9470*** -0.0219 -0.9400 (0.00153) (0.0365) (0.00802) (0.140) (0.0318) (0.948) Observations 659 659 526 526 526 526 Country FE No No Yes

Year FE Yes Yes Yes

LR-test p-value 0.0251** 0.0279 0.0043***

In addition, to the significant estimation of the average treatment effect reported by the maximum likelihood estimation, a significant effect for Heckman’s lambda (-0.0320) is also reported in the maximum likelihood estimation. Heckman’s lambda is the product of the correlation between the error terms of the selection and main models (ρ=-0.4307) and the variance of the error term of the selection model (σ=0.0775). The likelihood-ratio test indicates that we can reject the null hypothesis

(28)

28 of no correlation between the errors of the selection model and the errors of the main model. The estimated correlation between the errors of the selection model and the errors of the main model, ρ, is −0.4307. This negative relationship shows that the unobservable variables that increase acquirer’s CAR’s tending to incline with unobservable variables that lower initial anti-director of the target country increase of the acquirers being in a country of the treatment group after the enactment of the ETD.

To conclude, the most striking difference between using Maddala’s endogenous treatment model with maximum likelihood estimation and the normal difference-in-difference approach is the substantial difference in the average treatment effect of the interaction term treament and post-ETD. Whereas, in the difference-in-difference models the variable was either small and significant postive or even insignificant and negative, the estimations endogenous treatment model are substanially larger and positive significant. Yet, it is worth noting that if controlled for year fixed effects, several coefficient estimates become insignificant.

4.3 Results Better Target Selection

To test a potential channel that substantiates the hypothesis that the improvement of takeover law leads to higher acquirers returns, I will scrutinize the effect of the ETD on selecting better targets. Since the introduction of the ETD would lead to more transparancy in reporting of the targets, the implementation of the ETD would therefore lead to that the acquirers are better able to select more profitable targets. The research model of table 7 adopts the same difference-in-difference approach as used in table 4 to test wheter the introduction of the ETD contributes to picking better acquisition targets. To proxy for selecting better targets, I test whether firms of the treatment group avoid investing in low-Q target firms after the implementation of the ETD. In contrast to the regression of Table 4, the dependant variable in this test is changed to a dummy variable that takes the value of 1 if the target’s Tobin’s Q is in the first 25% quantile of the sample group and zero otherwise.

As shown by table 7, I find weak evidence that firms in countries of the treatment group after the implementation of the ETD avoid low Q-targets. In the first model, I find a weak significant negative effect of the difference-in-difference estimator, indicating that firms out of countries of the treatment group avoid investing in low-Q targets after the introduction of the ETD. Nonetheless, in all other models show that the DiD variable does not have any significant effect on investing in low-Q targets. In models 2 and 4 the deal and acquirers characterstics are added as a control variables, while models 3 and 4 control for year fixed effects.

(29)

29 Table 7

Target selection

Table 3 depict the tests whether the ETD has as significant effect on the likelihood to invest low – Q target firms by acquirers. The dependent variable is a dummy variable that takes the value of 1 for low-Q target firms, if a targets Tobin’s Q is in the first 25% quantile and zero otherwise. The data comprises 595 acquisitions over the sample period 2001-2017. In the first two models the effect is tested with OLS regression, in models 3 and 4 is controlled for year fixed effects and country fixed effects. Same as in Table 6 the independent variable of interest is DiD, interaction term of the dummies treatment group and Post-ETD. All standard errors are adjusted for heteroskedasticity and are reported in the parentheses. The ***, ** and * indicate statistical significance at 1; 5 and 10 percent, respectively.

Dependent variable: Tobin’s Q (1 in 25% quantile, 0 otherwise)

(1) (2) (3) (4) Treatment group 0.317*** 0.0131 0.1270 0.0164 (0.0623) (0.0603) (0.0611) (0.0655) Post-ETD 0.195*** -0.0649* -0.0612 -0.0555*** (0.0257) (0.0337) (0.0389) (0.0188) Post X Treatment -0.146* 0.0851 0.130 0.116 (0.0852) (0.0780) (0.0849) (0.0727) (ln)Assets 0.0376*** 0.0814*** (0.00581) (0.0143) (ln)Deal value -0.0210*** -0.0272*** (0.00707) (0.00804) Leverage ratio 0.125 0.192* (0.0912) (0.0988) Domestic 0.0208 0.119** (0.0314) (0.0470) Hostile -0.0975 -0.0680 (0.0802) (0.132) All-Cash deal -0.1170*** -0.1470*** (0.0301) (0.0322) Constant 0.264*** -1.072*** (0.0249) (0.274) Observations 595 595 595 595 R-squared 0.176 0.479 0.006 0.207

Year FE YES YES

5. Discussion

My findings suggest a positive, but weak relationship between improved shareholder rights and bidder’s returns. Acquiring companies from countries with initially weaker shareholder rights yield a 0,84-percentage point increase in returns to the acquirer after the introduction of the ETD. The increasing transparency imposed by the ETD seems to be the mechanism behind this effect. By looking at bidders’ target selection, I find that they are more likely to avoid less profitable targets, with lower Tobin’s Q, after the ETD. This result is robust because the coefficient is negative without controlling for other variables, but positive and insignificant after adding controls. Thus, only the model without

(30)

30 any control variable indicates that the improvement of shareholder rights leads to better target selection.

My analysis is limited since publicly traded companies can choose where to incorporate and which country’s legislation to abide. Since, companies have hidden incentives to incorporate or invest in companies whom are incorporated in countries are based out of countries of the treatment group. The first stage estimates of the endogenous treatment model imply that a lower level of anti-director’s rights increases the location choice of the acquiring company. To further optimize this research, one could potentially exploit the separating the effect of the transparency per deals and the relocation effect on bidder’s returns.

After taking self-selection into account, I find that the ETD led to an increase of 5,02 percentage points in acquirer’s returns, which corresponds to $69 million-dollar value on average. These results are in contrast with findings from (Humphery-Jenner, 2012). The author advocates that the implementation of the ETD lead to vagueness around legal shareholder rights and enhancing managerial entrenchement, therefore leading to more value destroying acquisitions after the implementation of the ETD. The variance between the findings of the difference-in-difference model and the endogenous treatment model are caused due to the fact that there are variables correlated with the dependent variable, acquirer returns, and the explanatory variable, improved legal shareholder rights. This study controls the correllating unobserved variables determing the acquirer’s returns in the first stage, by seperating out the effect of that the target’s country initial anti-director rights.

6. Conclusion

This paper tries to help resolve the question around the influence of country level legal shareholder rights on takeover efficiency. The implementation of the European Takeover Director served as a natural experiment to construct the research to examine the effect of new imposed regulation regarding takeover law and the allocation of assets. Due to substantial variations between countries corporate governance systems and takeover activity, I could exploit a difference-in-difference model to examine my hypothesis. I find weak significant evidence to support the fact that the improvement of country level legal shareholder rights increases the returns of acquiring companies. This research corrects for sample selection issues, by using a derivation of Heckman’s two- stage correction model. The results imply that the target’s country level of anti-director rights have an influence on the acquisition location of the acquiring company. Several issues remain unsolved, it is of importance to better understand and separate the effect on the acquirer’s returns between location of the target and the increased transparency generated by the implementation of the ETD.

(31)

31

References

Agrawal, A. (2013). The impact of investor protection law on corporate policy and performance: Evidence from the blue sky laws. Journal of Financial Economics , 417-435.

Alexandridis, G., Fuller, K., Terhaar, L., & Travlos, N. (2013). Deal size, acquisition premia and shareholder gains. Journal of Corporate Finance, 1-13.

Bebchuk, L. C. (2005). The costst of entrenched boards. Journal of Financial Economics, 78, 409-433. Bergman, N., & Nicolaievsky, D. (2007). Investor protection and the Coasian view. Journal of

Financial Economics, 84, 738-771.

C. Loderer, K. M. (1990). Corporate acquisitions by listed firms: the experience of a comprehensive sample. Journal of Financial Management, 17-33.

Clougherty, J. D. (2016). Correcting for Self-Selection Based Endogeneity in Management Research: Review, Recommendations and Simulations. Organizational Research Methods, 19(2), 286-347.

Commission, E. (2012). Report from the commission to the European Parliament, the council, the

European economic and social committee and the committee of the regions: application of the directive. Brussels.

de Jong, F. d. (2011). Event Studies Methodology. Tilburg: Tilburg University.

Doidge, C., Andrewkarolyi, G., & Stulz, R. (2007). Why do countries matter so much for corporate governance? Journal of Financial Economics, 86, 1-39.

Harford, J. H.-J. (2012). The sources of value destruction in acquisitions. Journal of Financial

Economics, 247-261.

Heckman, J. (1976). The common structure of statistical models of truncation, sample selection and limited dependent and a simple estimator for such models. Annals of Economic and Social

Measurement, 475-492.

Humphery-Jenner, M. (2012). The impact of the EU takeover directive on takeover performance.

Journal of Corporate Finance, 18, 254-272.

Humphery-Jenner, M. P. (2011). Firm size, takeover profitability, and the effectiveness of the market for corporate control: Does the absence of anti-takeover provisions make a difference?

Journal of Corporate Finance, 17(3), 418-437.

Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. American

Economic Review, 323–329.

Jensen, M. M. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 4(3), 305-360.

La Porta, R., Lopez‐de‐Silanes, F., & Shleifer, A. (2006). What works in securities laws? Journal of

Finance, 1-32.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. (1998). Law and Finance. Journal of

Do shareholder rights add value to M&As? : evidence from the European Takeover Directive