DO PRIVATE EQUITY FUNDS BENEFIT FROM THEIR RELATIONSHIPS WITH FINANCIAL ADVISORS IN M&A TRANSACTIONS?

Transcription

1 DO PRIVATE EQUITY FUNDS BENEFIT FROM THEIR RELATIONSHIPS WITH FINANCIAL ADVISORS IN M&A TRANSACTIONS? STEFAN MORKOETTER THOMAS WETZER WORKING PAPERS ON FINANCE NO. 2015/15 SWISS INSTITUTE OF BANKING AND FINANCE (S/BF HSG) AUGUST 2015

2 Do Private Equity Funds Benefit from their Relationships with Financial Advisors in M&A Transactions? Stefan Morkötter * and Thomas Wetzer ** August 2015 Abstract Financial advisors play an important role in M&A transactions. Private equity (PE) funds, in turn, are highly sought-after clients for financial advisors since they promise lucrative business due to their frequent engagements in acquisitions. Empirical research shows us that PE funds pay, on average, lower take-over premiums as compared to strategic acquirers. But do relationships between PE funds and their financial advisors have any impact on the size of these premiums? We show that PE funds indeed pay, on average, less for their portfolio companies when the sell-side advisors have worked for the acquiring PE fund on the buy-side in past transactions. We refer to this as indirect relationships. Our findings are stronger for larger PE firms with a high level of deal activity. Strategic acquirers do not benefit from these previous indirect relationships. At the same time, close relationships with their own buy-side advisors (i.e., always hiring the same investment bank) do not pay-off for both PE-related and strategic acquirers. We base our findings on acquirer-advisor relationship information of 53,552 M&A transactions with pricing information of 11,438 deals. Keywords: JEL Codes: Private Equity, Mergers and Acquisitions, Financial Advisors, Take-over Premiums G15, G24, G32, G34 * Morkötter: University of St.Gallen, [email protected]. ** Wetzer: University of St.Gallen, [email protected]. This paper was previously circulated under the title Private Equity Discounts in M&A Transactions Relationships Matter! We thank Manuel Ammann, Martin Brown, Roland Füss, Tim Jenkinson, Conor Kehoe, Markus Schmid, and the seminar participants of the First International St.Gallen PhD Conference in Singapore (2014), the 2014 and 2015 PhD in Finance Seminars at the University of St.Gallen, and the EFMA Annual Conference 2015 for helpful comments.

3 1 Introduction Recent studies have focused on the role of financial advisors in M&A transactions. Most research agrees that the choice of the financial advisor matters to the success of a transaction and to the post-transaction performance (e.g., Golubov et al., 2012). A different string of literature deals with the M&A performance of private equity (PE) firms 1 vs. strategic acquirers and concludes that PE firms manage to achieve a price discount (e.g., Bargeron et al. 2008). Despite the existing research both on PE transaction pricing and financial advisors, we know very little about the role of financial advisors in price discussions. In contrast to most strategic M&A players, PE firms are active acquirers. They are usually involved in numerous deals per year which allow them to acquire profound deal making expertise as well as relevant relationships with their financial advisors. Financial advisors are important decision makers in the deal making process and are highly interested in successful deal completion as they are remunerated based on feebased models: buy-side and sell-side financial advisors 2 only receive transaction fees, if transactions are completed. This creates a conflict of interest for buy-side financial advisors, as they only receive their fees in case their clients (acquirers) win the bid, which is usually when they are willing to pay the highest price. Prosperous relationships with PE firms and strategic acquirers positively correlate with more fee income for financial advisors. Acquirers, on the other hand, seek to minimize advisor fees while maintaining successful relationships with their financial advisors to obtain valuable target information. We investigate whether the high deal flow of PE firms (and their superior deal making capabilities over strategic acquirers) leads to more beneficial/lower transaction prices. Financial advisors support both acquirers and sellers throughout (friendly or hostile) transaction processes. If a buyer aims to acquire a company (target), usually both sides are advised by financial advisors. Sellers and buyers have to pay (significant) fees for these services. A question arising in this context is the value of these services. The existing financial literature offers some explanations and empirical evidence on the performance of financial advisors in M&A transactions. However, so far it has mainly centered on the question of whether (highquality) advisors lead to a better post-m&a performance of acquirers (see Golubov et al., 2012). Against this background, we investigate if acquirers in M&A transactions benefit from strong relationships with their financial advisors. For the purpose of this paper, acquirer benefits 1 In this paper, the terms general partner (GP) and private equity (PE) firm have identical meanings unless stated otherwise. 2 In this paper, we will refer to buy-side and sell-side advisors depending on whether advisors receive their mandates from acquirers or targets/sellers. Sell-side advisors advise the targets in a transaction. 1

4 are defined as lower transaction prices in comparison to similar transactions taking place at the same time. As a measure to define lower transaction prices, we use the EV/EBITDA multiple paid by acquirers. We distinguish between PE acquirers and strategic buyers. Acquisitions of portfolio companies are at the core of the PE business model, whereas strategic buyers (mainly) pursue transactions to optimize their product portfolio or to achieve inorganic growth. In addition, we introduce two types of acquirer-advisor relationships: a direct acquirer-advisor relationship exists between the acquirer and the financial advisor an acquirer hires for a particular transaction. The more often an acquirer has hired the same financial advisor in the past, the stronger this relationship. An indirect relationship in turn exists between the acquirer and the financial advisor of the sell-side. The more often an acquirer has hired the sell-side agent in the past as his own advisor in M&A transactions, the stronger is this indirect relationship. In our empirical analysis, we link these direct and indirect relationships PE firms and strategic acquirers maintain with financial advisors to the M&A performance of these acquirers. For this, we analyzed deal level advisory relationships for ~53,552 transactions undertaken by ~ 15,000 different acquirers (~2,000 PE firms and ~13,000 strategic buyers) that were advised by ~2,800 financial advisors. We collected all relationship information from Capital IQ and Thomson One and linked the information to a deal database also consisting of Capital IQ and Thomson One data. This deal database covers 1,004 PE deals and 10,434 strategic deals and includes information on deal characteristics, target characteristics, and, most importantly, targets EV/EBITDA multiples. Our findings confirm that PE firms benefit from strong relationships with their financial advisors, but only in the case of indirect relationships. In line with existing research (Golubov et al., 2012), we find that PE funds generally do not benefit from strong direct relationships with their own financial advisors in terms of transaction pricing. As outlined before, this comes as no surprise: the remuneration of buy-side financial advisors depends on the deal completion. The lower the price advisors are suggesting to their clients, the less likely their offer will be accepted by the target side and deal completion becomes unlikely. However, we find empirical proof that transaction prices decrease when a PE firm buys a company which is advised by a financial advisor with whom the PE firm has worked together on the buyer s side in the past (indirect relationship). This effect is mainly driven by larger PE firms with a high level of deal activity and intensifies the stronger the relationship is. We do not observe any of these effects for strategic buyers even when only taking a subsample of large, very active acquirers. Financial literature on the performance of financial advisors in M&A transactions has focused on the acquirer and target stock performance following an acquisition or the announcement and linked it to the role of (reputable) financial advisors. This hypothesis is often referred 2

5 to as the skilled-advice hypothesis. Most studies have stipulated that acquirers do not benefit from reputable advisors (e.g., McLaughlin, 1990, Rau, 2000). In an early study, Bowers and Miller (1990) found that advisors are good at identifying synergy potential but they often fail to capture these synergies. In a recent study, Golubov et al. (2012) outline the fact that top-tier advisors only deliver higher bidder returns than their non-top-tier competitors in public transactions where the reputation and skill set of advisors is larger (see also Ismail, 2010). They also find that advisors find it harder to capture gains in public acquisitions as public targets possess greater bargaining power than their unlisted counterparts. Servaes and Zenner (1996) compare deals in which acquirers are advised by external financial advisors with those solely managed by inhouse teams and conclude that hiring financial advisors (and paying them high fees) does not directly lead to any benefits for buyers. Bao and Edmans (2011) document significant investment bank fixed effects and conclude that financial advisors do matter for M&A transactions. Song et al. (2013) compare the M&A performance of boutique advisors versus full-service banks and find that investment boutiques are more likely to be hired in complex deal structures. In addition, they outline that deal premiums are lower when boutique advisors are involved. Fang (2005) reminds us that the selection of an advisor is not purely random. She explains this through the reputation effect. Some acquirers will only consider hiring reputable financial advisors (top-tier advisors), thus systematically ruling out a large base of financial advisors from their selection pool. Gompers et al. (2012) support Fang s idea that business partners are not chosen purely based on rational factors. They find that fund managers prefer to work together if they share either the same ability-based characteristics (e.g., university background) or the same affinity-based characteristics (e.g., ethnic background). Collaboration based on mutual ability-based characteristics increases fund performance while affinity-based characteristics have the opposite effect. Francis et al. (2012) find that existing banking relationships (e.g., lending business) do not influence acquirers choice of financial advisors. They also detect that active acquirers are more likely to switch their financial advisors following a poor deal outcome. Recent studies (e.g., Achleitner et al., 2011, and Guo et al., 2011) suggest that PE funds manage to buy portfolio companies cheap (with low EV/EBITDA multiples 3 ) and sell them dear (with high multiples). This so-called multiple arbitrage is achieved either by multiple surfing (buying at the low of a cycle and selling at the peak) or by multiple engineering (proving to the market that the firm is now worth more) or both. The literature suggests numerous reasons for this PE discount. Strategic acquirers are confident that they will generate synergies with their 3 In this paper, we refer to EV/EBITDA multiples when we use the term EBITDA multiples or simply multiples. 3

6 investments which encourage them to pay more for acquisition targets as long as operational improvement levers exist. Experience in price negotiations and deal making are also seen as reasons why PE firms manage to achieve price discounts. Deal pricing also varies depending on the deal setup and fund characteristics according to existing research (Achleitner et al., 2011). Axelson et al. (2013) argue that (the availability of) leverage drives buyout pricing levels up. Achleitner et al. (2011) add to the discussion and show that entry pricing levels are positively correlated with industry-wide public market valuations. Acharya et al. (2013) show that GPs with a banking background are particularly successful in transactions that involve significant M&A acquisitions. Based on our findings, we contribute to the existing literature on the role of financial advisors in M&A transactions, most notably by introducing indirect relationship networks between acquirers and financial advisors. In addition, we add to the literature analyzing the performance of PE firms in take-over situations. By doing so, we link acquirer-advisor relationships to transaction pricing and show that these relationship structures have an impact on prices paid by PE firms. To prove this we compare transaction multiples (EV/EBITDA multiples). This is an innovative approach: to our knowledge, few studies have focused on the relative deal-making performance of PE funds based on EV/EBITDA multiples. Most of the existing literature has applied abnormal returns experienced over a period of time prior to the transaction date. It has concluded that public firms are willing to pay a higher price in M&A transactions as compared to PE buyers (e.g., Bargeron et al., 2008). We suggest that the price discount over comparable strategic acquirers comes (partially) from the way PE firms efficiently manage their relationships with financial advisors. Their high deal activity means they are particularly attractive clients and financial advisors will be trying hard to maintain good working relationships with them. The remainder of this paper is organized as follows: Section 2 introduces the role and relationships of advisors in M&A transactions. Section 3 presents the data sample and explains the methodology. Section 4 displays the empirical results, which are discussed in Section 5. Finally, Section 6 concludes. 2 Financial Advisors in M&A transactions 2.1 Role of financial advisors In M&A transactions, financial advisors usually guide the involved parties through the entire transaction process. The selling party, the target, as well as the acquirers involved in a deal rely on their expertise and know-how. Financial advisors are typically investment banks, universal banks, investment boutiques, or consultancy firms. In general, their principle tasks in a trans- 4

7 action are a) to offer recommendations on the target's fair value and improve the quality of acquirer and target matches, b) to guide their clients through the acquisition process, and c) in case advisors are banks, to even act as credit lenders to acquirers by offering financing packages (Mortensen, 1982, Servaes and Zenner, 1996, and Chahine and Ismail, 2007). Often transaction parties employ more than one financial advisor depending on the advisors expertise and value proposition in each of the three tasks. Financial advisors collect valuable information on the target and its industry, reduce information asymmetry, and, thus, add value for their clients (Allen et al., 2002; Servaes and Zenner, 1996). Advisors can reuse data on a target obtained in previous deals and therefore often hold exclusive information. There is a relatively small group of (large) financial advisors who have developed a reputation of being transaction experts and who dominate M&A advisory services in particular for large acquisitions. This group mainly consists of so-called top-tier banks, which offer their advisory services on a worldwide scale. Their worldwide operations also give them access to a lot of industries and transaction structures. Interestingly, this competitive information advantage over smaller financial advisors (e.g., domestic investment boutiques) does not translate into higher bidder returns in M&A transactions (McLaughlin, 1990). It appears that smaller advisors in turn benefit from their local knowledge and networks. In the specific case of transactions undertaken by private equity funds, PE fund managers are usually very familiar with transaction processes as it is their daily business and many fund managers start their professional careers working as M&A advisors. Due to their lean organizational set-up, PE funds also rely on financial advisors as they lack internal resources (e.g., human resources) to work on the transactions on their own. PE firms require financial advisors for industry insights, execution of due diligence, acquirer and target matching (especially when a PE firm does a first-time investment in a specific industry), and credit lending. The remuneration of financial advisors is (to a large extent) performance-based, usually depends on the transaction size, and, most importantly, is only paid in case the transaction actually takes place (contingent fees). This is bad news for buy-side advisors, since in most M&A situations more than one party is bidding for the target. However, not all buy-side advisors involved in a transaction will be paid. Only the advisor working for the winning bidder will be. Sell-side advisors have the advantage of not usually facing such competition. Thus, the probability that sell-side advisors receive their fees tends to be higher: usually there is more than one bidder and sell-side advisors receive their fees independently of who the target is sold to (as long as it is actually sold). The size of advisor fees varies significantly and seems to have increased in recent years. Hunter and Jagtiani (2002) assumed on average USD 2.3mn for buy-side advisors (0.4% of transaction value) and USD 4.4mn for sell-side advisors (0.8% of transaction value) for 5

8 the period Chahine and Ismail (2007) reckon that financial advisors received USD 31bn in fees in 2005 out of USD 2.7tn in total M&A transaction volume, which suggests advisor fees of 1.2%. Buy-side advisor fees are usually only imposed when the buying party wins the bid. It is therefore obvious that buy-side advisors are highly interested in closing the deal for their clients. Understanding the remuneration policy of financial advisors will be important later in this paper when we discuss potential reasons for our results. Various researchers have analyzed the role of advisor fees in transactions. Hunter and Jagtiani (2002) find that the payment of larger advisory fees does not play an important role in determining the likelihood of completing a deal. Golubov et al. (2012) find that top-tier advisors receive a fee premium of 0.25% in absolute terms. Chahine and Ismail (2007) add that when acquirers pay higher fees than target firms, they acquirers pay lower premiums. 2.2 Acquirer-advisor relationship frameworks We are interested in the intensity of existing acquirer-advisor relationships in order to answer our main research question: do acquirers benefit from strong relationships with financial advisors? For our empirical analysis, we differentiate between direct and indirect acquireradvisor relationships. In the following, we will refer to a relationship as a direct relationship when an advisor is directly hired as an agent by an acquirer (buy-side mandate). With respect to acquirers, we will distinguish in the following between PE buyers (financial sponsors) and strategic buyers. An indirect relationship, on the other hand, means that an advisor works on the sellside of a transaction (offering advice to the target/seller company) and is not hired by the acquiring PE or strategic buyer for that specific deal (no buy-side mandate); yet the advisor has been hired by the buyer in the previous five years as a buy-side agent at least once. By introducing indirect relationships, we add to the existing literature focusing on financial advisors. Such existing research has largely been limited to direct acquirer-advisor relationships. By focusing on indirect relationships, we aim to fill a gap in the literature. We know from Fang (2005) and Golubov et al. (2012) that acquirer-advisor matching is not random. Due to the reputation effect, some acquirers prefer to hire reputable advisors and advisors with specific industry experience. This creates an endogeneity problem that we cannot eliminate completely. Nor can we fully take into account that fund managers might choose some financial advisors due to mutual ability and/or affinity-based characteristics (see Gompers et al., 2012). However, we address this endogeneity issue in robustness tests by running separate re- 6

9 gressions for a sample of top-tier and sample of non-top tier advisors as well as by controlling for industry fixed effects and industry-related deal activity. Based on the direct and indirect relationships described above, we define two relationship frameworks that exist between acquirers, target companies, and financial advisors both for PE and strategic buyers. With regard to the time period measuring the intensity of the relationships between an acquirer and an advisor, we assume that five years before a deal is reasonable. This approach is in line with existing literature (e.g., Francis et al., 2012) which also focuses on fiveyear relationships. High-ranked employees, usually partners in case of PE buyers or C-Level executives in case of strategic buyers, are the main contacts to advisors. These partners or executives typically tend to work for their employers for at least five years. In case of financial sponsors, a total period of five years also corresponds to the investment period of typical PE funds which is usually 3-5 years. Since large parts of a GP s income are attributed to the carried interest of the underlying fund, it is also reasonable to assume that partners stay with one PE firm for at least one fund s lifetime (e.g., ten to twelve years). The more often an acquirer has been working together with the same financial advisor, the stronger the direct acquirer-advisor relationship. Relationship (1): acquirers' relationships with their financial advisors (buy-side mandate) five years before deal D [Insert Figure 1A about here] The first relationship in our focus represents a direct acquirer-advisor relationship: we wish to know how many times an acquirer has hired the same financial advisor in the five years preceding a deal. Case example: KKR buys Unisteel on 23/09/2008 (date effective). Morgan Stanley advises KKR on the deal. How many times in the five years prior to 23/09/2008 did Morgan Stanley advise KKR? Can financial advisors put pressure on the selling party if they have particularly strong relationships with buyers? Do acquirers benefit from always having the same financial advisor? We investigate whether the direct relationship between financial advisors and their clients has any influence on pricing levels in M&A transactions (measured by enterprise value/ EBITDA multiples paid by acquirers). Relationship (2): acquirers' relationships with their financial advisors (buy-side mandate) five years before deal D if advisors worked with acquirers' targets (sell-side mandate) in t=0 [Insert Figure 1B about here] Relationship 2 is what we call an indirect relationship between the acquirer and the target's advisor. It focuses on how many times an acquirer was advised by a specific financial advi- 7

10 sor when acquiring a company in the five years before a specific deal if during the deal itself this advisor has a sell-side mandate and advises the target/seller and not the acquirer. Case example: KKR buys Unisteel on 23/09/2008. Macquarie Bank advises Unisteel on the deal (sell-side mandate). How many times in the five years prior to 23/09/2008 did Macquarie Bank advise KKR in a M&A transaction? Similarly with relationship 1, we count the amount of times an advisor advised an acquirer in the five years prior to a deal. But the constellation at the deal at t=0 is different: at t=0 the advisor is on the sell-side and not advising the acquirer yet, due to prior direct advisory mandates, a relationship has been established between these two parties. The more often an acquirer has been working in a direct relationship together with the financial advisor of the sell-side party, the stronger the indirect acquirer-advisor relationship. This type of relationship framework sheds light on the question of whether financial advisors have any incentive to push their clients for discounts when acquirers are involved in a deal with which they have maintained (strong) direct relationships in the past. This could result in a conflict of interest situation as the financial advisors are hoping for future business from acquirers. It could also result in trust creation with potential acquirers (e.g., the financial advisor knows that the acquirer is trustworthy and quick in executing the transaction and thus advises its client to accept the offer). 3 Data & Methodology 3.1 Data sample & statistics For a better understanding of our relationship analyses, it is important to keep in mind that we essentially worked with two different databases: the 53,552 M&A transactions (both PEbacked and strategic deals) in the relationship database for which we have information on (i) the deal completion date, (ii) the acquirer name, (iii) the acquisition/target advisor name, and (iv) the advisor mandate (buy-side or sell-side). In a second step, we collected deal, pricing, and target information of the transactions included in our relationship database. As this information was not available for all M&A transactions of our relationship database, we ended up with 11,438 M&A transactions (both PE-backed and strategic deals) for which we found detailed information on the underlying transactions, most importantly target EV/EBITDA multiples 4. Thus, the 11,438 deals are an intersection of the 53,552 deals included in our relationship database. In other words, for 21% of the 53,552 deals we also have deal, pricing, and target information. 4 Other information includes target financial information (ROA, leverage, net income, etc.), target characteristics (name, industry, region, public/private status, etc.) and deal characteristics (negotiation period, deal attitude, % of shares acquired, etc.). 8

11 Relationship database We drew our information on acquirer-advisor relationships from Capital IQ and Thomson One. As we were primarily interested in information on investment date, acquirer name, and advisor name, advisor mandate (buy-side or sell-side) we could use a broad range of deals 11,478 PE and 42,074 strategic transactions, i.e., 53,552 deals in total. Building up on our acquireradvisors relationships (see section 2.2), for each of the deals we matched acquirers with either buy-side and/or sell-side advisors. We could then quantify the intensity of advisor relationships: we counted how many times an acquirer had worked with an advisor in the five years prior to each of these deals. In a next step, we linked these deals from the relationship database with the M&A transactions of our deal database wherever possible. We ended up with 15,643 matches for the 11,438 deals (for some deals we found both the acquisition and the target advisors): for the strategic deals, we matched 7,356 deals for relationship 1 and 7,174 deals for relationship 2. For our PE deals we managed to match 698 deals for relationship 1 and 415 deals for relationship 2. Table 1 provides a detailed overview on our acquirer-advisor relationships (53,552 transactions in total). PE firms are involved in 6.8 deals on average, while strategic firms are involved in 3.1 deals. In the five years prior to a deal, a PE firm will have worked on average 1.0 times with that advisor on the buy-side (direct relationship). A strategic firm will have worked 0.5 with its advisor. When the advisor is on the sell-side in a specific deal (indirect relationship), a PE/strategic firm will have hired an advisor 0.2/0.1 times in the five years before that deal. There are some acquirers that maintain exceptionally strong relationships with their advisors: they worked on up to 32 deals together in the time period of five years (column E of Table 1). [Insert Table 1 about here] Table 2 adds practical evidence to our descriptive statistics on acquirer-advisor relationships. The tables show our league tables of acquirers and advisors that are most actively involved in PE and strategic acquisitions according to our relationship database 5. Table 2A shows that 3i is the PE firm that was most active in the investment period 1985 to 2013 (97 acquisitions). In 16% of its deals it issued the advisory mandate to PricewaterhouseCoopers (PWC). KPMG was the advisory firm that acted most often as the target advisor in acquisitions led by 3i (8% of all 5 This league table might not coincide with league tables from other sources as most other league tables use deal volume rather than deal activity. As outlined above, we do not have information on the deal volume for all the deals in our relationship database, thus, we relied on the absolute account. In terms of deal volume, 3i, for example, might not be the most active acquirer, as it might do on average much smaller deals than other PE players (e.g., KKR). The same observation holds for financial advisors: KPMG is more involved in smaller deals and not so much in larger transactions. 9

12 deals). Advisory firms that are not typical investment banks or investment boutiques, such as KPMG and PWC, engage in numerous, but usually relatively small, deals. They would surely not have appeared in the league table had we taken transaction size volume as a measure (see league tables of Golubov et al, 2012). The example of BNP Paribas shows us that whenever a PE firm is part of a larger financial corporation/universal bank, advisory mandates are usually issued internally. Siemens AG, a German multinational, is top of the league table among our strategic acquirers (44 acquisitions). It conducted 18% of all its acquisitions with Credit Suisse and found JP Morgan negotiating for the sell-side in 14% of all deals. In unreported analyses, we found that smaller strategic acquirers tend to diversify their advisor mandates much less. They usually work with the same advisor for all their acquisitions. Table 2A and 2B make clear that large acquirers tend to rely on the services of large international investment banks, such as the likes of Morgan Stanley, Merrill Lynch, and Goldman Sachs. There are only 13 different names in the list of the favorite advisors of the 20 largest acquirers (PE and strategic). Also the group of target advisors seems to be rather small: of the target advisors that the 20 largest acquirers negotiated with most there are also only 13 different names. These observations support existing literature that the group of high-quality advisors is fairly small (Golubov et al. 2012). [Insert Table 2 about here] We mentioned above that advisors are particularly useful to acquirers when they have exclusive knowledge in a particular industry. As a result, deal activity league tables look differently depending on the industry we are looking at. However, we observe that a selected group of toptier financial advisors are active in all industries. Figure 2 shows that there are even some top-tier advisors (e.g., Goldman Sachs, Morgan Stanley) who manage to be on the very top of the deal activity league tables in almost all industries. Others such as Deutsche Bank and especially KPMG seem to focus on specific industries: consumer products and industrials appear to be the main focus of KPMG, while far fewer deals are covered by KPMG in energy and healthcare. This observation also holds when we control for geographic regions. All top-tier advisors maintain a global footprint, yet some banks appear to be stronger in some regions (e.g., Deutsche Bank in Europe). It appears that acquirers are aware of advisors industry and regional expertise and do not randomly hire their advisors independent of the industry the deal is taking place. For the purpose of our regressions, it was therefore important to control for target industry/region when analyzing our acquirer-advisor relationships. [Insert Figure 2 about here] 10

13 We follow up on recent research on the impact of top-tier advisors on M&A transactions (Golubov et al., 2012). Golubov et al., following the approach of Fang (2005), define top-tier advisors based on size of transaction value. The eight advisors with the highest transaction value are top-tier advisors. As we lack transaction value data for many of our deals, we defined top-tier advisors based on deal activity instead. However, six of our eight top-tier advisors (mostly investment banks) are also found in Golubov et al s list. Only UBS and KPMG are not in their list (KPMG traditionally covers numerous smaller deals and, thus, does not rank high in terms of transaction volume). We find in unreported deal descriptive statistics that targets differ in terms of characteristics when top-tier advisors are involved in contrast to when non-top-tier advisors are involved: enterprise and transaction values are significantly higher, negotiation periods are longer (interestingly, Hunter and Jagtiani (2002) find the opposite), and leverage is higher. In the following empirical analysis, we therefore conducted tests to check the robustness of our results with regard to the impact of top-tier advisors. Deal database Of the 11,438 individual deals in our database, 1,004 are PE deals and 10,434 are strategic deals. With respect to PE deals, we focused on entry deals (deals in which a PE firm buys a target from a strategic seller). PE deals were sourced from Capital IQ and Thomson One. Strategic deals were sourced from Thomson One. For the PE deals, we manually ensured that we had no redundant deals in the deal list. We followed existing literature and removed deals with negative EV/EBITDA multiples (approx. 300 deals) to exclude pure restructuring cases from our data sample (see for example Achleitner et al., 2011). We also deleted real estate firms, financial institutions, and targets from the public services sector (approx. 600 deals) due to deal peculiarities in these three industries. We made sure that financial sponsors only included PE firms and excluded hedge funds and other financial sponsors such as sovereign wealth funds (approx. 400 deals). Moreover, we removed all deals in which we could not clearly identify the acquirer as a PE firm by matching them to the Preqin general partners list (approx. 350 deals) 6. We also ensured only completed deals are included in our list (removed approx. 350 canceled and announced deals). Lastly, we deleted any kind of repurchases and self-tenders (approx. 100 deals). Table 3 gives an overview of our data sample. Consumer product deals make up the largest industry group both in PE and strategic deals (34% and 22%). North America is by far the biggest market (43% and 47% of all deals). We collected deals from all over the world includ- 6 Preqin is a database focusing only on the financial sponsor industry with a particular focus on private equity. It maintains the most exhaustive list of general partners and has been widely used in financial literature for empirical research on private equity. 11

14 ing 13% emerging market deals to cover the full global deal spectrum. Most of our M&A transactions are majority takeovers (85% and 90%). Also, the deal attitude of most deals is friendly - both for PE and strategic buyers (94% and 89%). The number of listed targets is only slightly lower among our PE target group than among strategic targets (84% and 88%). [Insert Table 3 about here] The existing M&A literature argues that deals vary significantly by transaction and financial characteristics. PE and strategic deals differ significantly in terms of these characteristics and transaction prices are highly affected by them. Target profitability (though often negative) and total assets are systematically significantly larger in PE deals while transaction values are only slightly larger (e.g., Fidrmuc et al., 2012). It also makes substantial difference whether toptier advisors or non-top-tier advisors are involved in a deal. Top-tier advisors tend to cover larger deals than their less reputable counterparts (Golubov et al., 2012). Also other characteristics such as target industry (e.g., Fidrmuc et al., 2012) differ according to deal types. In our regression analysis, we therefore controlled for any deal and target characteristics that could affect our actual relationship analyses. We selected our deal control variables in line with existing M&A literature (e.g., Bargeron et al., 2008, Fidrmuc et al., 2012). Appendix 2 shows that these characteristics indeed differ significantly by deal type also in our dataset. Besides enterprise value, target industry, and profitability, we also controlled for negotiation period (time from deal announced to deal effective) (e.g., Bargeron et al., 2008), deal attitude (e.g., Flanagan and O'Shaugnessy, 2003), public/private target (e.g., Bargeron et al., 2008), and majority/minority takeover. Furthermore we controlled for regions, and investment years (Madura et al., 2012). 3.2 Regression models We used a standard OLS base regression model with various control variables as well as fixed effects to address the main research question of our paper 7. All regressions use the same dependent variable log(ev/ebitda multiples) and main independent variables (relationships 1 and 2) but vary in terms of data sample. In all regressions, we controlled for the seven target and deal characteristics C i described above to avoid selection bias regarding PE vs. strategic deals: enterprise value, ROA, leverage, negotiation period, deal attitude, listed target dummy, and majority takeovers. Target industry, region, and investment year were our fixed effects. Target industries were grouped based on SIC Codes, NAIC Codes, and overall company business descriptions (please refer to Appendix 1 for a detailed description of each control variable). In our base- 7 In this chapter, we only describe the regressions that are part of the main part of this paper. 12

15 line regression (see Table 4) we included the control variable advisor industry deal activity over total advisor deal activity to control for advisor industry expertise. In all regression models we estimated standard errors using the Huber-White sandwich estimators (Huber, 1967). This method allowed us to conduct OLS regressions with heteroscedasticity-consistent standard errors. To address the main research question of whether advisor relationships impact transaction prices in PE and/or strategic deals (Table 4), we used the following baseline regression: 7 log(m) = β 0 + β 1 R + β 2 II + i=1 β 3i C i + i=1 β 4i FF i + ε i, (3.1) where log(m) is the log of the EV/EBITDA multiple (see also Achleitner et al., 2011). The two acquirer-advisor relationships (R1 and R2) serve as independent variables R (one at a time analyzed in separate regressions) as we are interested in the impact of these relationships on deal multiples. IE is the industry expertise of financial advisors (defined as advisor industry deal activity over total advisor deal activity) that we also control for. In subsequent robustness checks, we removed the control variable IE and instead took subsamples of a) low/high acquirer deal activity, b) low/high financial advisor deal activity (both Table 5), and c) total funds raised in the last ten years (Table 6). Subsamples in all three regressions are roughly of equal size. These subsamples allow insights into whether the results of Table 4 are driven by deal activity and/or acquirer size. In all analyses we split the samples into PE deals and strategic deals to investigate the impact of the relationships on the two acquirer types separately. 3 4 Empirical Results We know from existing literature (e.g., Allen et al., 2002) that advisors have the power to utilize their information gathering to influence purchasing prices. However, our data shows that both PE and strategic acquirers do not benefit from strong direct relationships with their financial advisors in price negotiations: all regressions on relationship 1 in Table 4 suggest that there is no significant impact of direct acquirer-advisor relationships on transaction pricing. These results hold even when we control for the industry expertise of financial advisors (regression 2, 4, and 6). We believe this is not surprising as we expect a conflict of interest for financial advisors on the buy-side: acquisition advisors only receive their fees if the deal is executed (while target advisors get their fees no matter to which buyer the target is sold). The lower the price, the more likely the target will not agree to the deal, especially if there is more than one bidder involved. 13

16 When examining indirect relationships (R2) we see for our total deal sample that acquirers do not benefit from strong indirect relationships. However, a comparison of the deal types reveals that PE firms benefit from seeing familiar faces on the target advisor side while strategic firms do not: the more often PE firms have hired the sell-side advisors of a particular transaction themselves in the five years prior to a deal, the lower the target purchasing price. For an increase of the indirect relationship variable of one, we expect to see an 11% 8 drop in transaction multiple. These results are significant at the 5% level. As discussed earlier, industry expertise is one of the main decision criteria when hiring an advisor and might therefore bias our results. However, our results remain significant with almost the same economic magnitude even when controlling for industry expertise (Table 4, Column (8)). Our results imply that PE firms do not have to rely on their advisors industry experience when looking for lower M&A pricing levels. Advisors with different degrees of industry focus can grant this discount. We find no such discounts for strategic deals or the total sample of acquirers. One explanation relates to the notion of certainty: if a sell-side advisor knows the acquirer is likely to complete the deal (and there is usually a large chance that PE firms will close a deal as acquiring companies is their business model), he might be more inclined to close the deal with this acquirer. Another explanation is linked to potential future business with the acquirer: target financial advisors might push for a lower target price in order to remain in good terms with the PE firm that is seeking to acquire the target. Advisors do not want to jeopardize these fragile relationships with their PE clients (especially the larger ones with high deal flow). There exists empirical evidence suggesting that advisors treat PE buyers favorably in order to liaise with them in the long-term. Francis et al. (2012) find that financial advisors are becoming more and more aggressive in retaining existing clients and in winning future ones. [Insert Table 4 about here] Since we know that PE firms benefit from indirect relationships with financial advisors, we now focus on whether this applies to all PE acquirers and/or all types of financial advisors or only to a selected sample. By controlling for advisor industry deal activity over total advisor deal activity, we already established that advisor industry expertise plays no (major) role. In a second step, we controlled for deal activity of acquirers and total deal activity of advisors (see Table 5). Columns (1) (8) of Table 5 separates acquirer deal activity into two subsamples: low deal activity (1-5 deals) and high deal activity (more than 5 deals). We observe that in particular very exp(-0.109) =

17 active PE firms benefit the most from strong indirect relationships, whereas this is not true for very active strategic buyers. In addition, we find that with regard to the direct relationship variable, individual deal activity of the acquirer does not change the results of our first analysis: strong direct acquirer-advisor relationships do not influence the pricing levels of M&A transactions. Smaller and larger PE firms both do not seem to benefit from strong direct relationships with their financial advisors. Very active PE firms were involved in 181 or 43% of our PE deals. Unsurprisingly, larger PE firms are more attractive deal partners for advisors as they promise more deal flow and, thus, more fee income. Columns (9) (16) of Table 5 show similar results but this time for advisor deal activity. In order to separate very active advisors from less active financial advisors, we created two subsamples based on the overall deal activity in our relationship database of more than 50,000 transactions: low advisor deal activity (1-50 deals) and high advisor deal activity (more than 50 deals). We selected deal activity by each advisor as a measure to assess the experience of individual financial advisors. Our results show that acquirers again do not benefit from strong direct relationships with their financial advisors even when hiring very active, and thus experienced, advisors. With regard to the indirect relationships, we lose significance for PE acquirers in case of less active advisors. This result corresponds to our findings stated in the previous section, as large PE firms, in particular, benefit from strong indirect relationships. Large PE firms tend to engage in larger M&A transactions, which in turn are dominated by larger and more active financial advisors. In unreported regressions, we again picked up the results of Golubov et al. (2012) and checked whether our results differ when we create subsamples for top-tier advisors and non-top-tier advisors. Our results remain significant for both subsamples. [Insert Table 5 about here] When we redefine PE acquirer size as total funds raised in the last 10 years, we get a similar picture (see Table 6): smaller PE firms (total funds raised less than the sample median of USD 3.8bn) do not benefit from their indirect relationships with financial advisors, while larger PE firms (total funds raised more than the sample median of USD 3.8bn) do benefit and pay lower transaction prices. [Insert Table 6 about here] 15

18 We found that a) PE firms benefit from indirect advisor relationships (and strategic acquirers do not) and b) of these PE firms only the larger ones can secure this advantage. As our results show that the intensity of previous relationships (0, 1, 2, 3, relationships) has an impact on the pricing level paid by PE firms in M&A transactions, we were interested in whether the sheer existence of such relationships (previous relationship yes or no) leads to the same results. This analysis also served as an additional robustness test for our results as impact on transaction prices should be nil for all deals in which there were no previous relationships between PE acquirers and target advisors, which is what we observe in the regression analysis of Appendix 3. We ran binary regressions with our relationship variable being 1 if there had been a previous relationship (at least two) between acquirers and advisors and 0 otherwise (see Appendix 3). The outcome shows that our results were significant at the 1% level. Thus, our results seem to be driven by deals in which there have been at least two interactions between acquirers and advisors in the five years before the deal. So far our regressions have centered on the question of whether previous relationships between acquirers and advisors have any impact on transaction pricing. Consequently, we also checked whether future relationships have any impact: acquirers might reward the low price they paid for a target by hiring the target advisors in the future. For this, we swapped around dependent and independent variables (new independent variable: EV/EBITDA multiple; new dependent variable: the relationship variable) and kept the deal samples separated into PE and strategic. If future relationships mattered, then any decrease in transaction pricing should have led to an increase in acquirer-advisor relationships. Appendix 4 shows that this was not the case neither for PE firms nor for strategic acquirers. Robustness check: sample selection of M&A transactions All our analyses are based on the assumption that the selection of our M&A transactions for the multivariate analysis is unbiased. We faced the difficulty of detailed deal level information not being available for all transactions (e.g., EV/EBITDA multiples). However, we required deal-level and company-level control variables to ensure that we did not simply capture size or deal-type effects. As an additional robustness check we therefore checked whether the overall discount PE acquirers experience in M&A transactions also holds for a larger data sample of transactions for which, however, we lack the underlying acquirer-advisor relationship information, but do know the pricing levels (see Appendix 5). This approach allowed us to incorporate in particular a larger number of PE transactions than we used for our main analysis in order to overcome a potential selection bias. Based on a data sample of 15,433 deals including 16

19 transactions undertaken both by PE and strategic acquirers, we show that the PE discount for the larger, more comprehensive deal sample (-0.229/ ) is comparable to the deal sample ( / ) we use for our main analysis (see columns (1) & (3) vs. (2) & (4) in Appendix 5). Thus, we show that the PE-backed M&A deals (for which we were able to document the financial advisors relationships) are representative for a broader PE-driven M&A activity. 5 Discussion of Results Why do PE firms, and not strategic firms, benefit from strong relationships with financial advisors that are advising them in M&A transactions? As suggested above, the notion of certainty that a particular PE firm is more likely to complete a deal than other bidders might be an important reason. The more target advisors are familiar with buy-side PE firms, the more precisely they know when a PE firm will push for deal completion. Anecdotal evidence suggests that advisors even systematically calculate the likelihood of deal completion based on previous deals they worked on together with an acquirer (e.g., Morgan Stanley knows that KKR completes 90% of the deals the firm is bidding for). This information is likely to be communicated to the target company which will then be more inclined to sell to a PE firm even if this results in a price discount in comparison to a strategic buyer. No such information exists for most strategic acquirers as their deal flow is much lower and they usually buy for operational reasons. Target financial advisors have an incentive to encourage their clients to sell to the acquirer with the highest probability of deal completion: only upon deal completion advisors will receive their fees. There are in fact strong reasons why the probability of deal completion should be higher when a PE firm is involved. One reason is better access to financing (see Demiroglu and James, 2010). Another reason is the presence of antitrust regulations: PE firms are generally not seen as a regulatory threat as they acquire rather than merge operations. Strategic acquisitions are more strictly monitored by regulators (e.g., Anti-Monopoly Office (AMO) and other antitrust divisions and cartel offices). Target firms and advisors might be cautious to sell to a strategic acquirer if there is the possibility that the deal will be blocked by the regulators. This is also a potential reason why we see significance in our findings for PE firms but not for strategic acquirers, especially with high acquirer deal activity: KKR is unlikely to become a higher regulatory threat even after completing a large number of deals; but Telefonica SA, as an example, might become a more interesting target for antitrust divisions as they complete more deals (especially if these deals all take place in one region). Target financial advisors that usually work with the PE buying party (relationship 2) might also grant these buyers exclusive access to the deal (proprietary deal 17

20 sourcing) in which competitive bids are non-existent and lower prices can therefore be pushed through more easily. We do not question the fact that financial advisors strive to provide best services for their clients but conflicts of interest might arise as financial advisors are highly interested in maintaining long-lasting relationships with their PE clients. This is not because PE firms pay higher fees than their strategic peers. On the contrary, PE firms are particularly prudent in keeping fees low as they are involved in a large number of deals. A more likely reason is the sheer deal volume of PE firms. PE firms provide highly lucrative business for M&A advisors as they are constantly looking for new targets. Most PE funds have a closer look at targets at least four times per year and deal completion is often likely, while most strategic acquirers only buy occasionally. Furthermore, they are interested in financing their deals with high levels of leverage for which they will need debt from banks. Against this background, one may argue that the more favorably advisors treat PE firms, the more likely they are to work with them in the future. In addition, we provide evidence that PE firms indeed proactively manage their relationships with financial advisors. Aware of their importance to advisors, PE firms systematically maintain more diversified advisor portfolios than their strategic counterparts in order to keep advisor competition high. This allows them to constantly look for the best fee terms and highest quality. Despite our empirical findings that suggest a significant relationship between PE firmadvisor ties and transaction pricing, there are certainly also arguments that speak against these findings. Advisory is primarily a trust business. Why should banks jeopardize their reputation and feed rumors that they treat PE firms better than strategic acquirers? Financial advisors will lose credibility if they agree to fix deals at lower prices just to ingratiate themselves with the PE firms. Moreover, it is not clear whether advisors actually prefer to work with PE firms over strategic firms. It is true that, on average, they are more active in the M&A market but they often also pay lower advisor fees. The additional income for banks through commercial loans often does not compensate for the lower fees they receive from PE firms. 6 Conclusion We build on existing literature on the role of financial advisors in M&A transactions and on the price discounts achieved by PE firms in these transactions. Our data shows that the intensity of indirect acquirer-advisor relationships has a negative effect on the pricing level paid by PE firms in transactions. PE firms exploit value from their advisor relationships if a financial advisor, whom they have hired regularly in the past as a buy-side advisor, is on the opposite side of a transaction advising the sell-side. Large and very active PE firms benefit the most from 18

21 these indirect relationships. Strategic buyers in turn even the ones that are as active as PE firms in the M&A market are not able to benefit from these indirect relationships with financial advisors. Why may financial advisors grant a discount to PE acquirers with whom they worked closely in the past? Besides a potential conflict of interest, a major explanation is that sell-side financial advisors depend on a high probability of deal execution, which PE firms are usually more likely to satisfy than strategic acquirers. We do not find such results for direct acquireradvisor relationships. We believe this is due to the contingent fee system in which buy-side advisors are paid only when their clients win their bids (i.e., when they are willing to pay the highest price). 19

22 References Acharya, V. V., O. Gottschalg, M. Hahn, and C. Kehoe (2013): Corporate Governance and Value Creation: Evidence from Private Equity, The Review of Financial Studies 26, Achleitner, A.-K., R., Braun, and N., Engel (2011): Value creation and pricing in buyouts: Empirical evidence from Europe and North America, Review of Financial Economics 20, Allen, L., J. Jagtiani, S. Peristiani, and A. Saunders (2002): The Role of Bank Advisors in Mergers and Acquisitions, Journal of Money, Credit and Banking 36, Axelson, U., T. Jenkinson, P. Strömberg, and M. S. Weisbach (2013): Borrow Cheap, Buy High? The Determinants of Leverage and Pricing in Buyouts, Journal of Finance 68, Bao, J. and A. Edmans (2011): Do Investment Banks Matter for M&A Returns?, The Review of Financial Studies 24, Bargeron L. L., F. P. Schlingemann, R. M. Schulz, and C. J. Zutter (2008): Why do private acquirers pay so little compared to public acquirers?, Journal of Financial Economics 89, Bowers, H. M. and R. E. Miller (1990): Choice of investment banker and shareholders wealth of firms involved in acquisitions, Financial Management 19, Chahine, S. and A. Ismail (2007): Premium, merger fees and the choice of investment banks: A simultaneous analysis, The Quarterly Review of Economics and Finance 49, Degeorge, F., J. Martin, and L. Phalippou (2014): On Secondary Buyouts, ECGI, Working Paper. Demiroglu, J. and C. M. James (2010: The role of private equity group reputation in LBO financing, Journal of Financial Economics 96, Fang L. H. (2005): Investment Bank Reputation and the Price and Quality of Underwriting Services, Journal of Finance 60, Fidrmuc J. P., P. Roosenboom, R. Paap, and T. Teunissen (2012): One size does not fit all: Selling firms to private equity versus strategic acquirers, Journal of Corporate Finance 18, Flanagan, D. J. and K. C. O'Shaugnessy (2003): Core-related acquisitions, multiple bidders and tender offer premiums, Journal of Business Research 56, Francis, B., I. Hasan, and X. Sun (2012): Does Relationship Matter? The Choice of Financial Advisors, Bank of Finland Research Discussion Paper

23 Golubov A., D. Petmezas, and N. G. Travlos (2012): When It Pays to Pay Your Investment Banker: New Evidence on the Role of Financial Advisors in M&As, Journal of Finance 67, Gompers, P., V. Mukharlyamov, and Y. Xuan (2012): The Cost of Friendship?, NBER Working Paper Series. Guo, S., E. S. Hotchkiss, and W. Song (2011): Do Buyouts (Still) Create Value?, Journal of Finance 66, Huber, P. J. (1967): The behavior of maximum likelihood estimates under nonstandard conditions, Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Hunter, W. C. and J. Jagtiani (2002): An analysis of advisor choice, fees, and effort in mergers and acquisitions, Review of Financial Economics 12, International Monetary Fund (IMF): World Economic Outlook, October Ismail, A. (2010): Are good financial advisors really good? The performance of investment banks in the M&A market, Review of Quantitative Finance and Accounting 35, Leleux, B., H. Van Swaay, and E. Megally (2015): Private Equity 4.0 Reinventing Value Creation. Wiley Finance Series. Madura, J., T. Ngo, and A. M. Viale (2012): Why do merger premiums vary across industries and over time?, The Quarterly Review of Economics and Finance 52, McLaughlin, R. M. (1990): Investment-banking contracts in tender offers: An empirical analysis, Journal of Financial Economics 28, Mortensen, D. (1982): Matching as a non-cooperative bargaining game. In J. J. McCall (Ed.), The Economics of Information and Uncertainty. Chicago: University of Chicago Press. Rau, R. P. (2000): Investment bank market share, contingent fee payments, and the performance of acquiring firms, Journal of Financial Economics 56, Servaes, H. and M. Zenner (1996): The role of investment banks in acquisitions, Review of Financial Studies 9, Song, W., J. Wei, and L. Zhou (2013): The value of "boutique" financial advisors in mergers and acquisitions, Journal of Corporate Finance 20,

24 Figure 1: Acquirer-advisor relationships Figure 1 illustrates the two relationship frameworks that we are analyzing. Relationship 1 (Figure 1A) counts the number of times an acquirer was advised by a specific financial advisor when acquiring a company in the five years prior to t=0 (direct relationship). Relationship 2 (Figure 1B) looks at the number of times an acquirer was advised by a specific financial advisor when acquiring a company in the five years prior to t=0. At the deal itself, however, this advisor advises the target (indirect relationship). Figure 1A: Relationship 1 (R1) - DIRECT relationship Acquirer A buys target T & financial advisorf advises acquirer A -5 years t = 0 Figure 1B: Relationship 2 (R2) - INDIRECT relationship Fin. advisor F advises acquirer A Acquirer A buys target T -5 years t = 0 Financial advisor F advises target T 22

25 Table 1: Relationship statistics Table 1 provides background figures on our acquirer-advisor relationships. Columns A-C display information on the number of deals, acquirers and acquirers-advisor relationships. Note that not all deals in the relationship database can be linked to our deal database, thus we do not know the EV/EBITDA multiples of all these deals. Columns D-F show information on the number of deals per advisor with the same acquisition/target advisor in the five years prior to t=0 (average, maximum, 95 th percentile). Column G shows many of our relationship database deals could be linked to the deal database. Relationships Acquirer type R1: Relationships with own financial advisors in previous 5 years R2: Relationships with own financial advisors in previous 5 years that advise targets in t=0 A B C D E F G Deals (total relationship database) Acquirers (total relationship database) Average deals/ acquirer Deals/acquirer with same advisor (average) Deals/acquirer with same advisor (maximum) Deals/acquirer with same advisor (95th percentile) Deals (matches with deal database) PE 6, Strategic 19,442 6, ,356 PE 5, Strategic 22,632 6, ,174 Total PE 11,478 1, na na na 1,113 Strategic 42,074 12, na na na 14,530 23

26 Table 2: Most active acquirers and their advisors Table 2 provides descriptive information split in two tables on the top acquirers in our relationship database and their advisors. Table 3A lists the most active PE firms and the advisors that they worked most closely with. Example: we counted 97 acquisitions completed by 3i. PricewaterhouseCoopers advised 3i in 16% of these acquisitions, while KPMG advised 8% of the targets in these acquisitions (when 3i was acquiring). Table 3B follows the same methodology as Table 3A but lists the top strategic acquirers of our relationship database. Table A - Advisor relationships of top 10 GPs GENERAL PARTNER HEADQUARTER ACQUISITIONS STRONGEST ACQUISITION ADVISOR RELATIONSHIP % of total acquisitions STRONGEST TARGET ADVISOR RELATIONSHIP 1 3i United Kingdom 97 PricewaterhouseCoopers 16 KPMG 8 2 Blackstone Group United States 80 Deutsche Bank 21 Goldman Sachs 9 3 Carlyle Group United States 80 Credit Suisse 14 JP Morgan Chase & Co 8 4 Bridgepoint Advisers United Kingdom 69 N M Rothschild&Sons 17 Ernst & Young, N M Rothschild & Sons 4 5 Kohlberg Kravis Roberts United States 68 Credit Suisse, Morgan Stanley 22 Citigroup, Credit Suisse, JP Morgan Chase & Co 9 6 CVC Capital Partners United Kingdom 63 UBS, Deutsche Bank 14 Goldman Sachs 10 7 Equistone Partners Europe/ United Kingdom Barclays Private Equity 57 KPMG 23 KPMG, Lazard 5 8 BNP Paribas Private Equity France 55 BNP Paribas 71 JP Morgan Chase & Co, Morgan Stanley 4 9 Apax Partners United Kingdom 49 Merrill Lynch 16 UBS, JP Morgan Chase & Co 8 10 Ardian/AXA Private Equity France 49 Deloitte&Touche, PricewaterhouseCoopers 10 Ernst & Young, HSBC, N M Rothschild & Sons 4 AVERAGE % of total acquisitions Table B - Advisor relationships of most active strategic acquirers STRATEGIC ACQUIROR HEADQUARTER ACQUISITIONS STRONGEST ACQUISITION ADVISOR RELATIONSHIP % of total acquisitions STRONGEST TARGET ADVISOR RELATIONSHIP 1 Siemens AG Germany 44 Credit Suisse 18 JP Morgan Chase & Co 14 2 Telefonica SA Spain 38 Morgan Stanley 21 Citigroup 13 3 Vodafone Group United Kingdom 36 UBS 39 Merrill Lynch, Morgan Stanley 11 4 Coca-Cola HBC AG Switzerland 35 UBS 14 JP Morgan Chase & Co, Morgan Stanley 11 5 General Electric United States 35 Goldman Sachs 26 Goldman Sachs 23 6 NTT DOCOMO Inc Japan 35 Nomura Holdings Inc 17 Mizuho Bank Ltd 9 7 Procter & Gamble Co United States 33 Goldman Sachs 64 Goldman Sachs 21 8 Tyco International Ireland/Switzerland 32 Merrill Lynch 16 Morgan Stanley 19 9 Itochu Corp Japan 32 GCA Savvian Advisors LLC 25 Daiwa Securities Co Ltd, Mizuho Bank Ltd Schneider Electric SA France 30 Merrill Lynch 33 Credit Suisse 13 AVERAGE % of total acquisitions 24

27 Figure 2: League tables of the 10 most active advisors by industry Figure 2 ranks the most active financial advisors for a) all industries and b) split by industry. We define deal activity based on the number of deals these acquirers were involved with buy-side or sell-side mandates according to our relationship database. The labels on the lines refer to the deal activity ranking of the financial advisors in the respective industries Consumer Telecommunications All industries products Energy Healthcare Industrials Materials Technology Goldman Sachs Morgan Stanley Credit Suisse JP Morgan Chase & Co Merrill Lynch Lazard KPMG UBS Deutsche Bank Citigroup 25 25

28 Table 3: Deal summary statistics Table 3 includes summary statistics of all 11,438 PE and strategic deals between 01/01/1985 and 31/07/2013 in our data sample. All deals listed in our deal databases have a direct link to our relationship database, i.e., we know the acquisition and/or target advisor of these deals. Only deals with positive EV/EBITDA multiples are included in the data sample. Deals by industry are deal target industries, which are a combination of SIC Codes, NAIC Codes and overall company business descriptions (real estate firms, financial institutions, and targets from the public services sector are excluded). Developed markets include USA, Canada, Western Europe, Japan, and Australia/New Zealand. Emerging markets include: Asia (excl. Japan), Africa, Eastern Europe, Latin America, and Middle East. We define developed/emerging markets according to the criteria of the International Monetary Fund (2014). Majority takeovers are deals in which the acquirer purchased at least 51% of the target. Friendly takeovers are deals in which the deal attitude was flagged as 'friendly'. Listed targets are deals in which the target companies were publicly listed in one or more stock exchanges. Data for PE deals was consolidated from three different sources: Capital IQ and Thomson One. Data for strategic deals was collected from Thomson One only. Redundant deals are excluded. See Appendix 1 for variable definitions. PRIVATE EQUITY (PE) STRATEGIC TOTAL Acquisitions (%) Acquisitions (%) Acquisitions (%) DEALS 1,004 10,434 11,438 DEALS BY INDUSTRY Consumer products 337 (34%) 2,286 (22%) 2,623 (23%) Energy 54 (5%) 1,377 (13%) 1,431 (13%) Healthcare 84 (8%) 803 (8%) 887 (8%) Industrials 194 (19%) 1,599 (15%) 1,793 (16%) Materials 91 (9%) 1,379 (13%) 1,470 (13%) Technology 142 (14%) 1,614 (15%) 1,756 (15%) Telecommunications 102 (10%) 1,376 (13%) 1,478 (13%) DEALS BY REGION North America (NA) 434 (43%) 4,886 (47%) 5,320 (47%) Western Europe (WE) 366 (36%) 2,598 (25%) 2,964 (26%) Rest of world (RoW) 204 (20%) 2,950 (28%) 3,154 (28%) MARKETS VS. EMERGING MARKETS Developed markets (DM) 868 (86%) 9,040 (87%) 9,908 (87%) Emerging markets (EM) 136 (14%) 1,394 (13%) 1,530 (13%) MAJORITY TAKEOVERS Yes 852 (85%) 9,348 (90%) 10,200 (89%) No 152 (15%) 1,086 (10%) 1,238 (11%) FRIENDLY TAKEOVERS Yes 946 (94%) 9,318 (89%) 10,264 (90%) No 56 (6%) 1,086 (10%) 1,142 (10%) LISTED TARGETS Yes 841 (84%) 9,162 (88%) 10,003 (87%) No 161 (16%) 1,232 (12%) 1,393 (12%) TOP TIER ADVISOR Yes 358 (27%) 4,619 (27%) 4,977 (27%) No 977 (73%) 12,411 (73%) 13,388 (73%) DEALS BY SOURCE Thomson One 622 (62%) 10,434 (100%) 11,056 (97%) Capital IQ 382 (38%) - (-) 382 (3%) 26

29 Table 4: Impact of strong financial advisor relationships Table 4 displays the impact of our two acquirer-advisor relationships on transaction pricing (dependent variable: EV/EBITDA multiples) based on pooled ordinary least squares (OLS) regressions. Independent variables are the two types of advisor relationships (R1 and R2) we are investigating on. We run the regressions for a) PE deals only, b) strategic deals only, and c) for the PE and strategic deals taken together. In addition we run the regressions a) without the control variable advisor deal activity / total advisor deal activity and b) with the variable. In all regressions we control for our key deal characteristics (winsorized at the 1% significance level). We take fixed effects for industry, investment region, and investment year into account. Numbers in the upper rows represent the regression coefficients; numbers in brackets in the lower row represent respective standard errors. *, ** and *** indicate p-values of 10%, 5%, and 1% significance level, respectively. See Appendix 1 for variable definitions. Dependent variable: log(ev/ebitda multiple) PE PE Strategic Strategic Total Total PE PE Strategic Strategic Total Total Independent variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) R1: Relationships with buy-side financial advisors (5 years before a deal) (0.014) (0.014) (0.008) (0.008) (0.007) (0.007) R2: Relationships with buy-side financial advisors (5 years before a deal) that are on the target-side in t= ** ** (0.050) (0.050) (0.031) (0.031) (0.028) (0.028) Advisor deal activity by industry / total advisor deal activity No Yes No Yes No Yes No Yes No Yes No Yes Deal characteristics controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Constant 1.703*** 1.659*** 1.756*** 1.718*** 1.780*** 1.735*** 1.973*** 2.059*** 1.871*** 1.936*** 1.878*** 1.949*** (0.240) (0.256) (0.125) (0.130) (0.114) (0.121) (0.383) (0.404) (0.150) (0.162) (0.140) (0.152) Observations ,356 6,961 8,054 7, ,174 6,722 7,589 7,114 R-squared Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 27

30 Table 5: Impact of strong financial advisor relationships - by acquirer & advisor deal activity Table 5 is closely related to Table 4. It also displays the impact of acquirer-advisor relationships on transaction pricing for our PE and our strategic deals sample based on pooled ordinary least squares (OLS) regressions. Table 5 applies subsamples to investigate whether acquirer and/or advisor deal activity drive our result. Columns (1)-(8) group acquirer deal activity into a) <= 5 deals and >5 deals. Columns (9)-(16) group advisor deal activity into a) <=15 deals and >15 deals. Again, we control for our key deal characteristics (winsorized at the 1% significance level). We take fixed effects for industry, investment region, and investment year into account. Numbers in the upper rows represent the regression coefficients; numbers in brackets in the lower row represent respective standard errors. *, ** and *** indicate p-values of 10%, 5%, and 1% significance level, respectively. See Appendix 1 for variable definitions. Dependent variable: log(ev/ebitda multiple) 1a) Acquirer deal activity: 1-5 deals 1b) Acquirer deal activity: >5 deals 2a) Advisor deal activity: 1-50 deals 2b) Advisor deal activity: >50 deals PE Strategic PE Strategic PE Strategic PE Strategic PE Strategic PE Strategic PE Strategic PE Strategic Independent variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) R1: Relationships with buy-side financial advisors (5 years before a deal) R2: Relationships with buy-side financial advisors (5 years before a deal) that are on the target-side in t=0 (0.059) (0.020) (0.011) (0.010) (0.050) (0.017) (0.016) (0.009) *** ** (0.715) (0.064) (0.042) (0.036) (1.185) (0.198) (0.040) (0.028) Deal characteristics controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Constant 1.806*** 1.713*** 4.107*** 1.815*** 1.147** 2.554*** 1.381*** 2.637*** 2.101*** 1.546*** 2.209*** 1.525*** 1.189*** 2.157*** 1.867*** 2.189*** (0.329) (0.138) (1.475) (0.170) (0.455) (0.201) (0.528) (0.220) (0.476) (0.186) (0.422) (0.177) (0.289) (0.172) (0.477) (0.207) Observations 363 6, , , , , , , ,995 R-squared Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 28

31 Table 6: Impact of strong financial advisor relationships - by acquirer & advisor deal activity Table 6 is closely related to Table 5. This time we created subsample for PE firms with total funds raised in the last 10 years of a) less than USD 3.8bn (which is the median of all total funds raised) and b) more than USD 3.8bn. We control for our key deal characteristics (winsorized at the 1% significance level). We take fixed effects for industry, investment region, and investment year into account. Numbers in the upper rows represent the regression coefficients; numbers in brackets in the lower row represent respective standard errors. *, ** and *** indicate p-values of 10%, 5%, and 1% significance level, respectively. See Appendix 1 for variable definitions. Dependent variable: log(ev/ebitda multiple) PE Strategic PE Strategic Independent variables (1) (2) (5) (6) R1: Relationships with buy-side financial advisors (5 years before a deal) R2: Relationships with buy-side financial advisors (5 years before a deal) that are on the target-side in t=0 Total funds raised last 10 yrs below median (USD 3.8bn) Total funds raised last 10 yrs above median (USD 3.8bn) (0.016) (0.014) *** (0.218) (0.053) Deal characteristics controls Yes Yes Yes Yes Fixed effects Yes Yes Yes Yes Constant 1.538*** 1.694*** 1.852*** 2.976*** (0.366) (0.449) (0.251) (0.545) Observations R-squared Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 29

32 Appendix Appendix 1: Variable definitions Appendix 1 presents definitions for the variables used in this paper. When consolidating our databases, we paid great attention to making sure that variable definitions were the same in across all our databases. Variable Deal specifications EV/EBITDA-multiple Private equity (PE) deal PE investment Strategic deal Developed market deal Emerging market deal Transaction value Industry Region Investment year Portfolio company statistics log(enterprise value) ROA Description Ratio of the target's enterprise value (see definition below) and its EBITDA (for the last 12 months ending on the date of the most current financial information prior to the transaction). A private equity firm (GP) is the acquirer and/or the seller of a target company. For our deal sample, we identified PE firms either through their primary NAIC description or their primary VEIC code, and/or if they were listed as PE firms in the Preqin database. Note that we excluded deals involving other financial sponsors, such as hedge funds, from our data sample. All our PE deals are realized deals. A PE firm is the acquirer of the target and there is no PE firm on the target/seller side. Any deal in our data sample in which no PE firm is involved, i.e., where the target is purchased for strategic reasons only. All our strategic deals are realized deals. The target company is located in a developed market country. Our paper follows the developed market definition of the International Monetary Fund (IMF). The target company is located in an emerging market country. Our paper follows the emerging market definition of the International Monetary Fund (IMF). Total value of consideration paid by the acquirer, excluding fees and expenses in USD. Industries are categorized based on SIC Codes, NAIC Codes and overall company business descriptions. Our deal sample includes Consumer Products, Energy, Healthcare, Industrials, Materials, Technology, Telecommunications. Real Estate industry and Finance industry are excluded. Deals are grouped into 10 regions: Africa, Asia (excl. China & Japan), Australia & NZ, China, Eastern Europe, Japan, Latin America, Middle East, North America, Western Europe. In most cases we further aggregated this into 3 main groups (North America, Western Europe, Rest of World) as nearly 75% of our deals take place in North America or Western Europe. Deal effective year of our deal. We include deals between 01/01/1985 and 31/07/2013. Log of the target company's enterprise value at deal announcement in USD; winsorized at the 1%-level. Target company's return on asset of the last 12 months ending on the date of the most current financial information prior to the announcement of the transaction (LTM) - displayed as percentage and winsorized at the 1%-level. Re- 30

33 turn on assets is the ratio of net income (LTM) and total assets (LTM). Winsorized at the 1%-level. Leverage Majority takeover Negotiation period Friendly takeover Target is listed Advisor information Financial advisor Buy-side advisors Sell-side advisors Acquirer-advisor relationships R1: Relationships with own financial advisors in previous 5 years R2: Relationships with own financial advisors in previous & future 5 years that advise targets in t=0 Ratio of target company's total debt of the last 12 months ending on the date of the most current financial information prior to the announcement of the transaction (LTM) and its enterprise value at announcement. Winsorized at the 1%- level. The acquirer purchased at least 51% of the target. Time elapsed between deal announced date and deal effective date. Winsorized at the 1%-level. Deal attitude was explicitly friendly (as opposed to hostile, friendly-to-hostile, neutral, etc.). Target was publicly listed in one or more stock exchanges. Advisor that advised one of the parties on the deal's financial matters. Financial advisor that advised the deal's acquiring party. Financial advisor that advised the deal's target/seller. Number of times an acquirer was advised by a specific financial advisor when acquiring a company in the previous 5 years of t=0. Number of times an acquirer was advised by a specific financial advisor when acquiring a company within the previous 5 years of t=0. At the deal itself, this advisor advised the target. 31

34 Appendix 2: Portfolio company statistics - PE vs. strategic acquirers Appendix 2 compares deal and target characteristics by deal types: a) PE investments and b) strategic investments. All statistics are at deal announcement and are winsorized at the 1% significance level. Negotiation period is the time elapsed between deal announced date and deal effective date. Return on assets (ROA) is the ratio of total income over total assets. Leverage is the ratio of Total Debt over Enterprise Value. Financial statement statistics are as of last-twelve-months. All variables include positive figures only (except ROA). We performed a t-test on the mean difference between private equity and peer group deals. In the 'Mean (t-test)' Column, *, ** and *** indicate p- values of 10%, 5%, and 1% significance level, respectively. Statistics in Table 2 are not exhaustive - our data sample includes a large variety of further data. See Appendix 1 for more detailed variable definitions. ENTERPRISE VALUE (EV)/EBITDA MULTIPLE Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment 1, *** Strategic investment 10, Total 11, ENTERPRISE VALUE (USD mn) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment 1,004 1,012 *** 297 2, ,608 Strategic investment 10,434 1, , ,027 Total 11,438 1, , ,670 TRANSACTION VALUE (USD mn) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment *** 170 1, ,240 Strategic investment 10, , ,703 Total 11, , ,572 NEGOTIATION PERIOD (Days) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment 1, *** Strategic investment 10, Total 11, EBITDA (USD mn) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment *** Strategic investment 10, Total 11, RETURN ON ASSETS (ROA) (%) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment Strategic investment 10, Total 11, TOTAL ASSETS (USD mn) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment *** 250 2, ,497 Strategic investment 10,359 1, , ,584 Total 11,291 1, , ,386 LEVERAGE (%) Observations Mean Mean difference (t-test) 5th Median Std. Dev. percentile 95th percentile PE investment ** Strategic investment 8, Total 9,

35 Appendix 3: Impact of strong financial advisor relationships (dummy analysis) Appendix 3 is closely related to Table 4 in the main section of this paper. It also displays the impact of our acquireradvisor relationships on transaction pricing (dependent variable: EV/EBITDA multiples) based on pooled ordinary least squares (OLS) regressions. However, independent variables are modifications of our relationships: we are not counting the number of times a PE firm was advised by a specific financial advisor when acquiring a company but only if the GP was advised by the advisor before t=0 or not. The relationship dummies are 1 if there have been at least two relationships prior to the deal and 0 otherwise. We control for our key deal characteristics (winsorized at the 1% significance level). We take fixed effects for industry, investment region, and investment year into account. Numbers in the upper rows represent the regression coefficients; numbers in brackets in the lower row represent respective standard errors. *, ** and *** indicate p-values of 10%, 5%, and 1% significance level, respectively. See Appendix 1 for variable definitions. Dependent variable: log(ev/ebitda multiple) PE PE Independent variables (1) (2) R1 dummy (1 if at least two previous interaction, 0 otherwise) (0.069) R2 dummy (1 if at least two previous interaction, 0 otherwise) *** (0.120) Deal characteristics controls Yes Yes Fixed effects Yes Yes Constant 1.688*** 1.962*** (0.239) (0.382) Observations R-squared Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 33

36 Appendix 4: Impact of EV/EBITDA multiples on future financial advisor relationships Appendix 4 proves whether transaction prices paid in a deal has any impact on future relationships between acquirers and advisors based on pooled ordinary least squares (OLS) regressions. Relationships 1 and 2 are our dependent variables. The EV/EBITDA multiple is the independent variable. We run the regressions separately for our two sample PE and strategic. We control for our key deal characteristics (winsorized at the 1% significance level). We take fixed effects for industry, investment region, and investment year into account. Numbers in the upper rows represent the regression coefficients; numbers in brackets in the lower row represent respective standard errors. *, ** and *** indicate p-values of 10%, 5%, and 1% significance level, respectively. See Appendix 1 for variable definitions. Dependent variable: Relationships R1 and R2 R1 R2 PE Strategic PE Strategic Independent variables (1) (2) (3) (4) EV/EBITDA multiple (0.180) (0.013) (0.045) (0.004) Deal characteristics controls Yes Yes Yes Yes Fixed effects Yes Yes Yes Yes Constant * 0.562** *** (1.023) (0.237) (0.331) (0.055) Observations 698 7, ,174 R-squared Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 34

37 Appendix 5: Regression results on EV/EBITDA multiples Appendix 5 presents the results of pooled ordinary least squares (OLS) regressions on the log of EV/EBITDA multiples for PE and strategic investments for the period 1985 to We include two groups of samples in our regressions. Columns marked as Relationship sample' include those deals for which we know the acquirer-advisor relationships. Columns marked as Control sample includes deals for which we do not necessarily know the advisor-acquirer relationships but we know all other relevant information to conduct the regressions (EV/EBITDA multiple, deal characteristics, fixed effects). By showing that we yield the same significances for both sample groups, we demonstrate that our reduced relationship sample does not carry a selection bias. Regressions (1) and (2) show the effect of PE investments on the log(ev/ebitda multiples) without controlling for our deal characteristics (winsorized at the 1% significance level). Regressions (3) and (4) show the effect of PE investments on the log(ev/ebitda multiples) with controlling for our deal characteristics (winsorized at the 1% significance level). We take fixed effects for industry, investment region, and investment year into account. Numbers in the upper rows represent the regression coefficients; numbers in brackets in the lower row represent respective standard errors. *, ** and *** indicate p-values of 10%, 5%, and 1% significance level, respectively. See Appendix 1 for variable definitions. Dependent variable: log(ev/ebitda multiple) Relationship sample Control sample Relationship sample Control sample Independent variables (1) (2) (3) (4) PE Investments *** *** *** *** (0.026) (0.020) (0.026) (0.020) log(enterprise value) 0.085*** 0.088*** (0.005) (0.004) Return on assets *** *** (0.001) (0.001) Leverage *** *** (0.000) (0.000) Majority takeover (0.027) (0.021) Negotiation period ** ** (0.000) (0.000) Friendly takeover (0.027) (0.020) Target is listed *** *** (0.031) (0.024) Fixed effects Yes Yes Yes Yes Constant 1.637*** 1.547*** 1.849*** 1.904*** (0.112) (0.103) (0.117) (0.101) Observations 11,438 18,588 9,672 15,433 R-squared Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 35