The impact of fairness opinions on bid premiums

Thesis

MSc Business Administration

Specialisation: Finance

University of Groningen

Faculty of Economics and Business

Author: Jeroen Linthorst Student number: 1542672 Supervisor: prof. dr. B.P. de Bruin

JEROEN LINTHORST*

ABSTRACT

This study examines the influence of fairness opinions on bid premiums. Moreover, the impact of independent as opposed to dependent financial advisors is analyzed for Austria, Belgium, France, Germany, the Netherlands and Switzerland between 2004 and 2010. Although univariate analysis shows a higher bid premium in deals with a fairness opinion, regression analyses show that fairness opinions used by target companies do not significantly influence the bid premium. Additionally, the independence of the financial advisor issuing these fairness opinions does not affect the bid premium, although univariate analysis shows a negative relation. Thus, the conflict of interest between a target and a dependent financial advisor does not seem to have an impact. Target boards could use an independent fairness opinion in deals with a low bid premium, to convince shareholders that they will receive a fair price.

JEL classification: G24, G34, G38

Key words: fairness opinions, bid premium, mergers and acquisitions, conflict of interest, regulations, corporate governance

*

1. Introduction

The objective of this study is to empirically examine the influence of a fairness opinion used by a target company on the bid premium in a public M&A deal in Continental Europe. Additionally, the influence of the independence of financial advisors issuing fairness opinions on this relation is studied. The bid premium is defined as the percentage difference between the offer price and target share price one month prior to the announcement date. A fairness opinion is a third-party assessment about the fairness of the terms of the deal, from a financial point of view. They are most often issued by investment banks, but they can also be issued by consulting or accountancy firms. A fairness opinion is a letter stating whether the proposed price is in the range of fair values calculated by the financial advisor and can be addressed to either the management or the supervisory board or to both. In most cases a fairness opinion is issued after a transaction has been structured (bid price and type of payment is determined etc.), but before it has been approved by the shareholders (Caddell, 1997).

A fairness opinion can signal a good deal for investors and shareholders. When a deal is negotiated, this signal can influence the premium that is paid for the target company. This can be the case when the proposed bid price does not fall within the fair values calculated by the fairness opinion provider and therefore the price is renegotiated, so that the price does fall within this range. However, a dependent financial advisor with a fee contingent on completion of the deal can have an incentive to “rubber stamp” a deal (Bebchuk and Kahan, 1989), which can lead to a lower bid premium. Moreover, corporate governance, like ownership structures and investor protection, can influence both the use of fairness opinions and the bid premium. It is expected that these two are different in Continental Europe than in the U.S., which was the main focus of previous research. The problem statement of this study is as follows:

Does a target fairness opinion lead to a higher bid premium in a public M&A deal in Continental Europe? Moreover, is the bid premium higher if the fairness opinion is issued by an independent instead of a dependent financial advisor?

companies. When the usage of a fairness opinion leads to a higher bid premium, they can decide to use one in future M&A deals. Moreover, when an independent fairness opinion leads to a higher premium, the board can decide to only ask an independent financial advisor to issue a fairness opinion.

This study only looks at fairness opinions used by target companies, since only three fairness opinions used by acquiring companies are found in the sample deals. The sample consists of 340 public M&A deals in Austria, Belgium, France, Germany, the Netherlands and Switzerland in the period between January 1, 2004 and December 31, 2010. Thus far, this period has not been the subject of other research involving fairness opinions. Moreover, the influence of corporate governance on the bid premium and regulations involving fairness opinions are new contributions to the literature.

Another contribution to the literature is that this is the first study that looks at the influence of fairness opinions on bid premiums in Continental Europe instead of the U.S. or Australia. It is expected that differences exist between these geographic areas. One difference lies in different shareholder protection. Continental European countries have code law, while Anglo-Saxon countries like the U.S. have a common law system (La Porta, Lopez-de-Silanes, Shleifer, and Vishny, 1998). Code law countries are characterized by low shareholder protection, while common law countries have high shareholder protection (La Porta, Lopez-de-Silanes, Shleifer, and Vishny, 2000). The outcome hypothesis predicts higher premiums in countries with higher investor protection, while the efficiency hypothesis predicts that these are higher in countries with lower investor protection (Rossi and Volpin, 2004).

The EU Directive 2004/25/EG about public offers can also influence the bid premium. It is aimed at improving the protection of minority shareholders. The EU member states implemented this directive in laws, with some countries regulating the use of fairness opinions. Given that the studies on investor protection are performed before this directive was issued, the difference between code law and common law countries could be reduced afterwards. Therefore, it is expected that the bid premium in this study is lower than in the previous studies that looked at common law countries.

independent if, first, he or she is not involved in other aspects of a deal besides issuing a fairness opinion and, second, receives a fixed fee upon providing this fairness opinion.

The results of this study show that target fairness opinions and the independence of fairness opinions providers do not significantly influence the bid premium. Although univariate analyses show that the bid premium is higher in deals with a target fairness opinion than without a target fairness opinion, and lower in deals with an independent target fairness opinion than with a dependent target fairness opinion, this result is not confirmed by the regression analyses. Return on equity and market-to-book value of the target have a significantly negative impact on the bid premium. A deal that is cross-border and the percentage of stock of the target company that is acquired influences the bid premium significantly positive. A possible explanation for the univariate results can be that target boards ask for an independent fairness opinion in deals where a lower premium is offered, to convince shareholders that they will receive a fair price.

2. Literature review

This section elaborates on the theoretical background of the relation between fairness opinions and the bid premium. First, the conflict of interest between managers and shareholders that might exist is examined. This is followed by the possible conflict of interest between financial advisors and company shareholders. After that, the corporate governance structure in Continental Europe and the relation to fairness opinions and the bid premium is explained. Moreover, the influence of regulations in the sample countries are discussed. This section ends with an overview of the existing empirical literature on the relation between fairness opinions and the bid premium.

2.1 Conflict of interest between managers and shareholders

An agency problem can exist between managers and shareholders of a company (Jensen and Meckling, 1976). In the case of a corporate control transaction, managers of a target company could favor a deal in which they personally gain, but that is bad for shareholders. DeAngelo (1990) argues that this agency problem can be resolved by fairness opinions: it can signal to the shareholders that management acted in the best interest of the shareholders. Information asymmetry between managers and shareholders can also motivate the use of fairness opinions (DeAngelo, 1990). Share prices do not reflect the inside information of managers due to information asymmetry. Therefore, shareholders cannot rely on share prices to assess the value of their shares. An independent fairness opinion that uses this inside information in its valuation can diminish this information asymmetry.

2.2 Conflict of interest between financial advisors and company shareholders

Financial advisors that issue a fairness opinion can have conflicts of interest with the company shareholders, when they are not independent. A possible conflict of interest can arise when a financial advisor is also involved in other aspects of the deal like negotiations, besides providing a fairness opinion. The financial advisor may have a strong incentive to give a ‘fair‘ opinion when they negotiated the price, to prevent harming its credibility (Parijs, 2005).

what critics call “rubber stamp” a deal (Bebchuk and Kahan, 1989). He can present a low ‘fair’ price, when the contingent fee is a linear function of the deal value (“total-value fee”) (Mclaughlin, 1990). This would lead to a lower price and thus premium. A financial advisor can try to raise the price by presenting a high ‘fair’ price, when the contingent fee is paid when the price is above a specific amount (“incremental fee”) (Mclaughlin, 1990). However, a financial advisor will only try to raise the price, if the chance of ‘killing the deal’ is low (Bebchuk and Kahan, 1989).

There are several reasons why these potential problems with financial advisors may not lead to a conflict of interest (Bebchuk and Kahan, 1989). First, a financial advisor has “reputational capital”. A financial advisor can lose its reputation of being a top-tier financial advisor when they issue a ‘fair’ opinion on a financial inadequate transaction. Consequently, since Rau (2000) found a positive relation between investment banker market share and their reputation, the market share of a banker will decline if he loses its reputation. However, Cleveland (2006) argues that there are several reasons why the reputation of the financial advisors will not suffer if they issue an inadequate fairness opinion. First, the information market for detecting low quality fairness opinion is inefficient, because very little is disclosed about the underlying analyses and there is no consensus about what a ‘fair’ price is. Second, financial advisors often cannot be punished by a regulation agency, since in many countries nothing is prescribed about the content of a fairness opinion. Thus, the argument of investment banks stating that their reputation would lead them to issue unbiased fairness opinions can be questioned. Moreover, although investment bankers might be willing to preserve the reputation of the bank, they are subject to cognitive biases that can lead to irrational behavior (Cleveland, 2006). For instance, when a financial advisor is overconfident in his fairness opinion, he can underestimate the possible risk on his reputation.

underestimating the negative effect of issuing a ‘fair’ opinion on an inadequate deal) and the self-serving bias (predictions of the fair value are biased towards what is in the interest of the banker) (Cleveland, 2006). Also, investment banks claim that their contingent fee aligns the interests of the shareholders and advisors instead of causes conflicts of interest. Finally, the financial advisor will not receive the additional pay, if a financial advisor that receives a contingent fee issues a ‘fair’ opinion for a bad deal and the transaction is not consummated (Calomiris and Hitscherich, 2005).

2.3 Corporate governance

Anglo-Saxon countries with a common law system like the U.S. are characterized by widely held companies and high shareholder protection (La Porta, Lopez-de-Silanes, Shleifer, and Vishny, 2000). In contrast, European companies with code law often have a majority shareholder like a family, and thus lower shareholder protection. This controlling shareholder frequently has control without having equal cash flow rights, by using pyramidal ownership, shareholder agreements, and dual classes of shares (La Porta, Lopez-de-Silanes, and Shleifer, 1999). The interests of minority shareholders and a controlling shareholder may not be aligned. Fairness opinions can reduce this agency problem by delivering an independent valuation of the company.

The bid premium can also be affected by the different corporate governance systems in the U.S. and Continental Europe. First, the outcome hypothesis predicts that the bid premium will be higher in countries with higher investor protection (Rossi and Volpin, 2004). Ownership is concentrated (Bebchuk, 1999) and the private benefits of control are large (Nenova, 2002; Dyck and Zingales, 2002) in countries with low investor protection, which lead to an imperfect market for corporate control. On the other hand, ownership is dispersed and private benefits of control are small in countries with high investor protection, which lead to a lively corporate control market (Manne, 1965; Jensen, 1993). This competition will drive the premium up. Conversely, the efficiency hypothesis states that the bid premium will be lower in countries with higher investor protection. La Porta, Lopez-de-Silanes, Shleifer, and Vishny (2002) and Bhattacharya and Daouk (2002) find more inefficient companies (e.g. lower valuations and higher cost of capital) in countries with lower investor protection. Change of control in these companies trough M&A can create value.

2.4 Regulation

protect the rights of minority shareholders. This directive includes a definition of the price that must be offered: it has to be an “equitable price”. This is defined as: “the highest price paid for the same securities by the offeror, or by persons acting in concert with him, over a period, to be determined by the Member States, of not less than six months and not more than twelve before the bid. If, after the bid has been made public and before the offer closes for acceptance, the offeror or any person acting in concert with him purchases securities at a price higher than the offer price, the offeror must increase his offer so that it is not less than the highest price paid for the securities so acquired.” Hence, no mention is made about a fairness opinion, but some countries use this guideline to regulate the use of it.

In Austria the board of directors of the target company have to appoint an independent expert that evaluates the bid and the report of the board that comments on the bid. In addition, the complete fairness opinion must be disclosed by the target company.

In Belgium an independent expert must evaluate the bid price for the target shareholders in the case of a public offer of a controlling shareholder or in the case of a squeeze-out. The report has to be disclosed to the public. The expert is defined as independent when he is not involved in other ways with the deal. Some companies cannot act as independent advisor, like the accountant of the target or acquirer, since this could lead to a conflict of interest. Moreover, the financial advisor is not allowed to receive a fee that is contingent on completion of the deal.

Since 2006, France has even stricter regulation in place about fairness opinions. These regulations state that a fairness opinion is mandatory in deals where a potential conflict of interest could arise, like in a merger, when a controlling shareholder makes an offer for the outstanding shares or in a squeeze-out. Also, the regulations deal with the content of the opinion and the independence of financial advisors. A fairness opinion provider must provide a statement of independence, which holds that they cannot be involved in the deal other than providing the fairness opinion and that they cannot have been an advisor to the acquirer in the previous two years. In addition, they have to be paid a fixed fee, of which the amount also has to be disclosed. Next to that, the financial advisors have to monitor the quality of the fairness opinion by creating committees and internal procedures to comply with regulations.

In the Netherlands it is not mandatory to get a fairness opinion for acquirers or targets. Since 2007 however, a fairness opinion has to be disclosed to the shareholders if an acquirer or target got one.

Finally, in Switzerland it is not mandatory to use a fairness opinion, but when it is used it also has to be disclosed to the public. The issuer of the fairness opinion has to be ‘suitably qualified an independent of the offeror, the offeree company and persons acting in concern with them'.

2.5 Empirical literature

2.5.1 Relation between bid premium and fairness opinion

Few empirical literature has focused on the relation between bid premiums and fairness opinions. Bugeja (2007) finds no significant relationship between the use of a fairness opinion and the bid premium. His sample consists of 397 deals for publicly listed Australian companies in the period between 1990 and 2000. The 190 deals where fairness opinions were issued due to regulation are deleted from his sample. Bugeja finds a mean bid premium of 31.28%, which is defined as the percentage of the offer price over the target stock price two months prior to the announcement of the deal. Using logit regression models, the frequency of using a fairness opinion is related to several variables, including the bid premium. Although the coefficient for the bid premium is negative, the effect is not significant.

Two other studies investigate the United States. Makhija and Narayanan (2007) find that in about 53% of deals involving listed companies a target company uses a fairness opinion. This percentage is concluded after consulting the Securities Data Company’s (SDC) M&A database. They examine 1,927 deals in the period 1980 to 2004. They exclude 4,561 out of the total 6,488 deals where no information about advisors and financials are available, or when the deal is private, not completed, a minority acquisition (where the acquirer does not hold a majority after the deal) or is a ‘cleanup offer’ (where the acquirer holds more than 50% of the target shares prior to the deal). The definition of the bid premium in their study is expressed as the percentage of the offer price over the target stock price four weeks prior to the announcement of the deal. They find a mean (median) bid premium of 50.85% (40.67%). Univariate tests indicate no difference in the bid premium between deals with a target fairness opinion and deals without one. However, when a regression analysis is used to control for company and deal characteristics the bid premium is significantly 8% higher when a target fairness opinion is used.

opinions. However, online SEC files are used to check the results of the SDC database. In one-third of the deals the SEC files identified the use of a fairness opinion, while this was not recorded in the SDC database. Thus, the sample of Makhija and Narayanan (2007) and Kisgen, Qian and Song (2009) substantially differs from each other. This might also explain why the latter authors in an ordinary least squares (OLS) regression framework find no difference in the bid premium when a target fairness opinion was used.

Overall, studies in the US and Australia show mixed results about the relation between the bid premium and fairness opinion. Makhija and Narayanan (2007) report a positive relationship, while Bugeja (2007) and Kisgen, Qian and Song (2009) find no significant effect. This difference can be attributed to different sample periods, type of deals and method of data collection.

2.5.2 Relation between bid premium and independent fairness opinion

Four studies also take the effect of advisor independence into account. They define ‘independence’ in different manners. The criterion used by Kisgen, Qian and Song (2009) to distinguish dependent (“affiliated”) and independent (“unaffiliated”) fairness opinion providers is whether they are affiliated with the advisor group that is involved in other aspects of the deal. Makhija and Narayanan (2007) use a similar definition. Moreover, the two studies also look at the influence of fees that are contingent on deal completion. However, they do not state this as a necessary condition for independence. Chen (2010), who investigates the impact on announcement returns and operating performance, states that an advisor is independent if he or she only issues a fairness opinion and receives a fixed fee upon providing this fairness opinion. The strict definition of Chen (2010) is used in this research because a financial advisor that receives a fixed fee, but also performs other advisory services (with fees that are contingent on completion of the deal), will be more prone to issue a ‘fair’ opinion. This will be the case in practice, because they will lose the part of the fees that is contingent on completion when they do not give a ‘fair’ opinion. This definition is the strictest definition of independence, which would lead to the most unbiased opinion about the fairness of the deal. This is in the best interest of company shareholders who want to receive a fair price for their shares.

Calomiris and Hitscherich (2005) investigated independence by looking at the effect of the proportion of fixed fees on deal premiums. A sample of 170 cash deals with a value of at least $100 million, which were announced between 1994 and 2002 in the US, is used. Moreover, the sample only included friendly, two-step transactions (an acquisition of shares in the first step and in the second step the companies merge without needing the approval of the shareholders) with a public target. The SDC database is used to identify deals. A mean bid premium of 54.60% is found in these deals. Calomiris and Hitscherich report no significant relation between the proportion of fixed fees and acquisition premiums in a regression framework. In addition, previous business relationships of targets do not influence the premium.

Makhija and Narayanan (2007) find that merger premiums are lower when a dependent target financial advisor issues the fairness opinion instead of an independent financial advisor. In addition, this result also holds when both targets and acquirers use a fairness opinion in a transaction. This provides some information about the effect of fairness opinions on deal premiums, but Makhija and Narayanan (2007) stress the fact that this not automatically means that targets lose and acquirers win when using a dependent fairness opinion. For instance, mergers premiums and fairness opinions are jointly determined and endogenous. This holds that, for instance, a fairness opinion is asked because the premium in a deal is low and the target board wants to go ahead with the financially inadequate transaction. A fairness opinion that states that the price that shareholders receive is fair, can lead to consummation of the deal although it would harm the shareholders of the target firm.

Kisgen, Qian and Song (2009) report no significant effects of target fairness opinions on bid premiums. Concluding the previous research on the impact of independent fairness opinions on the bid premium, ambiguous results are reported. Three studies find no significant effect, while Makhija and Narayanan (2007) find a positive relationship. One possible explanation of these different results is that Eddey (1993) and Calomiris and Hitscherich (2005) exclude non-cash deals, since previous research has reported that target shareholders receive significantly higher premiums when cash is the method of payment (Waynsley et al., 1983; Huang and Walkling, 1987; Franks et al., 1991; Da Silva Rosa et al., 2000). Moreover, Makhija and Narayanan (2007) report that fairness opinions are used less frequently in cash offers.

Table 1: overview of previous empirical literature on the relation between bid premium and target (independent) fairness opinions

The presented signs for the relation between the bid premium and (independent) fairness opinion are from the author’s study. A + means that a positive relation is found between the use of a fairness opinion and the bid premium, N.S. means that no significant relation is found and n.a. means that this relation is not examined in the study. Also, it is mentioned whether the author gives a definition of the independence of the financial advisor used in his study.

Authors Domain Period # Deals Mean (median)_{bid premium %} Independencedefinition available

Relation bid premium and fairness opinion

Relation bid premium and independent fairness opinion

Eddey (1993) Australia 1988 - 1991 364 39.40 (28.74) Yes n.a. N.S.

Bugeja (2007) Australia 1990 - 2000 397 32.28 No N.S. n.a.

Calomiris and Hitscherich (2007) U.S. 1994 - 2002 170 54.60 No n.a. N.S.

Makhija and Narayanan (2007) U.S. 1980 - 2004 1,927 50.85 (40.67) Yes + +

3. Hypotheses and methodology

The theoretical literature and previous empirical results lead to the following hypothesis about the relation between target fairness opinions and the bid premium:

H1: The bid premium is higher when a target fairness opinion is used in a deal than when no target fairness opinion is used.

The hypothesis about the independence of financial advisors issuing a target fairness opinion and its impact on the bid premium is as follows:

H2: The bid premium is higher when an independent target fairness opinion is used in a deal than when no independent target fairness opinion is used.

3.2 Methodology 3.2.1 Univariate analysis

As univariate analysis the parametric two-sample t-test (Newbold et al., 2003) and the non-parametric Mann-Whitney U test (Mann and Whitney, 1947) are used to investigate whether there are differences between variables in subsamples of the dataset. The t-test is used to test for the difference in means, while the Mann-Whitney U test tests the medians. This test is used next to a t-test, because the Mann-Whitney U test does not require the variables to be normally distributed (Newbold et al., 2003).

3.2.2 Multivariate analysis

In order to perform an OLS regression the following assumptions have to be fulfilled (Brooks, 2008).

1. The errors have zero mean;

2. The variance of the errors is constant and finite over all values of ݔ௧; 3. The errors are linearly independent of one another;

4. There is no relationship between the error and corresponding ݔ variate; 5. ݑ௧is normally distributed.

The first assumption of zero mean errors is met, since a constant is included in every regression equation (Brooks, 2008). The second assumption is dealt with by correcting the standard errors for heteroscedasticity using White’s heteroscedasticity consistent standard errors (White, 1980). When heteroscedasticity is present, the variance of the coefficients, and therefore the standard errors, in a regression can be biased (Brooks, 2008). This can lead to wrong rejection or acceptation of hypotheses. However, the coefficients estimated will be unbiased. The third assumption can be checked using a Durbin-Watson test for autocorrelation. No problem with this assumption was found for the regressions in this study. If the correlation of the fourth assumption is present, the OLS estimator will not be consistent, because explanatory power is assigned to the variables, while it is actually is because of the correlation between the error term and ݕ௧(Brooks, 2008). The final assumption is that the error terms are normally distributed. However, because of the relative large sample size the central limit theorem suggests this is not a problem (Brooks, 2008).

To test the hypotheses on the impact of (independent) target fairness opinions on the bid premium, the methodology of Makhija and Narayanan (2007) is followed and bid premiums are regressed on two fairness opinion variables and several control variables:

PR = α + β1FO + β2FO ind + β3Log(Deal size) + β4ĞĂůĂƫ ƚƵĚĞĨƌŝĞŶĚůLJ + β5Method of payment stock +

β6Cross-border + β7Log(ROE Target) + β8Log ൬D_ETarget൰ + β9Log ൬M_BTarget൰ + β10Percent acquired +

β112004 + β122005 + β132006 + β142007 + β152008 + β162009 + ϵ (1)

Chidambaran et al., 2010; Alexandridis, Mavrovitis, and Travlos, 2011). These fixed year effects appear as dummy variables in the regression, with 2010 as the omitted variable.

3.2.2.1 Dependent variable

The bid premium is the dependent variable in the regression analyses, and is defined as the percentage premium of the offer price over the target price one month prior to deal announcement. Most previous literature defines the bid premium as the percentage difference between the offer price and target share price four weeks prior to the announcement date (e.g. Makhija and Narayanan, 2007; Kisgen, Qian and Song, 2009), but a month instead of four weeks is used because of data limitations. It is not expected that this slightly altered definition gives substantially different results. As a robustness check, alternative definitions of the bid premium are also examined.

3.2.2.2 Independent variables

The main independent variables in the OLS regression are a dummy stating if a target fairness opinion is found for the deal, and a dummy stating if an independent target fairness opinion is found.

sign in regression analyses for the target debt-to-equity ratio. D/E Target is added as the next control variable. Bid premiums are lower when the target market-to-book ratio is high, according to Makhija and Narayanan (2007) and Kisgen, Qian and Song (2009). The control variable M/B Target is expected to have a negative sign. Finally, the bid premium can be higher when a transaction involves the purchase of a controlling stake in the target company. This so called control premium is on average 14% (Dyck and Zingales, 2004). Percent acquired is the next control variable.

The variable M/B Acquirer is deleted from the regressions, because this variable has many missing values and thus would cause the regression to omit a lot of data. Moreover, the variables Deal size, ROE Target, D/E Target and M/B Target are transformed into a log variable with base 10. This is done because the original variables were not normally distributed. After the log transformation they are more normally distributed. The variables mentioned above are presented in Table 2 and show their predicted sign for this current research.

Table 2: predicted signs of control variables influencing the bid premium

Control variable Predicted sign

Deal size

-Deal attitude friendly

-Method of payment stock

4. Data

Data on M&A deals and financial information of companies involved in these deals are retrieved using Capital IQ. This is a database comprising data on more than 60,000 public companies and worldwide transactions. The deals included in this research satisfy the following selection criteria:

1. The deal is a tender offer; 2. The deal is closed;

3. The deal size is at least 5 million Euros;

4. The deal was announced between January 1, 2004 and December 31, 2010; 5. The target is a public company;

6. The target company is based in Austria, Belgium, France, Germany, the Netherlands or Switzerland.

4.2 Descriptive statistics 4.2.1 Sample distribution

Table 3 below presents an overview of the deals in the sample across years and countries. The total number of deals is 340, with most deals from the countries France (115) and Germany (96) and the years 2006 (58), 2007 (81) and 2008 (64). The lowest number of deals is observed in 2010. This can be caused by lower M&A market intensity, but also by the fact that the deals that were announced in that year are not closed at the time of this research.

Table 3: Number of tender offers in sample

The sample consists of closed tender offers of at least 5 million Euros on public targets announced between 2004 and 2010 in Austria, Belgium, France, Germany, the Netherlands and Switzerland.

2004 2005 2006 2007 2008 2009 2010 Austria 2 2 1 1 2 3 2 13 Belgium 2 5 4 7 4 4 1 27 France 9 16 21 31 21 9 8 115 Germany 12 4 13 20 26 14 7 96 Netherlands 6 4 8 13 5 6 4 46 Switzerland 2 7 11 9 6 6 2 43 33 38 58 81 64 42 24 340

Table 4: Number of tender offers where a fairness opinion is found

The sample consists of closed tender offers of at least 5 million Euros on public targets announced between 2004 and 2010 in Austria, Belgium, France, Germany, the Netherlands and Switzerland. The percentages denote the percentage of total tender offers where a fairness opinion is found in that year or country.

2004 2005 2006 2007 2008 2009 2010 Austria 2 100% 2 100% 0 0% 1 100% 1 50% 3 100% 2 100% 11 85% Belgium 0 0% 4 80% 1 25% 1 14% 2 50% 2 50% 0 0% 10 37% France 2 22% 7 44% 8 38% 26 84% 12 57% 6 67% 8 100% 69 60% Germany 3 25% 1 25% 2 15% 12 60% 12 46% 6 43% 3 43% 39 41% Netherlands 5 83% 3 75% 3 38% 7 54% 4 80% 4 67% 4 100% 30 65% Switzerland 0 0% 6 86% 7 64% 6 67% 6 100% 5 83% 2 100% 32 74% 12 36% 23 61% 21 36% 53 65% 37 58% 26 62% 19 79% 191 56% The target fairness opinions can be divided in independent and dependent fairness opinions. Table 5 hereafter summarizes the deals where an independent fairness opinion is found. In an average of 37% of the deals an independent fairness opinion is found. The highest percentage can be seen in Austria (85% of the deals), where all fairness opinions are from independent advisors. In both Switzerland and France, the independence of the financial advisor cannot be confirmed with only one fairness opinion. These high percentages can be explained by the strict regulations on fairness opinions in these countries. What also stands out is that there are more independent than dependent fairness opinions in the sample (126 versus 66, respectively). This is largely caused by Austria, France and Switzerland where almost every fairness opinion is issued by an independent financial advisor.

Table 5: Number of tender offers where an independent fairness opinion is found

The sample consists of closed tender offers of at least 5 million Euros on public targets announced between 2004 and 2010 in Austria, Belgium, France, Germany, the Netherlands and Switzerland. The percentages denote the percentage of total tender offers where an independent fairness opinion is found in that year or country.

2004 2005 2006 2007 2008 2009 2010 Austria 2 100% 2 100% 0 0% 1 100% 1 50% 3 100% 2 100% 11 85% Belgium 0 0% 1 20% 1 25% 0 0% 1 25% 1 25% 0 0% 4 15% France 2 22% 7 44% 7 33% 26 84% 12 57% 6 67% 8 100% 68 59% Germany 0 0% 0 0% 1 8% 0 0% 0 0% 2 14% 2 29% 5 5% Netherlands 1 17% 0 0% 1 13% 3 23% 0 0% 1 17% 1 25% 7 15% Switzerland 0 0% 6 86% 7 64% 6 67% 6 100% 4 67% 2 100% 31 72% 5 15% 16 42% 17 29% 36 44% 20 31% 17 40% 15 63% 126 37% 4.2.2 Dependent and independent variables

from 9.04% to 25.58% in the different sample countries. It is the highest for Austria (25.58%), followed by Belgium (24.50%) and the Netherlands (20.36%) and the lowest for France (9.04%). As expected, the mean and median bid premium is lower than in previous studies that focus on the U.S. and Australia.

Figure 1 shows a plot of the mean and median bid premium across years. There clearly is a pattern observable, with a decline from 2005 to 2007 and subsequently a sharp increase until 2008, after which it rises again. The sharp increase after 2007 could be caused by the credit crunch or by regulations. Due to the credit crunch, and therefore the drop in share prices, managers felt that the value of their company was not reflected in the stock price and therefore demanded a higher bid premium. However, at the same time regulations following the EU directive that are implemented can influence the bid premium.

Figure 1: mean and median bid premium of the total sample across years

We see from Table B.1 that the average deal size is 1.2 billion Euros, with a standard deviation of 5.45 billion. The median deal size is only 120 million Euros. This skewness is causes by a few very high outliers, for instance the deal between ABN AMRO and the consortium of RBS, Santander and Fortis of almost 72 billion Euros. Figure 2 graphically points out that the mean deal size is much more volatile the median deal size.

Figure 2: mean and median deal size in € million across years

The growth options are somewhat larger for acquirers than target companies, as indicated by the higher median market-to-book value for acquirers (1.91 versus 1.77). Almost every deal is perceived as friendly, with a percentage of 98.2%. Cash is the preferred method of payment, since in only 7.4% stock is used for payment of the deal. Moreover, most deals involve the acquisition of less than 50% of stocks (181 deals out of the total of 340, or 53%). This can be the acquisition of a minority share in a company or a majority shareholder buying additional shares of the target company.

Next, the total sample is divided in two subsamples: one consisting of the deals where a target fairness opinion is found and the second consists of the deals where no target fairness opinion is found. The descriptive statistics of these subsamples are presented in Table B.2 and B.3. The median bid premium is higher for deals where a fairness opinion is found (16.64% versus 13.98%). Also the median deal size is larger in deals when a fairness opinion is found (192.10 million versus 161.63 million). Another noticeable difference between the subsamples is that the market-to-book value of the median target company is lower when a fairness opinion is found (1.64 versus 2.01). This holds that firms with less growth opportunities ask for a fairness opinion when being acquired more often than companies with more growth opportunities.

Lastly, the subsample of deals where fairness opinions are found are further divided in subsamples where an independent fairness opinion is found and one where a dependent fairness opinion is found. In Table B.4 and B.5 are the descriptive statistics of these subsamples presented. The median bid premium is lower in deals with an independent fairness opinion than in deals with

a dependent fairness opinion (13.03% versus 24.50%). In addition, the median deal size is lower in these deals (152.88 million versus 352.12 million). The distribution of the percentage of acquired stock is more evenly spread in the subsample with deals with a dependent fairness opinion than in the subsample of independent fairness opinions. In the latter case 55% of the times the deal concerns less than 50% of the shares, while this percentage is only 29 in the former case.

4.3 Correlations

When some independent variables are highly correlated, multicollinearity can cause the coefficient estimates in a regression to be biased (Brooks, 2008). The correlation matrix for the independent variables used in this research is presented in appendix C. What stands out is that the variable Percent Acquired is significantly correlated to almost every variable, ranging from -0.264 to 0.208. Since these are not very high correlations, this variable can be left in the analysis.

4.4 Overview

Table 6 provides a summary of the mean and median bid premiums in the total sample and subsamples.

Table 6: summary of mean and median bid premiums

Sample Mean Median

Total sample 23.85 14.71

Fairness opinion 25.48 16.64

No fairness opinion 22.00 12.50

Independent fairness opinion 21.41 13.03

Dependent fairness opinion 33.37 24.50

5. Results

5.1.1 Bid premium

The results for the univariate analysis of subsamples are presented in Table 7. The mean and median bid premium are significantly higher in deals where a target fairness opinion is found at the 10% and 5% level, respectively. The higher bid premium when a fairness opinion is found contradicts previous literature where univariate results are available (e.g. Kisgen, Qian and Song, 2009; Makhija and Narayanan, 2007), which find no difference in bid premiums between deals where a target fairness opinion is used and where it is not used. These different results can be caused by the fact that these studies focus on the U.S. instead of Continental Europe. Moreover, their sample period end in 2003 and 2004, while this current research is the first to investigate 2004 to 2010.

Table 7: Results for univariate analysis of relation between bid premium and (independent) fairness opinions

Mean Median Test for difference between groups T-statistic U-statistic

Fairness opinion 25.48 16.64 _{-1.666 * 12,508.5 **}

No fairness opinion 22.00 12.50

Independent fairness opinion 21.41 13.03

2.294 ** 3,274.5 **

Dependent fairness opinion 33.37 24.50

***. Significant at the 0.01 level (2-tailed). **. Significant at the 0.05 level (2-tailed). *. Significant at the 0.10 level (2-tailed).

5.1.2 Control variables

Table D.1 in Appendix D shows the univariate test results for the control variables used in this study. Panel A displays the results for fairness opinions versus no fairness opinions. The mean and median M/B Target are significantly lower in deals with a fairness opinion. The mean and median Percent acquired are significantly higher when a fairness opinion is used. Moreover, the median Deal size is higher in deals with a fairness opinion. Makhija and Narayanan (2007) find among other things that fairness opinions are used less when the deal attitude is hostile, the method of payment is cash or have few growth options (a low market-to-book value). These variables are not found to be significantly different between the ‘fairness opinions found’ and the ‘no fairness opinions found’ subsample. It has to be noted that Makhija and Narayanan (2007) only test for differences in means. Panel B shows that the mean and median Deal size and Percent acquired are significantly lower in deals with an independent fairness opinion than a dependent fairness opinion.

5.2 Multivariate analysis

Table 8 on page 27 presents the results of the regression analyses with White heteroskedasticity-consistent standard errors. Every model is checked for multicollinearity, autocorrelation and non-normality of the residual errors. No model was subject to these potential specification problems, except from the residuals being non-normal.

First, a regression with only the variables FO and FO ind as the independent variables is run. The use of a fairness opinion has a positive effect, while an independent fairness opinion negatively influences the bid premium. The model itself is significant at the 5% level as indicated by the F-statistic, but the explanatory power is low (adjusted R² = 0.015). Thus, this model is not very good in explaining the variability of the bid premium and another specification of the model is needed.

opinion leads to an 8% higher premium. Also the use of an independent fairness opinion is not a significant determinant of the bid premium. This is in line with the results of Calomiris and Hitscherich (2007) and Kisgen, Qian and Song (2009), but in contrast to Makhija and Narayanan (2007), who find a positive relation.

In model 3 and 4, insignificant control variables are omitted from the regression. Results of model 4 are discussed here, since this model is the most significant as indicated by the F statistic. This model has an adjusted R² of 0.204 and has an F statistic of 7.632. The results for the fairness opinion variables are consistent with model 2, so the model is robust for omitting insignificant variables. Log (ROE Target) is negatively related with the bid premium. An increase of 1% in return on equity of the target leads to a 12.11%1_{lower premium for this company. This is in line with}

Kisgen, Qian and Song (2009), who indicate a negative influence of -4.5% of this variable in tender offers. The Percent acquired by the acquirer is positively related to the bid premium. When this percentage is one higher, the bid premium goes up with 0.28%. Thus, the literature on the control premium is confirmed by the data in this research. Log (M/B Target) has a negative relation with the bid premium. The bid premium is 8.01% lower when the market-to-book value of the target is 1% higher. Also the variable Cross-border is significant in explaining the bid premium. When a deal goes cross-border instead of being domestic, the bid premium of the deal is 6.90% higher. This percentage is not as high as Dyck and Zingales (2004) find in their research, but it is a significant premium. The year dummies of 2004 until 2007 have a significantly negative effect on the bid premium compared to 2010, while the dummies for 2008 and 2009 are not significant. This indicates that the bid premium was lower in 2004-2007 than in 2008-2010. This was already suggested by Figure 1 in section 4. A possible explanation is the credit crunch. Moreover, regulations can also have influenced the bid premium after 2007. However, it is difficult to separate these effects in this regression. Moreover, there is a transition period regarding the regulations between 2004 and 2007 where companies can anticipate to forthcoming regulations. In section 5 the effect of the different countries’ regulations are further examined.

5.3 Robustness checks

Previous results could be unstable to different specifications of the regression model. To investigate whether this is the case, several robustness checks are performed.

1

Table 8: OLS regression results for bid premiums and fairness opinions use

The sample for the regressions consists of tender offer from 2004 to 2010 in Austria, Belgium, Germany, France, The Netherlands and Switzerland. The dependent variable is the bid premium one month prior to the announcement of the deal. FO and FO ind are dummy variables equal to one if a fairness opinion or an independent fairness opinion, respectively. is found. The other variables have been explained in other tables before. T-statistics are in parentheses and computed after correcting the standard errors for heteroskedasticity.

Expected sign 1 2 3 4 Constant 21.774 *** 101.037 *** 105.783 *** 106.678 *** (8.679) (3.507) (3.898) (3.905) FO + 11.598 ** 4.684 4.590 4.786 (2.384) (0.894) (0.883) (0.925) FO ind + -10.637 ** -5.530 -5.424 -5.836 (-2.196) (-1.149) (-1.134) (-1.223)

Log (Deal size) - 0.757

(0.363)

Deal attitude friendly - 8.551

(0.763)

Method of payment stock - -8.393 -7.740

(-1.302) (-1.223)

Cross-border + 6.003 * 6.449 ** 6.898 **

(1.851) (2.020) (2.184)

Log (ROE Target) - -28.512 *** -27.638 *** -28.031 ***

(-3.337) (-3.218) (-3.235)

Log (D/E Target) + -1.911

(-0.964) Log (M/B Target) - -18.413 ** -18.135 ** -18.527 ** (-2.059) (-2.049) (-2.103) Percent acquired + 0.284 *** 0.284 *** 0.277 *** (5.091) (5.537) (5.606) 2004 -18.558 ** -18.059 ** -18.323 ** (-2.009) (-2.003) (-2.014) 2005 -18.890 ** -19.518 ** -19.825 ** (-2.085) (-2.184) (-2.195) 2006 -18.726 ** -18.817 ** -18.496 ** (-2.157) (-2.205) (-2.141) 2007 -21.419 ** -21.478 *** -21.106 ** (-2.560) (-2.596) (-2.524) 2008 -8.988 -9.401 -8.901 (-1.010) (-1.068) (-1.000) 2009 0.363 -0.589 -0.988 (0.037) (-0.060) (-0.099) N 340 310 310 312 Adjusted R² 0.015 0.200 0.205 0.204 F-statistic 3.625 ** 5.838 *** 7.113 *** 7.632 ***

5.3.1 Different definitions of the bid premium

To test the influence of different definitions of the bid premium, regressions are run with two alternative definitions of the bid premium. Specifically, the target share price is taken at different points in time. First, in the robustness check the bid premium is defined as the percentage difference between the offer price and target share price one week prior to the announcement date. The second definition looks at the percentage difference between the offer price and target share price one day prior to the announcement date. The results of these regressions are presented in appendix E, Table E.1. Models 1 and 2 show the results for the ‘week definition’, while models 3 and 4 show the ‘day definition’ results.

Models 1 and 2 are both significant with an F-statistic of 4.860 (p<0.01) and 7.614 (p<0.01) respectively. The adjusted R² of both models is slightly lower than in the regressions with the month definition, but still comparable to other previous research. The main conclusion of the fairness opinion and the independent fairness opinions having no significant effect on the bid premium is robust to the ‘week definition’ of the bid premium. Also the control variables do not show a major change compared to the original model. The variables Cross-border, Log(M/B Target), Percent acquired and the year dummy for 2007 are still significant.

Next, looking at models 3 and 4 we see that both models are significant (F=4.537, p<0001 and F=7.682, p<0.01). Again the adjusted R² of both models is slightly lower than in the original regressions. The variable for the use of a fairness opinion or an independent fairness opinion are also not significant for the ‘day definition’. The year dummies of 2006 and 2007 are significant. The control premium is again the most important explanatory variable for the bid premium. In conclusion, the original model is robust to changes in the definition of the bid premium.

5.3.2 Influence of countries’ regulations

To check if the bid premium changed after implementing the national regulations, a regression is run including a dummy for the whole sample period with value equal to one from the moment new regulations have been issued in that country. Between 2006 and 2007 the sample countries incorporated the EU directive into their laws. Switzerland is excluded from this analysis, since they are not part of the European Union and do not use the EU directive. Table E.2.2 gives the results from the regression analyses. Models 1 and 2 include year dummies and models 3 and 4 leave them out. As in the original models, all four models are significant and the variables for the fairness opinions are not significant. Thus, the effect of fairness opinions on the bid premium is robust for including a dummy about the EU directive. However, it is difficult to separate the effect of the year dummies and that of the regulation dummy. Since the sample countries incorporated the EU directive into their laws in 2006 and 2007 and the credit crunch began in 2007, the regulations dummy can measure the same effect as the year dummies. When regressions 1 and 2 are run including year dummies, the EU directive dummy is not significant, while it is significant when the year dummies are removed in regression 3 and 4. It is hard to tell why the EU directive dummy is significant in the latter regressions because now they measure also year effects. Therefore, no conclusion can be drawn about the impact of the regulations on the bid premium. Untabulated results show that only in the Netherlands the median bid premium is different (e.g. higher) at the 5% level before and after regulations are put in place. Again, this can be a proxy for year effects.

5.3.3 Alternative definition of independence

The definition of independence in this research is two-fold: the financial advisor may not be involved in other aspects of the deal besides the fairness opinion and must receive a fixed fee. To check if the regression results are robust to another definition, a wider definition of independence is used in a new regression. This wide definition defines the fairness opinion to be independent when the fairness opinion states that:

1. the financial advisor only provides a fairness opinion and is not involved in any other way with the deal, but the kind of fee is not mentioned, or;

3. providing a fairness opinion is a separate service of the financial advisor for which they receive a fixed fee, but the financial advisor also performs other services to the company in the deal for which they receive a contingent fee.

The results of this robustness check are presented in table E.3 in the appendix. These regressions have a similar adjusted R² than the original regression: between 0.200 and 0.204. The dummy for an independent fairness opinion is again negative but statistically insignificant. Concluding, the original regression is also robust to a wider definition of independence of the financial advisor. 5.3.4 Self selection bias

To control for the possible self selection bias mentioned in section 4, regressions are run without deals from 2004 and 2005. This way, the years for which finding a fairness opinion on the websites and with search engines proved to be difficult are deleted. Table E.4 in the appendix shows the results for this regression; leaving out these years does not alter the main results. Model 1 runs the complete regression with all control variables and model 2 leaves out the insignificant control variables. Both models are significant (F=6.338, p<0.01 and F=7.346, p<0.01) and the adjusted R² is higher than the original regression (0.233 and 0.236). The coefficient for the fairness opinion and the independent fairness opinion are not significant different from zero. Deal attitude friendly is highly significant (p<0.01), while this was not the case in the original regressions. So, the original regression is robust for the main variables, but the control variables are influenced by the sample period.

5.4 Interpretation of results

The univariate analysis shows that a higher bid premium is found in deals where a fairness opinion is found. This analysis also concluded that a lower bid premium is present in deals with an independent fairness opinion. However, in the multivariate analyses the insignificant coefficients of the use of a fairness opinion and an independent fairness opinion indicate that fairness opinions do not influence the bid premium. Although the positive sign for the use of a fairness opinion and the negative sign for an independent fairness opinion are in line with the univariate results, their high standard errors and as a result insignificance makes that we cannot confirm their influence on the bid premium.

premium. Target board members might choose a fairness opinion to protect themselves against possible lawsuits from shareholders who do not agree with the deal. Since fairness opinions are observed in larger deals, there is more at stake for the target board members. Larger deals also attract the attention of shareholders more than smaller deals do. The data also shows that fairness opinions are used more often in deals where a higher percentage of the target’s stock is acquired. These things give the target board members an incentive to prevent being sued by shareholders. One way of doing this is seeking a fairness opinion.

Hypothesis 2 can also be rejected. There is no evidence that the bid premium is higher for independent instead of dependent fairness opinions. In the contrary, there is some evidence that points in the other direction, although this could not be statistically confirmed in the regression analyses. Target companies might seek an independent fairness opinion instead of a dependent one in larger deals where a low bid premium is offered. The fairness opinion would then be used to convince shareholders that the offered premium is indeed a fair bid, since an independent financial advisor confirms this. Bebchuck and Kahan (1989) state this as a possible reason for directors to use a fairness opinion. Thus, the causality of the relation between the bid premium and the use of an independent fairness opinion could be reversed.

Table 12 shows the predicted and found signs for the control variables. The predictions for Cross-border, ROE Target, M/B Target and Percent acquired are confirmed in this research. The other variables proved to be insignificant in explaining the bid premium.

Table 12: predicted and found signs of control variables influencing the bid premium

Control variable Predicted sign Found sign

Deal size - Insignificant

Deal attitude friendly - Insignificant

Method of payment stock - Insignificant

Cross-border + +

ROE Target -

-D/E Target + Insignificant

M/B Target -

-Percent acquired + +

M/B Acquirer + Insignificant2

2

6. Conclusions

This study empirically examines the relationship between fairness opinions used in tender offers and the bid premium. Moreover, this study looks at the impact of independent financial advisors as opposed to dependent financial advisors issuing a fairness opinion. It is the first to look at this relation in Europe and the period between 2004 and 2010. Financial advisors are classified as independent, when they receive a fixed fee and are not involved in any other way with the deal besides providing a fairness opinion. Included countries are Austria, Belgium, France, Germany, the Netherlands and Switzerland. Fairness opinions seem to be used infrequently by acquiring companies, or at least are not disclosed, since only three fairness opinions for acquirers are found in the sample of 340 deals. For target companies, fairness opinions are found in 56% of the deals. This is in line with previous research of Makhija and Narayanan (2007), who report that 53% of U.S. targets use a fairness opinion. Independent fairness opinions are more often used than dependent fairness opinions: 37% versus 19% of the deals. This is partly caused by regulations where an independent fairness opinion is mandatory in certain deals. In line with the outcome hypothesis of Rossi and Volpin (2004), the mean and median bid premium are lower in this research than in previous studies that look at the U.S. and Australia, who are characterized by widely-held companies and high investor protection.

Results show that there is no convincing evidence of fairness opinions influencing the bid premium. Although the median bid premium is significantly higher in deals where fairness opinions are used than in deals where they are not present, regression results indicate that this difference is caused by control variables like deal and firm characteristics. The bid premium consists mainly of a control premium. When the percentage of acquired stock in the deal is one higher, the bid premium goes up with 0.28%. In addition, consistent with Kisgen, Qian and Song (2009), a negative relation is found between the bid premium and the return on equity of the target company. An increase of 1% in return on equity leads to a 12.23% lower premium.

economical meaningful item. Moreover, target boards could ask for an independent fairness opinion in deals where a lower premium is offered, to convince the shareholders that they will receive a fair price.

6.2 Limitations and recommendations for further research

This study is subject to some limitations that have to be considered. First of all, it is possible that not all fairness opinions that are used in deals are also found on the internet. Because of a lack of databases in many countries like the ones in Austria or Switzerland, where all fairness opinions are available, we cannot be completely sure that every fairness opinion is recovered. Therefore, the number of fairness opinions and conclusions in this research can be biased.

Second, the classification of the independence of financial advisors can be prone to errors. Advisors are classified as independent when they state in the fairness opinion that they comply with the criteria mentioned earlier. It is assumed that they are not independent, if they do not explicitly state their compliance to these criteria. However, it could be the case that they are independent, but they do not state it in the fairness opinion. On the other hand, investors only have this information at hand to make their decisions so they will also not know if a financial advisor is independent if this is not stated explicitly in the fairness opinion.

Lastly, the small sample size compared to previous research could be a reason that the regression results for the fairness opinion variables are not significant. Therefore, it could prove to be valuable to expand the dataset to other countries or a longer time period and to do the analyses again at a later time.

REFERENCES

Alexandridis, G., C. F. Mavrovitis and N. G. Travlos, 2011, How Have M&As Changed? Evidence from the Sixth Merger Wave, Working Paper, Available at SSRN:http://ssrn.com/abstract=1632263. Bebchuk, Lucian A. and Marcel Kahan, 1989, Fairness Opinions: How Fair are They and What Can Be Done

About it?, Duke Law Journal, Vol. 27, pp. 27-53.

Bebchuk, Lucian A., 1999, A Rent-Protection Theory of Corporate Ownership and Control, Harvard Law School, Cambridge, MA.

Bhattacharya, Utpal and Hazem Daouk, 2002, The world price of insider trading, Journal of Finance, 57, 75-108.

Bris, A. and C. Cabolis, 2004, The Value of Investor Protection: Firm Evidence from Cross-Border Mergers,

Review of Financial Studies, 21, 605-648.

Brooks, C., 2008, Introductory Econometrics for Finance, 2nd edition, Cambridge University Press, Cambridge

Bugeja, M., 2007, Voluntary Use of Independent Valuation Advice by Target Firm Boards in Takeovers,

Pacific Basin Finance Journal, Vol. 15, 368-387.

Caddell, R., 2007, SEC to banks: You’re Liable to Shareholders for Fairness Opinions, Corporate

Financing Week, 23 (39), 1-2.

Calomiris, Charles W. and Donna M. Hitscherich, 2005, Banker Fees and Acquisition Premia for Targets in Cash Tender Offers: Challenges to the Popular Wisdom on Banker Conflicts. Available at SSRN:

http://ssrn.com/abstract=708222.

Chen, L.H., 2010, The Use of Independent Fairness Opinions and the Performance of Acquiring Firms,

Journal of Accounting, Auditing & Finance, June 1, 323-350.

Chidambaran, N. K., K. John, Z. Shangguan and G. Vasudevan, 2010, Hot and Cold Merger Markets. Review

of Quantitative Finance and Accounting, 34, 327-349.

Cleveland, Steven J., 2006, An Economic and Behavioral Analysis of Investment Bankers When Delivering Fairness Opinions, Alabama Law Review, 8, 299-348.

Da Silva Rosa, R., H. Izan, A. Steinback and T. Walter, 2000, The method of payment decision in Australian takeovers: an investigation of causes and consequences, Australian Journal of Management, 25, 67–94.

Davis, A., 2004a, NASD scrutinizes conflicts in bankers’ fairness opinions, Wall Street Journal (Eastern

edition), June 11, pg. A.1.

Davis, A., 2004b, Input is sought on “fairness” rule, Wall Street Journal (Eastern edition), November 12, pg. C.3.

independent; banks that do them often are advisers on transactions and have fees at stake; a high-profit margin item, Wall Street Journal (Eastern edition), December 29, pg. A.1.

Davis, A. and D.K. Berman, 2005, Checkup prompts search for second opinions; as regulators heighten scrutiny of Wall Street’s views, outside merger assessments gain, Wall Street Journal (Eastern

edition), January 24, pg. C.1.

Davis, A., 2005, Wall Street’s “fairness opinions” draw fire from CALPERS, Wall Street Journal (Eastern

edition), February 8, pg. C.1.

DeAngelo, L., 1990, Equity valuation and corporate control, The Accounting Review, 65, 93–112.

Dyck, Alexander and Luigi Zingales, 2004, Private Benefits of Control: An International Comparison, Journal

of Finance, 59, 2, 537-600.

Eddey, P., 1993, Independent expert's reports in takeover bids, Accounting and Finance, 33, 1–18.

Franks, J., R. Harris and S. Titman, 1991, The postmerger share-price performance of acquiring firms, Journal

of Financial Economics, 29, 81–96.

Huang, Y. and R. Walkling, 1987, Target abnormal returns associated with acquisition announcements: payment, acquisition form, and managerial resistance, Journal of Financial Economics, 19, 329-349. Jensen, Michael C., 1993, The modern industrial revolution, exit, and the failure of internal control systems,

Journal of Finance, 48, 831-880.

Jensen, M.C. and W.H. Meckling, 1976, Theory of the firm: managerial behavior, agency costs and ownership structure, Journal of Financial Economics, 3, 305−360.

Kisgen, D.J., Jun ‘‘QJ’ Qian and Weihong Song, 2009, Are fairness opinions fair? The case of mergers and acquisitions, Journal of Financial Economics, 91, 179–207.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 1998, Law and finance,

Journal of Political Economy, 101, 678-709.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 2000, Investor protection and corporate governance, Journal of Financial Economics, 58, 3-27.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 2002, Investor protection and corporate valuation, Journal of Finance, 57, 1147-1170.

Makhija, A. K. and R. P. Narayanan, 2007, Fairness opinions in mergers and acquisitions, Working Paper, The Ohio State University.

Mann, H. B. and D. R. Whitney, 1947, On a Test of Whether one of Two Random Variables is Stochastically Larger than the Other, Annals of Mathematical Statistics, 18 (1), 50–60.

Manne, H.G., 1965, Mergers and the market for corporate control, Journal of Political Economy, 73, 110−120. McLaughlin, R. M., 1990, Investment Banking Contracts in Tender Offers: An Empirical Analysis, Journal of

Financial Economics, 28, 209-232.

Financial Economics, 68, 325-352.

Newbold, Paul, William L. Carlson and Betty Thorne, 2003, Statistics for business and economics, Pearson

Education, New Jersey.

Parijs, S., 2005, Fairness opinions and liability, Kluwer, Amsterdam.

Rau, P. R., 2000, Investment Bank Market Share, Contingent Fee Payments, and the Performance of Acquiring Firms, Journal of Financial Economics, 56 (May): 293–324.

Rossi, S. and P. Volpin, 2004, Cross-country determinants of mergers and acquisitions, Journal of Financial

Economics, 74, 277-304.

Sweeney, P. and C. Sechler, 1999, Who Says It Is a Fair Deal?, Journal of Accountancy, 188 (2), 44-50. Van Gessel, Ronald, 2011, Bank belang bij goede afloop - Waarderingsrapport bij overnames is gekleurd, De

Financiële Telegraaf, January 22.

Waynsley, J., W. Lane and H. Yang, 1983, Abnormal returns to acquired firms by type of acquisition and method of payment, Financial Management, 12, 16–22.

Appendix A

Table A1: definition of variables used in this research

Variable Definition

PR Premium paid for the target company, defined as the percentage premium of the offer price over the target price 1 month prior to deal announcement FO A dummy variable that takes the value 1 if a fairness opinion is found for the

deal, 0 if there is not found a fairness opinion

FO ind A dummy variable that takes the value 1 if the financial advisor is independent, 0 otherwise

Deal size The total transaction value (enterprise value) in million of Euros, which includes amounts paid for the target common stock, preferred stock, net debt, options/warrants and other deferred payments

Deal attitude friendly A dummy variable that takes the value of 1 if the deal attitude is friendly, 0 if the deal attitude is hostile.

Method of payment stock A dummy variable that takes the value of 1 if at least 50% of the transaction is paid by stock, 0 otherwise

Cross-border A dummy variable that takes the value of 1 the transaction is cross-border, 0 if it is domestic

ROE Target The ratio of earnings to average equity for the prior fiscal year D/E Target The ratio of debt-to-equity of the target for the prior fiscal year

M/B Target The target’s market-to-book ratio, measured as the ratio of the year-end market value of common stock to the book value of equity for the prior fiscal year

M/B Acquirer The acquirer’s market-to-book ratio, measured as the ratio of the year-end market value of common stock to the book value of equity for the prior fiscal year

Appendix B

Table B.1: Descriptive statistics total sample

Austria Belgium France Germany Netherlands Switzerland Total Premium Mean 31.25 31.34 20.06 23.64 32.56 21.67 23.85 Median 25.58 24.50 9.04 15.35 20.36 14.21 14.71 Std. Deviation 29.27 37.90 29.19 30.02 33.35 21.51 29.56 Minimum -0.19 -5.21 -26.00 -20.67 -3.07 -7.85 -26.00 Maximum 113.33 161.90 127.45 145.90 139.64 74.27 161.90 N 13 27 115 96 46 43 340 Deal size Mean 762.67 705.37 971.21 1,013.04 3,238.15 663.94 1,221.78 Median 110.59 136.37 72.14 87.30 828.74 221.56 120.35 Std. Deviation 1,144.08 2,145.01 5,377.41 3,526.88 10,741.98 1,290.80 5,457.75 Minimum 8.50 9.16 5.35 5.20 7.48 5.55 5.20 Maximum 3,720.52 11,202.35 56,566.01 23,175.32 71,937.01 7,112.90 71,937.01 N 13 27 115 96 46 43 340.00

Deal attitude friendly 100.0 96.3 99.1 97.9 97.8 97.7 98.2

Method of payment stock 7.7 0.0 7.8 4.2 8.7 16.3 7.4

Table B.2: Descriptive statistics 'fairness opinion' subsample

Austria Belgium France Germany Netherlands Switzerland Total Premium Mean 31.12 23.92 20.95 28.49 35.39 22.50 25.48 Median 25.58 19.25 10.09 21.46 19.98 16.01 16.64 Std. Deviation 28.66 26.25 28.02 31.53 36.25 21.21 28.70 Minimum 6.90 -5.21 -21.89 -20.67 -3.07 -7.85 -21.89 Maximum 113.33 66.00 124.25 145.90 139.64 67.85 145.90 N 11 10 69 39 30 32 191 Deal size Mean 865.92 1,515.59 447.02 1,612.90 2,331.73 521.08 1,073.59 Median 110.59 152.37 89.56 143.76 1,341.55 222.08 192.10 Std. Deviation 1,220.54 3,462.55 1,325.18 4,626.27 3,326.64 859.28 2,807.36 Minimum 8.50 9.96 5.35 5.53 10.91 11.57 5.35 Maximum 3,720.52 11,202.35 9,799.68 23,175.32 12,568.79 3,530.78 23,175.32 N 11 10 69 39 30 32 191

Deal attitude friendly 100.0 90.0 98.6 94.9 100.0 96.9 97.4

Method of payment stock 9.1 0.0 5.8 0.0 10.0 18.8 7.3

Table B.3: Descriptive statistics 'no fairness opinion' subsample

Austria Belgium France Germany Netherlands Switzerland Total Premium Mean 31.95 35.71 18.73 19.25 27.25 19.25 22.00 Median 31.95 25.13 6.48 11.95 20.71 13.98 12.50 Std. Deviation 45.45 43.49 31.13 30.30 27.41 23.24 31.71 Minimum -0.19 -0.13 -26.00 -52.84 -1.15 -2.40 -52.84 Maximum 64.09 161.90 127.45 140.38 95.34 74.27 161.90 N 2 17 46 57 16 11 149 Deal size Mean 194.84 228.78 1,757.51 602.62 4,937.69 1,079.54 1,411.76 Median 194.84 136.37 61.19 67.91 147.34 161.63 74.62 Std. Deviation 218.03 284.86 8,339.55 2,481.24 17,889.97 2,110.37 7,618.78 Minimum 40.67 9.16 6.20 5.20 7.48 5.55 5.20 Maximum 349.01 973.47 56,566.01 18,386.34 71,937.01 7,112.90 71,937.01 N 2 17 46 57 16 11 149

Deal attitude friendly 100.0 100.0 100.0 100.0 93.8 100.0 99.3

Method of payment stock 0.0 0.0 10.9 7.1 6.3 9.1 7.5