Does it pay to hire an Investment Banker?
New evidence on the role of financial advisors in M&A transactions in the EU 15
countries
ALBÈRT BLOEMHOF1
ABSTRACT
This thesis deals with the question whether or not it is beneficial for European acquiring companies to hire a top-tier investment bank to accompany them in merger and acquisitions (M&As). Using a sample of 4,724 transactions completed by EU 15 based companies acquiring public, private, or subsidiary companies, we find that for UK based bidders acquiring listed and private transactions it is beneficial to hire a top-tier investment bank. For bidders who reside in the other European countries (EU 14) the involvement of top-tier investment banks does not lead to superior returns. In addition, the recent financial crisis did not have a spoiling effect on the perception of top-tier investment bank reputation. Next, a prior relationship between acquirer and a top-tier investment bank is of significant positive importance in the decision making process whether or not to employ a top-tier financial advisor. Furthermore, it is demonstrated that top-tier investment banks do not succeed in shorten the time needed to complete an M&A deal. Also, top-tier advisors are capable of limiting takeover premiums in UK based listed deals whereas the opposite is true for EU 14 based listed transactions. Regarding the method of payment, top-tier investment bank employment is positively related to the likelihood of a transaction which is entirely financed with stock.
Key words Investment bank reputation, mergers and acquisitions, EU 15, short-term announcement effects, credit crisis, information asymmetry
JEL classification G24, G32, G34
1 Student at the University of Groningen, Faculty of Economics and Business, Department of Economics, Econometrics and Finance. This thesis is written to fulfil the requirements for the degree of Master of Science in Business
I. Introduction
REPUTATION OF FINANCIAL ADVISORS plays a key role in facilitating acquisitions by firms in a world of imperfect information. Foremost since the advisors derive their abilities to charge fees and attract new businesses by their perceived capabilities, talents, trustworthiness – ultimately, in depends all on reputation. Also, reputation is the main motivating factor in bank’s hard work to bring into being high-quality information since reputation is a virtuous circle. A client recognize them for their high-quality work and consequently grants them the capital to composite their competitive advantage. This leads to increased effort to produce superior information due to the higher reputation the bank has at stake. Next to that, reputation is built over time and is the result of good performance of bankers; therefore, reputation induces investment banks to place their long-term relationship with investors above short-term benefits created through opportunistic behavior.
Mergers and acquisitions (M&As) are one of the most fiercely debated issues in corporate finance. Acquisitions are the firm’s largest investment; they involve a change in the firm’s capital structure, as well as, in many cases, a strategic reorientation of the company. Typically financial advisors play an important role in the M&A process. McLaughlin (1990), among others, elaborate on the core activities that M&A-advisors carry out for clients: (i) identifying possible bidders and targets, (ii) negotiating, finishing offers, looking for higher bids, protecting against hostile offers, (iii) advising on the bidding strategy and offer price, on the accept/reject decision, and evaluating the potential for competitive bids. Servaes and Zenner (1996) find support for three reasons for the employment of investment bankers in M&As: firstly, investment banks have a comparative cost advantage in analyzing transaction. Second, they reduce information asymmetry. Lastly, the certification of the M&A transaction by the investment bank reduces agency costs. According to Schiereck, Sigl-Grüb, and Unverhau (2009), practitioners emphasize the role of investment banks in providing liquidity and therefore the increase in efficiency on the market for corporate control.
May be only as implementation offices to make sure that the presumed deal is completed? On the other hand, are there conditions in which it pays off to employ a top-tier financial advisor? Golubov et al. (2012) used a dataset of US based M&A activities containing public, private, and subsidiary deals announced over the period from 1996 to 2009. Importantly, they found that top-tier financial advisors generate higher bidder gains than non-top-tier advisors but in public acquisitions only. But does this hold for a sample of European acquisitions? And, why does it, or why does it not? And is there any difference between the perception of investment bank reputation in market based economies such as the United Kingdom on one hand, and EU 14 countries on the other hand where relationship banking and commercial banking is much more prevalent?
The existing body of research conducted in the past examined the US market very thoroughly in contrast to the European market for corporate control. More specifically, there is a total lack of evidence on the relation between advisor reputation and gains generated by Europe domiciled bidders. Therefore, this study focuses on Europe, and makes distinctions between the UK and the EU 14 domiciled bidders based on reasons which are presented hereafter.
Culpepper (2011) argues that the UK and US have an active market for corporate control whereas the rest of the EU or Continental Europe is designated as a passive market for corporate control, where political and business leaders collude to prevent large companies from being treated as disposable assets. To illustrate this, the management of companies in the EU 14 often mobilized enough political support to neutralize any attempt of hostile takeovers. Next to that, takeovers within Continental Europe still have specific characteristics, different from the Anglo-Saxon countries and Asia. These differences mainly arise in their attitude, acquisition techniques, payment methods, and premiums. For example, with respect to their attitude, Rossi and Volpin (2004) demonstrate that hostile deals and competing bids are more common in countries with better shareholder protection such as the UK. Also, stock deals are more frequent in the UK as all-cash bids are more prevalent in the EU 14 (Rossi and Volpin, 2004; Faccio and Masulis, 2005). Investment banks are seemingly increasingly hired when deal complexity increases such as in all-stock financed transactions since these deals demand strong advisor professional qualifications. Therefore, it is justifiable that the reputation mechanism is more prevalent in deals announced by UK based acquirers. This is line with the findings of Golubov et al. (2012) that the reputation mechanism leads to superior bidder returns in public deals. Part of their way of thinking is that services delivered by investment bank associated with public deals are more complex and publicly observable, and therefore the reputation mechanism is relatively more important.
announced between 1997 and 2002. However, these authors do not divide their sample into listed, private, and subsidiary acquisitions, nor make geographical distinctions between e.g. the UK and the other European countries. A distinction between the UK and the other European countries is rational based on the reasons put forward by Schiereck et al. (2009) to validate their focus on European bidders apart from US acquirers. Yet the characteristics of the US market for corporate control are very similar to the UK’s market for corporate control and specifically the investment banking working environment. As mentioned above, their arguments include competition level in the market of corporate control, possible conflicts of interest, and the potential influence of financial advisors on M&A outcomes. Schiereck et al. (2009) had a rather limited number of observations available covering the whole EU 15; since this study considers more than 4,500 M&A transactions announced between 1996 and 2012 by EU 15 based acquirers, it is practically feasible to draw comparisons.
The US and UK market for corporate control is thus especially mature and the number of M&A activities are high, in contrast to the EU 14 M&A market. Billet and Qian (2008) argue that severe competition for target firms tend to increase managerial hubris-related effects that in turn could lead to reduced bidder gains. Next to that, fierce competition may worsen the negative aspects of agency problems (Jensen 1986; 2005). Mandelker (1974), Asquith (1983), and Rossi and Volpin (2004) argue that the US and UK markets for corporate control are more competitive; they for instance document that in the US and the UK takeover premiums are substantially higher; as a result target companies tend to capture most of the M&A benefit. Therefore, the potential for value creation for bidders is limited in the most competitive markets. Alexandridis, Petmezas, and Travlos (2010) demonstrate that acquirers within the most competitive markets for corporate control (US, UK, and Canada) significantly underperform those in the rest of the world. These findings lead to increased challenges for investment banks advising acquirers within the US, UK, and Canada, since we noted that bidder value creation is difficult in those countries and that top-tier advisors presumably pursue the best deals for their employers (which is among others reflected in bidder gains). Since Golubov et al. (2012) document that top-tier investment banks are able to provide superior services in a highly competitive market for corporate control such as the US market, we expect that the same holds for the UK market as the circumstances for this markets are very similar. Given that the European M&A market is found to be less competitive, it is reasonable that it is less difficult for financial advisors to discover the premium targets for their clients and that it is harder for top-tier advisors to distinguish from the non-top-tier advisors by delivering superior performances in the EU 14 market for corporate control.
Furthermore, Abrahamson, Jenkinson, and Jones (2011) found an entry barrier around investment bank reputation in the US while in the European countries investment banks price their IPO services strategically in order to increase market share; this indicates that reputation only is not sufficient to increase market share in Europe. The different market circumstances around investment bank reputation affect the perception of reputation and presumable the outcomes of the employment of top-tier investment banks, such as bidder gains.
financial statements, e.g. financial resources, cash flows, et cetera. Next to that, they may have access to other exposures based on its long-term relationship with the companies involved. The authors find that bidder firms are more likely to employ a commercial bank if they have had a prior relationship with the bank. The EU 14 based companies are generally much more centered around commercial banks and more based on relationship banking compared to their UK domiciled counterparts. Therefore, commercial banks may have a competitive advantage over investment banks for EU 14 based transactions, in contrast to UK based deals where commercial banks have a minor role in the M&A market. Since the market participants are well aware of this situation it is expected to be reflected in bidder gains.
In sum: based on the lower level of competition in the EU 14 corporate control market, the assumed increased deal complexity for UK based deals, the finding of Allen et al. (2004) that commercial banks may have a competitive advantage over investment banks in the EU 14, and the result of Abrahamson et al. (2011) that in Europe reputation alone is not sufficient to hold and expand market share, it is reasonable to expect that the reputation mechanism for top-tier investment banks is less important for transactions announced by EU 14 based bidders than for M&A deals announced by UK based acquirers.
We re-examine the function of financial advisors in M&A activities by investigating the connection between investment bank reputation and the quality of the services they provide. A large and comprehensive dataset is used containing more than 4,500 acquisitions of public, private, and subsidiary firms announced over the period from 1 January 1996 to 1 February 2012. Next to partition based on geographical origin, we separately examine different types of M&As acknowledged on the basis of the target company’s listing status since Golubov et al. (2012) showed that the relation between financial advisor reputation and M&A outcomes differ between target listing status. This reputation mechanism is more present in transactions that lead to larger reputational exposure. According to Rhee and Valdez (2009), greater visibility creates greater potential reputational damage. Secondly, Golubov et al. (2012) argue that public acquisitions call for more capabilities and endeavor on the part of the advisors as: (1) it is more difficult to detain abnormal returns in acquisitions of public targets due to superior negotiating skills of publicly held targets compared to that of unlisted firms (Fuller, Netter, and Stegemoller, 2002; Officer, 2007); (2) this type of deals require more disclosures and normally necessitate regulatory and/or shareholder approvals, growing deal complexity and therefore demanding strong advisor professional certification; and (3) given the dispersed ownership of publicly held target firms, there is characteristically no identifiable entity to back any concealed or disclosed liabilities of the target after the deal is closed, which hinders the aptitude of the acquirer to put together any form of post deal indemnification from the seller, and consequently puts ex ante stress on the acquiring company’s investment banker to perform. Summarily, it is reasonable to expect that advisor reputation is comparatively more important in acquisitions of public firms.
premium fees for more reputable financial advisors. Also, firms may want to know if increased bank reputation is of any importance for them in the M&A initiating, guidance, and execution processes. They are presumably looking for any benefits. On the other hands, banks themselves get insights in their performances, and the factors influencing their acts. Next to acquirers and banks, investors are highly interested in the general reliability of the quality and reputation signaled by banks through e.g. the league tables and their performances.
For practitioners, we provide justification for the construction of league tables based on transaction values they advice on, however, only for UK based transactions. In this respect, for the M&A market in the UK investment bank’s position in these top-ranking indicate the quality of its provided services. On the other hand, for EU 14 based transactions, we do not find justification for the construction of these tables since employment of top-tier investment banks do not lead to superior returns; moreover, we do not find signs of premium quality in terms of shorter times need to complete a deal, or their ability to limit the transfer of wealth from the acquirer’s shareholder towards the target owners through superior negotiation skills.
This study is related to several studies conducted in the (recent) past such as McLaughlin (1990, 1992), Chemmanur and Fulghieri (1994), Servaes and Zenner (1996), Rau (2000), Kale, Kini, and Ryan (2003), Hunter and Jagtiani (2003), Ismail (2010), and Golubov et al. (2012). Most of them look at the relation between the reputation of financial advisors and the results of conducted M&A activities; however, they all study US based transactions. Apart from these authors, Schiereck et al. (2009) examine the reputation mechanism for European transactions without any significant result. This study makes contributions to the literature in several ways. First of all, the study offers new facts on the effect of investment bank reputation on acquirer returns. With respect to the EU 15 countries, no prior evidence of the importance of the reputation mechanism in M&A context was found. Schiereck et al. (2009) did not examine differences between EU 14 domiciled bidder companies on one hand and UK based bidder firm on the other hand, nor division of transactions by target listing status. Specifically, we show that UK based public and private bidder companies gain more when they employ a top-tier investment bank. Alternatively, employment of top-tier investment banks by EU 14 domiciled buyers does not lead to superior bidder gains. We also find that it is reasonable consider the endogenous nature of bidder-advisor matching. Next, our study provides new evidence on the determinants of the choice to make use of a financial advisor in Europe. We show that the prior relationship between acquirer and financial advisor is an important determinant. Lastly, top-tier investment bank employment is positively related to the likelihood of transactions financed with entirely stock.
premiums supporting the superior skills hypothesis. The findings regarding bidder gains lead to the same conclusion as the result with respect to takeover premiums: top-tier investment bank employment leads to higher (lower) bidder returns and lower (higher) takeover premiums. Top-tier advisors are not capable of limiting the time needed to complete a deal.
The thesis is organized as follows. First, the existing body of literature related to the subject central in this study is introduced. Section III presents the dataset and the methodology. In the subsequent section the descriptive statistics and the results are presented. In Section V the limitations and recommendations are presented. Finally, in Section VI we discuss the conclusions of this study and its implications for the M&A practice.
II. Literature Review & Hypotheses
This thesis deals with the role of financial advisors in merger and acquisitions activities by investigating the relation between investment bank’s reputation and the quality of their merger advisory services. The role of financial advisors servicing clients in corporate finance has been subject to a large body of research; the vast majority of the conducted research is focused on the underwriting business (e.g., Krigman, Shaw, and Womack, 2000; Dunbar, 2000; Hansen, 2001; Garner and Kale, 2001; Chalmers, Dann, and Hartford, 2002. Less well covered is the role of investment banks in M&A. In this theoretical background we depict the available literature on the following subjects, and we also present the hypotheses which subsequently emerge. We mainly focus on bidder returns and the likelihood of hiring a top-tier investment bank.
We firstly consider the relationship between reputation, quality, and price in general terms, and more specifically with respect to investment banking services. Then we depict the European M&A context, including the role of financial advisors. We argue that it is valuable to divide the sample along target listing status and geographical differences. We also consider the available literature related to the recent financial crisis, the method of payment applied in the context of M&As associated with information asymmetry and the role of financial advisors. Also, the time needed to complete an M&A is considered.
A. Reputation, quality, and price
specifically the equity underwriting services: high reputation financial advisors offer higher-quality services and charge higher fees. Their theoretical model could be extended to M&A services provided by investment banks, according to these authors, since the quality of investment bankers is ex ante observable and the services delivered are sold repetitively.
The reputation of financial advisors in mainly driven by how much value they create for their customers. Therefore, the key question is: “how beneficial have top-tier investment banks been to their clients”?
B. Financial Advisors’ reputation in mergers and acquisitions
Prior evidence on the relationship between advisor reputation and shareholder wealth effects associated with M&As is mixed. The literature mainly focuses on two conflicting hypotheses: the superior
deal hypothesis and the deal completion hypothesis. The first hypothesis argues that high reputated investment
banks have the capabilities to identify better target companies for the acquirer who employed them and have the ability to create greater financial and operational synergies. Based on the presumed expertise of the top-tier advisors this first hypothesis expects a positive relationship between bidder gains and the reputation of financial advisors. On the other hand, the deal completion hypothesis does not per se expect such relation between acquirer wealth gains and investment bank reputation. Investment banks have strong incentives to complete M&A transactions since their fees are partially contingent on deal completion. Rau (2000) therefore argues that investment bank reputation, or their market share, will depend on the number of deals they complete. However, McLaughlin (1990) put forward that a fee-driven motivation would not increase the acquisition price since this behaviour reduces the value of their reputational capital.
A number of authors focus on the effects of diverse categories of investment banks. E.g. Bowers and Miller (1990), Servaes and Zenner (1996), Rau (2000), Rau and Rodgers (2002), Hunter and Jagtiani (2003), Kale et al. (2003) and Ismail (2010) find that top-tier investment banks provide no greater shareholder wealth effects compared to non-top-tier financial advisors. Lowinski, Schiereck, and Thomas (2004) also failed to discover evidence when looking into Swiss acquirers. Da Silva Rosa, Lee, and Walter (2004) and Walter, Yawson, and Yeung (2008) as well fail to hit upon a positive relationship between the quality of investment banks and returns. However, Golubov et al. (2012) find that top-tier advisors deliver higher bidding returns than their non-top-tier counterparts using a US based dataset, but in public acquisitions only. Next to these US based investigations, to my knowledge there is one study available focusing on the European M&A market: Schiereck et al. (2009). They do not find any significant results. This study does not differentiate between bidder country or regions, nor between target listing status.
use of boutique financial advisors leads to lower takeover premiums paid in public deals. Nevertheless, bidder gains, completion rate, and fees to be paid to boutique advisors are indifferent to the fees charged by investment banks, the bidder returns and the completion rates when acquirers use investment banks.
Chang, Shekhar, Tam, and Amy Zhu (2010) focus on the individual bank level while considering the determinants of the financial advisor choice. This study demonstrates the industry expertise of the financial advisor, and the prior relationship with the banks, have a positive influence on the likelihood of employing a particular advisor for a specific deal. A second study focusing on performance persistence of M&A advisors on a bank-level discovers significant bank-level fixed effects, next to a finding that financial advisors in the top quintile of buyer achievements persist to deliver better services than the advisors in the bottom quintile of buyer achievements for the next numerous years. Ertugrul and Krishnan (2009) focus on the individual bankers’ characteristics on the outcomes of M&A activities. They find for example that investment bankers have a statistically and economically significant effect on bidder gains, deal completion time, and the operating performance of acquirers.
Another interesting study puts the role of investment banks as insiders in the M&A market at the center of attention. Bodnaruk, Massa, and Simonov (2009) show that financial conglomerates allied with the investment banks which advise the buyers frequently build up a stake in the firm to be acquired prior to the merger and acquisitions announcement. The banks make a substantial profit by doing so. They also find a negative relationship between the size of the stake and the post merger bidder profitability.
C. Investment banks, reputation, and M&A in Europe
In the above section we showed that previous evidence on the relation between advisor reputation and shareholder wealth effects associated with M&As is mixed. With respect to empirical evidence on the market reactions to an M&A activity announced by either bidders domiciled in the EU 14 or in the UK is limited and mixed. Goergen and Renneboog (2004) demonstrate that UK acquirers obtain a CAR of 1.5% whereas the bidder gain is only 0.9% for Continental European acquiring companies. Martynova and Renneboog (2006) do not find support for this result; these authors show that UK bidder gain is lower compared to CARs experienced by EU 14 bidders.
Since we investigate the role of investment banks in M&A transactions where acquirers are domiciled in the EU 15, we depict the European banking and M&A context; in part contrary to the US M&A market. The empirical setting of this study is based in Europe, more specifically the EU 15. Investigating the European M&A market and advisor reputation may be constructive for some reasons. Some of them are in depth described in the introduction section.
Firstly, it provides an institutional setting of European countries; EU 14 on one hand, and the UK on the other hand, but different from the US market.
on the function of financial advisors in M&A in a bank-centered system since we also consider the EU 14 countries apart from the UK, which makes comparisons possible.
Third, very little attention has been devoted to the role of investment banks in the European merger and acquisitions market. However, the US market has been thoroughly investigated in this respect. The US market is very mature and the M&A activities are high, compared to the European M&A market. According to Schiereck et al. (2009), it is not always possible for a client within this market environment to take the advice of his most preferred financial advisor as this investment bank might have conflicts of interest which prohibit the mandate. Therefore, the probability of conflicts of interest should be significantly lower in less mature markets with less M&A activity. Also, the potential influence of the investment banks on M&A outcome should be less biased according to these authors.
Fourth, the reputation of top-tier financial advisors employed in M&A transactions based on transactions values may be of different importance in the EU compared to the US and the UK. The investment banks conceived to be top-tier in the US are perceived to be top-tier in Europe as well: based on the league tables constructed by Thompson based on transaction values, in this study 8 leadings banks are considered to be top-tier investment banks in Europe. 5 of these 8 top-tier investment banks are labeled as top-tier investment banks in the study conducted by Golubov et al. (2012), where the authors examine the US M&A market. Thus great similarities are observed in the leadings banks in both markets. Interestingly, there is a study digging into differences between the US and Europe related to investment banking services, although they focus on IPO services specifically. Abrahamson et al. (2011) investigate the difference in IPO fees between the US and European countries. Given the similarities in the leading banks in both IPO markets, the same underwriters charge different fees in the two markets. They show that in the US fees are tightly clustered around 7% while European fees have fallen in the last decade. There is no evidence for arguments such as legal costs, retail offerings, litigation risk, and sell-side analyst differences. Nor are US offerings under-priced less than in Europe. The only plausible explanation for the premium fees charged in the US in an industry where there appears little price competition is a lack of competition, reflecting barriers to entry around reputation contrary to the situation in Europe. There is no study which questions whether these outcomes hold for the M&A market as well, although Abrahamson et al. (2011) show that it holds for at least one business environment of investment banks. Possibly the lack of entry barriers around reputation exists for the European M&A market since it holds for at least the IPO market; therefore, there is more competition among investment banks: reputation alone is not sufficient to hold and expand market share. Based on that, perceived reputation of the top-8 top-tier investment banks as experts in capital markets transactions may be of less importance in the EU 14.
based on relationship banking compared to their UK domiciled counterparts. Therefore, commercial banks may have a competitive advantage over investment banks for EU 14 based transactions, in contrast to UK based deals where commercial banks have a minor role in the M&A market. Since the market participants are well aware of this situation it is expected to be reflected in bidder gains.
In sum: based on the lower level of competition in the EU 14 corporate control market, the assumed increased deal complexity for UK based deals, the finding of Allen et al. (2004) that commercial banks may have a competitive advantage over investment banks in the EU 14, and the result of Abrahamson et al. (2011) that in Europe reputation alone is not sufficient to hold and expand market share, it is reasonable to expect that the reputation mechanism for top-tier investment banks is less important for transactions announced by EU 14 based bidders than for M&A deals announced by UK based acquirers.
This being said, in the end the delivered performances by high reputed financial advisors are important since superior performances could justify premium fees. For the US M&A market Golubov et al. (2012) found that high reputed top-tier investment banks deliver superior performances when they advise buyers on listed transaction, while they charge premium fees to these clients. In this study we dig further into financial advisor reputation in Europe hoping to reveal whether perceived top-tier reputation leads to superior performances.
Next to reasoning above, the acquiring companies operate in a different home economy compared to US based buyers, both economically as well as regulatory. Although the (continental) European market is becoming more integrated, its starting point was a series of fragmented market. In contrast the US which was already a well established unified market. Another aspect is that European companies operate in a different economic (banking) environment than US firms do. For example, a comparative analysis of the US and EU banking markets conducted in 2003 by the Worlds Savings Bank Institute and the European Savings Bank Group reveal that the GVA1 financial intermediation/Total GVA was 5.4% for EU 15 countries while the US exhibits an 8.7%. Next to that, the total assets of the banking sector as portion of GDP comprise 277% in EU 15 countries. With respect to the US banking environment, this percentage is 88%.
The same authors argue that differences can be observed in the strategies applied to achieve the stability objectives pursued by the regulatory and supervisory bodies. The US and the EU pursue similar goals: to ensure that the banking sector is safe, both at the entity level and at the industry level. Furthermore, investors need to be protected, the integrity of the financial markets has to be guaranteed, and it has to be made sure that insurance companies can fulfill their obligations. However, differences can be observed in the strategies applied to reach this stability. E.g., European regulators say that a level playing field must be guaranteed for all companies operating in the EU area, their US equivalents believe that regulatory competition is the best approach. Furthermore, the powers and roles accredited to the regulators are very different which translates into differences between the EU and the US supervisory frameworks.
The US market is market oriented, where Europe’s financing services are considered to be relationship or bank-based. Relationship-based financing is more prevalent, but academics agree that arm’s length financing is increasingly more widespread in Europe during the last years due to the process of integration (both monetary integration at the European level and financial integration at the worldwide level), and the revolutionary nature of innovation (Rajan and Zingales, 2003). The literature (e.g. DeLong, 2003) suggests that market reactions to an M&A transaction between US (‘market-based’) and Europe (‘bank-based’) would differ. The differences between Europe and the US are very clear and are commonly agreed on. This being said, in general the UK financial system is lumped together with the US in the same category of market-based financial system. Both the US and UK are clear examples of market-based banking (Rajan and Zingales, 2003), but with significant differences such as that the UK’s banking systems is much more concentrated than the US banking system. Recent studies on the political economy of European finance, e.g. Hardie and Howarth (2010) and Howell (2007), demonstrate the marginalia of the traditional view of the UK as a market-based economy. While considering the differences between Europe and the US, we mentioned the bank assets to GDP ratio. Bank assets to GDP in the US are the lowest available, and bank loans represent a lower proportion of company financing than in Europe or Japan. The US financial system has moved towards market-based banking while the UK in contrast moved more towards a bank-based model as generally perceived. Bank lending to UK non-financial companies increased dramatically, while equity financing has been negative meaning that companies are repurchasing equity. The private bonds markets remain small. Since 2001 the customer lending grew faster than their holdings of securities (Bank of England, 2008) and since 1980 UK bank assets have increased approximately five times relative to the size of the UK economy, even excluding the banks’ large foreign currency assets (Miles, 2009). Furthermore, US commercial banks were continuing to depart from history through their national expansion of retail banking, and the shift into investment banking permitted by waves of deregulation in the 1980s and 1990s.
In this sub-section and in the introduction section we argued that it is reasonable to investigate the role of investment banks in M&A transactions in the EU 15 in order to fill the gap in the literature. Since Schiereck et al. (2009) and Golubov et al. (2012) did not find significant results for the EU 15 and US based transactions, we do not expect a significant positive or negative relation between investment bank reputation and bidder gains regarding the total sample of transactions. However, the consequence of reputation is presumably not equally important for all types of deals as suggested and demonstrated by Golubov et al (2012). These authors found that top-tier financial advisor involvement leads to higher bidder retains for public deals only; in the introduction section we elaborated more on their reasoning. We therefore investigate the reputation mechanism by separating the sample into public, private, and subsidiary acquisitions. Consequently we propose the following hypothesis where we expect that the effect of advisor reputation on
bidder CAR is significantly more important in acquisitions of public firms compared to transactions of private
H1: Top-tier investment bank employment leads to higher bidder CARs for listed transactions compared to private and subsidiary transactions
As indicated in the introduction section and above, we also examine the effects of advisor reputation for sub samples of EU 14 bidder companies and UK domiciled bidder firms. Both for the total sample containing all three types of deals (listed, private, and subsidiary) as well as for the sub samples which covers listed, private, or subsidiary transactions, respectively. The expectation is translated into the following hypothesis based on the given reasoning above and in the introduction section where it was argued that the reputation mechanism is less important for EU domiciled acquirer companies:
H2: Top-tier investment bank employment leads to lower bidder CARs for EU 14 based transactions compared to UK based deals
Furthermore, we focus on the relation between advisory fees and the reputation of top-tier investment banks. Golubov et al. (2012) found a positive relation between top-tier investment bank employment and advisory fees charged to the acquirer. Abrahamson et al. (2011) found that fees charged by top-tier investment banks in the IPO market are higher in the US than in Europe due to a lack of competition in the US which reflects an entry barrier around reputation. There may be a difference in the relation between advisory fees charged by the same banks in the UK and the fees paid for by the buyer in the EU 14.
H3: Top-tier investment bank employment is positively related to the % fee of transaction value
E. Financial crisis
a significant positive coefficient of the variable crisis, indicating that top-tier investment bank involvement in a deal which toke place during the recent crisis is positively related to bidder CARs. We translate this expectation into the following hypothesis:
H4: The financial crisis has a positive influence on the perception of investment bank reputation.
F. Method of Payment
The literature identifies the use of investment banks as mitigating problems associated with information asymmetry and agency costs. For example, well known is the notion that stock payment is used in an M&A transaction to facilitate mitigating valuation problems due to information asymmetry between the bidder company and the target firm. Next to that, stock is also offered when the acquirer has precious investment prospects, reflecting the increased discretion implied in equity offerings relative to debt payments. Nevertheless, equity payments become less striking to the management of the acquiring company if the management has an important stake in their company or of the offering of an equity payment generates a blockholder. Funds that are internally generated deliver the highest managerial discretion; consequently companies with large cash flows are more likely to use cash payments in merger and acquisitions.
Prior conducted studies indicate that bidder companies face negative stock price reactions if they pay with stock for an acquisition of a publicly held target (e.g. Travlos, 1987; Wansley, Lane, and Yang, 1987). On the other hand, Asquith, Bruner, and Mullins (1987) and Servaes (1991) show that bidders using cash experience significantly higher returns compared to stock transactions.
The listing status of target firms is of importance when looking at stock as the method of payment, since the market examines an equity issuance differently based on whether it is a public or private placement. In case the placement is public, the market may view it as an equivalent of a public issuance of equity. Consequently, the acquirer faces a negative stock price reaction to the announcement. On the other hand, if a target is privately held and is paid for by stock, then it is a private placement of equity. Chang (1998) observe that in such cases bidder abnormal returns are significantly positive in contrast to the use of cash. Explanatory reasons for the positive abnormal returns include the creation of a blockholder which can convey positive information about the firm’s future to the market, since the blockholder has potentially the ability and incentive to monitor the performances of the acquirer.
Essentially there are two information asymmetry problems in M&As. The first problem concerns the value of the acquirer company; the second problem is about valuing target companies.
There are incentives for bidder companies and their financial advisors to finance an M&A transactions with stock. According to Myers and Majluf (1984) overvalued bidder companies could exploit information on their own value by offering stock payments in an M&A transaction. Next to that, investment bankers could increase the likelihood of completing M&A deals by encouraging stock paid transactions in case acquirers do not have sufficient cash to pay the target value. Also, the involvement of financial advisors could lead to application of stock payments because the inclusion of stock leads to more complex deals which require more working hours for investment bankers. Therefore, it is reasonable that involvement of top-tier investment banks leads to more stock based offers. We elaborate more on these issues below.
We mentioned that there are two problems around information asymmetry in M&As. Next to the problem of valuing acquirer companies, there is the problem of valuing target companies. Acquirers and their financial advisors make an estimation of the value of a target company (including the expected synergy gains); their offer is based on this valuation process. The problem is that the target company has an information advantage concerning their value. According to Hansen (1987), acquirers could mitigate the problem by offering stock payments to make the offer contingent on future performances. If the acquirer offers stock, the subsequent value of the offer depends on the market reactions which ultimately lead to risk-sharing between the acquiring company and the target firm (the so called “risk sharing hypothesis”). On the other hand, in transactions financed with cash acquirers carry the risk that the acquisition premium will not be realized on their own. Based on this rationale on which we elaborate further below, one could argue that the use of the services of top-tier investment banks leads to less need of risk-sharing since top-tier investment banks are better able to value target companies.
Acquirer and investment bank incentives
transactions based on stock financing require additional working hours of the bankers involved. That may justify higher transaction fees. Finally, due to the opaqueness of stock financed transactions it is more difficult to find failures of stock based M&A deals. Negative stock prices reaction can be assigned to temporary overvaluation of stock prices. Bankers and/or managers may suggest that these overvaluations were used in the best interest of bidder company’s shareholders as transactions currency.
The probability of using top-tier investment banks may increase in case of more complex deal structures such as applying stock based financing. We argued that top-tier investment banks could show their assumed competitive advantage in complex and stock based transactions. Furthermore, stock based financing could be in interest of top-tier investment banks and also in favor of bidder company’s top managers since acquirer CEOs often obtain large amount of new stock and options grants after acquisitions. This reasoning is translated in the following hypothesis:
H5: Bidder companies are more likely to use top-tier investment banks in stock based acquisitions
Information asymmetry between acquirer and target
Travlos (1987) found that announcement returns are on average significantly negative when acquirers use stock to acquire publicly traded targets. Researchers indicate that the rationale for the negative returns associated with stock payments for public targets is adverse selection: managers use stock as currency to pay for acquisitions only when the target company’s stock is overvalued. According to the Myers and Maljuf (1984) paper, the higher the information asymmetry, the more negative is the reaction of the market to equity issues. Several authors contribute to the literature by presenting empirical evidence for this result (Moeller, Schlingemann, and Stulz (2007), Fuller et al. (2002). On the contrary, recent evidence identifies a sub-sample of takeover announcements that result in positive abnormal returns for acquiring firm stockholders if transactions are financed with stock payments. According to Officer, Poulsen, and Stegemoller (2009), this sub-sample comprises of targets that are difficult to value, e.g. private target companies. The pattern that acquirers use their own overvalued stock to pay for an acquisition, and that it will lead to a negative abnormal return, is one explanation for the application of stock in M&A transactions.
implication of the method of payment choice is post-acquisition ownership. In case a bidder company finances the transaction with shares, the owner of the target becomes a shareholder in the bidder company. When the relative deal size and the portion paid with shares is substantial, the target owners can become a significant minority shareholder. Alternatively, in case of a complete cash consideration the post-merger company risk will exclusively be taken on by the bidder shareholders.
There is likely to be substantially more information asymmetry concerning a privately held target’s value relative to a publicly held target company, because privately held companies harder to value and are more opaque than publicly held firms. Also, Beard (2004) show that private firms are more opaque than subsidiary firms. Secondly, takeovers of private companies are much less likely to elicit competing bids and do not experience the information generated by an auction for the firm. Next to that, Officer (2004) shows that in acquisition of public firms the length of time between announcement and completion influences the method of payment choice. This consideration is generally less important in acquisitions of privately held firms as those acquisitions are typically announced only when completed. Therefore, there are fewer factors affecting the method of payment decision in acquisitions of privately held targets thus we have a cleaner measure of the method of payment choice if we consider private acquisitions.
One could expect that top-tier investment banks are able to reduce information asymmetry when an acquirer buys a private company. It is incumbent on top-tier financial advisors to choose the correct method of payment in a contracting situation potentially fraught with information asymmetry. If top-tier investment banks succeed in reducing information asymmetry on the target company’s value, there is less need to share risks and therefore to offer contingent payments. This brings us to the hypothesis that the method of payment is more likely to be of cash form if the bidder company is advised by a top-tier investment bank in a private transaction.
H6: Top-tier investment bank involvement is positively related to the cash portion in private transactions
F. Time to completion
superior skills hypothesis. Therefore, we expect that top-tier investment bank involvement leads to a shorter time necessary to complete the transactions.
H7: Top-tier investment bank involvement shortens the time needed to complete a deal
G. Previous relationship
Previous relationships can affect the price (and demand) of services via a reduction of information asymmetry and a subsequent reduction in cost of these services. An investment bank or financial advisor might have acquired information about a company as a financial advisor in a prior acquisition or even as an underwriter in a bond or equity issuance or syndicated loan. Chang et al. (2010) focus on the individual bank level while considering the determinants of the financial advisor choice. This study demonstrates the industry expertise of the financial advisor, and the prior relationship with the banks, have a positive influence on the likelihood of employing a particular advisor for a specific deal.
A study conducted by Forte, Iannotta, and Navone (2010) finds that target companies’ choice to use an investment bank conditional on hiring an advisor is influenced by the closeness of the relationship. They also find that hiring an advisor with a strong prior relationship significantly increases the average cumulated abnormal return up to 40 days after the deal’s announcement. We investigate these issues for acquirer companies. Firstly, the likelihood of hiring a top-tier investment bank is examined, next to that the effects of prior relationships on the wealth of the acquirer’s shareholders is considered. We measure the extent to which the acquiring company used the services of a top-tier financial advisor across an assortment of previous capital market transactions, i.e. equity issues, bonds issues, and M&As.
III. Data and Methodology
This section includes a description of the sample selection process and criteria associated. Firstly, the sample selection process applied in order to find deals to be used in the empirical analysis is described. Secondly, in order to execute our analysis on top-tier investment banks and non-top-tier financial advisors the technique used to classify financial advisors into top-tier investment banks and non-top-tier financial advisors is explained. Third, the methodologies applied are described in depth.
A. Data selection
We gather a sample of M&As announced between January 1, 1996 and February 1, 2012 from the Thomson One Database. Bidders are EU 15 public firms, and targets are public, private, or subsidiary firms domiciled in the EU 15 as well as in non-EU 15 countries.
Successful deals are included while repurchases are excluded. Deal values higher than USD 1m are included. The original dataset included 24,924 deals. This sample is cleaned of privatizations, going private transactions, liquidations, leveraged buyouts, reverse takeovers, restructurings, and bankruptcy acquisitions, leaving a dataset of 23,888 transactions. Given that transactions that represent a transfer of control are of interest, a requirement is that the acquiring firm owns less than 10% of the target company before the deal and requests to buy more than 50% after the transaction as in Faccio, McConnel, and Stolin (2006), which brings us a dataset of 15,666 transactions. Also, the bidder advisor should be reported by Thomson One Database, which results in a sample of 5,300 transactions. We further require that the bidder be covered in the Datastream database, which result in a sample of 4,724 deals. For this dataset, announcement returns are measured for a five day time fence (-2, +2) around the announcement. The methodology is described below.
B. Financial Advisors
TABLE I
Top-25 EU 15 Financial Advisor Ranking by Transaction Value
This table presents financial advisor ranking of the top-25 investment banks according to the value of deals on which they advised for a sample of M&A transactions with EU 15 acquiring companies during the period January 1, 1996 to February 1, 2012 drawn from the Thomson One Banker Merger and Acquisition Database. Transaction value is in USD m. The number of deals advised by each advisor is also presented. Credit is allocated fully to both bidder and target firm advisors and to each eligible advisor in the case of multiple advisors for a single party. Equity carve-outs, exchange offers, and open market purchases are excluded.
Rank Financial Advisor Transaction Value Number of Deals
Top-Tier (1st to 8th)
1 Goldman Sachs & Co 4,438,249 1,632
2 Morgan Stanley 4,092,201 1,841
3 JP Morgan 3,567,464 2,132
4 Bank of America Merrill Lynch 3,185,603 1,336
5 UBS 2,857,652 1,793 6 Citi 2,789,612 1,675 7 Deutsche Bank AG 2,461,598 1,898 8 Rothschild 2,450,079 2,541 Non-Top-Tier (9th to 25th) 9 Credit Suisse 2,412,826 1,635 10 Lazard 2,283,579 1,997 11 BNP Paribas SA 1,741,939 1,528 12 Nomura 1,394,781 805 13 RBS 1,022,875 1,395 14 Commerzbank AG 975,048 852 15 HSBC Holdings PLC 785,974 900 16 Societe Generale 721,864 822
17 Credit Agricole CIB 690,949 757
18 Mediobanca 587,842 518
19 Barclays Capital 460,646 301
20 Santander 428,185 419
21 KPMG 409,130 4,072
22 Greenhill & Co, LLC 341,380 174
23 Intese SanPaolo 320,539 281
24 PricewaterhouseCoopers 269,563 2,766
25 SEB Enskilda 242,431 544
one of the advisors fits in the top-8 cluster. This approach is standard in the literature (Rau, 2000; Servaes and Zenner, 1996; Golubov et al., 2012). In summary, we classify 1,382 transactions as advised by a top-tier investment bank, while 3,342 deals are advised by a non-top-tier financial advisor. A full description of the classification and the tracking of M&As among financial advisors themselves is described in Appendix B.
C. Abnormal returns
This thesis studies the bidder announcement returns of companies acquiring public, private, or subsidiary firms. The bidders are listed companies domiciled in the EU 15. We divide the samples in several sub samples, e.g. EU 14 transactions and UK transactions, furthermore we examine separately the sub samples determined by the target status (listed, private, or subsidiary). The listed status of the bidders implies that we can observe the share price of the bidder around an announcement of an M&A transaction. This announcement is usually the first signal to the market that the bidding firm is intended to acquire another company, which translates in a share price movement by the market. We refer to this phenomenon as the announcement return. We measure the acquirer’s return compared to the market return in order to exclude general market movement effects on the return. The result is called the cumulative abnormal return (CAR), which is the announcement return excluding the effect of general market movements measured over a few days.
A standard event study methodology is used in which the cumulative abnormal return (CAR) is calculated for each bidding company in a five day time fence (-2, +2) around the announcement. In conjunction with Fuller et al. (2002) we apply the market adjusted return method and calculate the CAR using the following formula:
, 2 2
∑
− = = t it i AR CAR (1)where CARi is the cumulative abnormal return of the bidding firm i for the five days adjacent to the acquisition announcement (-2, +2), and AR, known by:
, mt it
it R R
AR = − (2)
where ARit is the abnormal return of the bidder firm i at time t, computed by deducting the return of the market index (Rmt) from the return of the individual bidding company (Rit). The market index applied is the
where
CAR
is the average cumulative abnormal return for the total dataset of firms n. In order to investigatethe significance of the average CAR, which indicates an abnormal bidder announcement return comparative to the market index, the t-statistic is tested according the formula for an independent one-sample t-test:
,
n
CAR
t
σ
=
(4)where t is the statistic, n is the number of deals in the sample, and σ is the sample standard deviation. This t-test gives a direction on the significance of the potential abnormal returns compared to the market. Furthermore, the test assumes a normal distribution in each sample. Next, we consider whether the mean CAR of the acquiring companies which are advised by top-tier investment banks are significantly different from the mean CAR of the bidder companies advised by non-top-tier financial advisors by conducting a t-test assuming unequal sample sizes and variances which is given by:
, nTT TT CAR CAR nTT TT S CAR CAR t − − = (5)
where CARTT is the average cumulative return for the sample of acquiring companies advised by top-tier financial advisors,
CAR
nTT is the average cumulative return for the dataset consisting of firms advised by non-top-tier investment banks, andnTT TT CAR CAR S − is given by: , 2 2 nTT nTT TT TT CAR CAR n n S nTT TT
σ
σ
+ = − (6) where 2 TTσ
and 2 nTTσ
are the variances of the dataset with acquiring companies advised by top-tier investment banks respectively and the sample with bidder companies advised by non-top-tier investment banks, andn
TTand nnTT are the sample sizes of the two sets.
This parametric t-test assumes normality of returns. However, the Shapiro-Wilk test for normality indicates that the return data in this study follows a normal distribution. Therefore we apply the non-parametric Mann-Whitney U test in order to examine whether the median CARs of the sub samples significantly differ from each other.
D. Multivariate analyses
bidder- and deal-specific characteristics are controlled for; those are found to affect bidder returns as described in the previous section. Also, year fixed effects (coefficients suppressed) are taken into consideration. Next to that, heteroskedasticity-robust standard errors are used. Our main variable of interest is
top-tier. As indicated before, this is the indicator taking the value of one if a top-8 financial advisor advised on
the deal, and zero otherwise. The ordinary least squares regression assumes i) error term (ei) is independently
and identically distributed in i with a standard normal distribution; ii) {xi: i, …,n} is independent of ei; iii) the
error term (ei) and unmeasured latent continuous random variable (
δ
i) are independent with i.δ
iis normaldistributed and represents the unmeasured characteristics not captured by the independent variables. The OLS regression is specified as follows:
, i i i i X top tier y =
β
+γ
− +µ
(7) ), , 0 ( ~ 2 i I eσ
(8)where yi is the dependent variable bidder CARs, and X a vector of independent variables and control variables
(which are described in Table II), top-tier is the dummy variable indication whether the bidder company employed a top-tier investment bank, and
µ
i is the error term.In order to fully capture the interaction of target firm’s listing status and the payment method effect in the total sample analysis, we create 6 mutually exclusive categories: we have 3 types of target companies (listed, private, and subsidiary companies) and 2 methods of payment (all cash, and stock included). This methodology is in line with Golubov et al. (2012) and Masulis, Wang, and Xie (2007). In order to avoid multicollinearity with the intercept, the subsidiary stock deal indicator is excluded from the regression equations. For all regressions the Variance Inflation Factors (VIFs) are checked.
E. Heckman Two-Stage Procedure
In order to test our hypothesis we use a top-tier dummy variable as a right-hand-side variable. There is an obvious self-selection problem since bidder companies and top-tier banks choose each other. In order to deal with sample selection bias we apply the Heckman two-stage procedure. The main issue associated with the Heckman procedure is that we add a derived variable called the Inverse Mills Ratio as an extra independent variable in the second-step OLS regression. By examining the significance of the coefficients on the Inverse Mills Ratio we test the existence of a self-selection bias1.
For the OLS regression form of equation (7) and (8) and its estimates to be reliable, this setup implicitly requires that the independent variable top-tier is exogenous. Equation (7) cannot be consistently estimated by OLS if the independent variable top-tier is endogenous. Heckman (1979) showed that unreliable OLS estimates can be produced as a result of self-selection bias; the author developed a two-step procedure to control for it and presented in the classical paper of Heckman (1979). The model is utilized in this study to correct for selection bias in order to estimate the unbiased effects of independent variables on the dependent
variable. With respect to the model, the following ordinary least squares regression assumptions remain the same: i) the error term (ei) is independently and identically distributed in i with a standard normal distribution;
and ii){xi: I, …n} are independent of {ei: i=1, …, n}. The OLS assumption of independence between error
terms (ei) and unmeasured latent characteristics (
δ
i) in i is violated; therefore, unobserved determinants of thefinancial advisor selection can now influence the conclusion of the variable of interest.
The first equation models the choice between a top-tier investment bank and a non-top-tier financial advisor while the second-stage equations corrects for the selection bias. According to Li and Prabhala (2007) and Fang (2005) one preferably incorporates a variable that is included in the fist-stage equation while excluded in the second-stage equations. In line with Golubov et al. (2012) and Fang (2005) we include de variable prior relationship to serve as a so called “identification restriction”. Prior relationship measures to what extent the bidder company made use of the services of top-tier investment banks, including advice on M&As, equity issues, and bond issues during 5 years prior to the announcement. Furthermore, in line with Golubov et al. (2012) we exclude the dummy variable for tender offers and the premium variable from the selection equation. The authors argue that the acquisition technique and the valuations are normally found out by the financial advisor; the choice of the advisor is generally not determined by the acquisitions technique and the valuations.
For this purpose, a probit model is estimated as the error terms of this model is normally distributed (which is one of the assumptions underlying the Heckman model):
, i i Z tier top− =
δ
+ε
(9)where Zi is a vector of characteristics that has influence on the decision to employ a top-tier advisor or a non-top-tier financial advisor.
ε
i is the error term of the selection equation. Noteworthy, the variable top-tier is of a binary nature and was constructed to measure reputation:1 = − tier
top iff Zi
δ
+ε
i >0 and top− tier=0 iff Ziδ
+ε
i ≤0. (10)The OLS estimates resulting from equation (7) will be biased when
µ
i andε
i are correlated. It is shown by Heckman (1976) and Lee (1978) that when we replace equation (7) by:,
)
1
(
)
(
1
)
(
)
(
)
(
i i i i i i i i itop
tier
Z
Z
tier
top
Z
Z
X
y
ν
δ
δ
φ
ω
δ
δ
φ
ω
β
−
−
+
Φ
−
−
+
−
Φ
+
=
(11)(2012), Gande, Puri, and Saunders (1999), Gande, Puri, Saunders, and Walters (1997), and Puri (1996) make use of this model1. Golubov et al. (2012) uses it for considering the benefits and costs of investment bank involvement in US based M&As, while the latter three papers examine the entry of commercial banks to the bond issuing markets.
This setup described above can be generalized in order to create a model which is appropriate for examining the differences in the influence of bidder- and deal characteristics on the outcome variables between the two types of financial advisors. Campa and Kedia (2002), Lokshin and Sajaia (2004), Fang (2005), and Golubov et al. (2012) make use of a so called switching regression model with endogenous switching: equation (11) is replaced by two equations,
i i i X y1 =
β
1+µ
1 (12) i i X y2 =β
2 +µ
2 . (13)The first equation (12) is the equation for the group of transactions where a top-tier investment bank was employed, the second (13) for the non-top-tier advisors. We examine y1i or y2i, dependant on the type of advisor employed. Therefore,
i i y
y = 1 iff top-tieri =1 and yi =y2iiff top-tieri = 0. (14)
As indicated above, the allowance of correlation between the residuals of the selection and outcome equations (
ε
i andµ
1i(µ
2i)) leads to the modelling of endogeneity. Therefore, the outcome variable of interest is now influenced by unobserved determinants of the choice whether or not to employ a top-tier financial advisor. This yields the following non-diagonal covariance matrix:cov(
µ
1i,µ
2i,ε
i) =
1
2 1 2 22 21 1 12 11 ε ε ε εσ
σ
σ
σ
σ
σ
σ
σ
. (15)The result of equation (9) determines whether we observe equation (12) or equation (13). Therefore, the observed yi becomes a conditional variable as we never observe both of them; also, the error terms in these
equations do not have zero mean. However, the inverse Mill’s ratios,
1There is a handbook available (Handbook of Corporate Finance, Chapter 2) with a very thoroughly written description of the models,
) ( ) (Z i Z i i =
φ
÷Φλ
, for y1i , (16) )) ( 1 ( ) (Z i Z i i =−φ
÷ −Φλ
, for y2i , (17)where Z is the vector of independent variables;
φ
1 is the density probability function andΦ
1 is the cumulative probability function, adjusts equations (12) and (13) for the nonzero mean ofµ
1i andµ
2i. The equations can subsequently be used for OLS regressions1.This methodology is based on the Heckman two-stage procedure (1979) and applied in the studies of Golubov et al. (2012), Fang (2005), Dunbar (1995), and Lee (1978). The first two studies examine investment banking reputation in relation to the quality and price of bond underwriting services and in relation to quality of M&A advice, respectively.In order to further find out what the impact is of the employment of top-tier investment bank on yi we
compare the outcome of the model when a particular deal was advised by top-tier advisors with the potential outcome when a non-top-tier financial advisor was applied. Ever after, so far we only observe that a certain deal is guided by a top-tier investment bank or by a non-top-tier financial advisor. More practically, as mentioned above, we execute two second-stage regressions (since we observe outcomes for both selected and non-selected groups, e.g. top-tier and non-top-tier) both with Inverse Mills ratios. Then we perform a linear prediction for both categories using coefficient estimates of the opposite category. Lastly, we subtract the predicted value from the actual value; this will give us the hypothetical improvement of using a top-tier investment bank or a non-top-tier financial advisor. Econometrically, the hypothetical improvement can be estimated by first evaluating Xi in the alternative advisor equation
[
y
2iTop
−
tier
i=
1
]
=
E
[
X
i 2+
u
2iZ
i+
i>
0
]
E
β
δ
ε
Φ
+
+
=
)
(
)
(
)
,
cov(
2 2 2δ
δ
ϕ
ε
β
i i i i i iZ
Z
u
u
X
E
. (20)For comparison reasons, we measure the difference between the actual outcome and the hypothetical value
[
y
iTop
tier
i]
y
iE
2−
=
1 −
1 . (21)
1 We applied the described methodology using Eviews. The program we have written is depicted in the Appendix. Alternatively, the same approach could be followed using Stata or SPSS (see e.g. Smith (2003) where the following formulas are given to calculate the Inverse Mills Ratios:
) ( ) 5 . 0 (exp( ) 2 / 1 ( i i i i i π ξ ξ φ ξ λ = ⋅ ⋅ − ⋅ ⋅ ÷ , for top-tier = 1 (18) )) ( 1 ( ) 5 . 0 (exp( ) 2 / 1 ( i i i i i π ξ ξ φ ξ λ =− ⋅ ⋅ − ⋅ ⋅ ÷ − , for top-tier = 0, (19)
