Conflict of Interest and the Credibility of M&A Advisor Recommendations

Transcription

1 Conflict of Interest and the Credibility of M&A Advisor Recommendations Kshitij (Ken) Shah Department of Economics and Finance Albers School of Business & Economics Seattle University and Mark Simonson Department of Finance Arizona State University October 2006

2 Conflict of Interest and the Credibility of M&A Advisor Recommendations ABSTRACT We study the advice provided by investment banks to the board of directors and stockholders of large firms considering mergers between 1998 and Advisors working for the acquiring firms provide estimates of target firm value that are significantly higher than those provided by the target advisors by 21% to 28%, depending on the valuation technique used. Regardless of the technique used, the average valuation of acquiring firm advisors is above the merger price, while the average valuation of the target firm by the target firm advisors is below the offer price. Based on these valuations, such expert opinions would be interpreted by both target and acquirer stockholders as a recommendation to vote to approve the merger. The average acquirer s stock falls by 6% at the merger announcement, reducing acquirers stockholders wealth by $600 million and implying that the stock market believes the merger price is 10% too high. Nevertheless, the acquirers advisors estimate the value to be 6% to 22% above the merger price. Our results are consistent with the existence of a conflict of interest between firms seeking fairness opinions and advisors that favor deal completion to generate the majority of their fee as pointed out by McLaughlin (1990). 1

3 1. Introduction The idea that equity research analysts tend to provide biased recommendations in order to increase the fees they generate for their corporate finance unit was first documented in the finance literature by Michaely and Womack (1999). An awareness of potential conflicts of interest among equity analysts has recently led regulators to investigate the objectivity of their research. 1 The investigation eventually led to the Global Research Analyst Settlement in 2003 between ten investment banks and the Securities and Exchange Commission, the NASD, the New York Stock Exchange, New York Attorney General Eliot Spitzer, and other state regulators. Under the settlement, the ten participating Wall Street entities paid a total of $875 million in penalties and were required to spend $432.5 million to fund independent research to their clients and make third-party research available to their clients along with in-house research for all stocks they cover. 2 Similar to equity analysts, M&A dealmakers compensation generally depends, even more directly, on the amount of investment banking fees they generate, and they are subject to potential conflicts of interest similar to the ones faced by equity analysts when issuing investment recommendations and earnings forecasts. The press, citing investment bankers that recommend eye-popping deals that in retrospect look excessive, has begun to take notice of a potential lack of objectivity by M&A advisors. 3 Recently, the NASD, the main self-regulatory authority for brokerage firms, has begun an inquiry into the possible conflicts of interest involving investment banks that provide fairness opinions in merg- 1 Investors Want Cops on the Street, New York Times, May 28, The Global Research Analyst Settlement 3 Firms That Lived by the Deal in the 90s Now Sink by the Dozens, Wall Street Journal, June 6,

4 ers and acquisitions. The NASD has already sent letters to several Wall Street firms earlier this year requesting information about their recently issued fairness opinions in an attempt to examine possible conflicts in past deals. 4 The presence of a potential conflict of interest in client-advisor relationships in M&A transactions is well known. McLaughlin (1990) finds that 94% of tender offer advisory contracts between bidding firms and investment banks have fees that increase if the deal is successful. He shows that, on average, over 80% of the total bidder and target advisory fees are due only after the successful completion of a deal and suggests that such agreements could provide a conflict of interest between the stockholders, who want to accept value-increasing deals, and advisors, who are biased to ensure deal completion to collect their fee. However, he does not provide any evidence of a bias and also suggests that reputation building concerns of the advisor could serve to nullify the potential agency problem. A typical advisory contract, from the sample of mergers used in our study below, is illustrated in the following excerpt from Morgan Stanley s fairness opinion presented to the board of directors of Household International regarding its bid for Beneficial Corporation: Household has agreed to pay Morgan Stanley: (i) an advisory fee estimated to be between $200,000 and $300,000 which is payable if the Merger is not consummated, (ii) an exposure fee of $3.75 million and (iii) a transaction fee equal to $19,720,500, which is payable upon the consummation of the Merger. In this transaction, 80% of the fee is due on the contingency that the deal is completed. Rau (2000) finds that first-tier investment banks charge an average of 73% of their fees on a contingency basis for tender offers and 55% on a contingency basis for mergers and suggests that deal completion incentives are stronger in tender offers than mergers. 4 See NASD Scrutinizes Conflicts in Bankers Fairness Opinions, Wall Street Journal, June 11,

5 The board of directors of a bidding firm presumably hires an advisor on behalf of the firm s stockholders to provide target firm evaluation services and negotiating skill. Rau (2000) finds that acquirer advisors with greater experience and prestige are able to charge higher fees and are more likely to complete the transaction. Hunter and Jagtiani (2000) find that the greater the proportion of the advisory fees that are paid upon deal completion, the faster the deal tends to be completed. Bowers and Miller (1990), Mikhel, Shaked and Lee (1991), Servaes and Zenner (1996), and Rau (2000) find that neither the acquirer s use of a financial advisor nor the advisor s reputation is related to the abnormal stock return of the acquirer at the announcement of a deal. Rau (2000) also investigates whether the deal completion incentives provided by advisory contracts with high proportional contingent fees tend to lead to valuedecreasing acquisitions. Consistent with this idea, he finds that, after controlling for factors that have been shown to affect post-acquisition abnormal returns, the post-issue performance of acquirers in tender offers is negatively related to the proportion of contingent fee payments in the advisory contracts and concludes that such deal-completion incentives may motivate advisors to recommend bad deals. Further evidence supporting the deal completion hypothesis by top tier investment banks is presented by both McLaughlin (1992) and Rau (2000), who find that bidders using higher quality investment banks, which charge higher proportional contingent fees, offer about 50% higher premiums than the lowest tier of advisors. Servaes and Zenner (1996) examine the role that buy-side M&A advisors serve by comparing deals completed with an advisor to deals completed with only in-house personnel. They report that firms tend to use advisors for large, complex, hostile acquisitions by firms with less acquisition experience and less insider ownership that use at 4

6 least some securities in their payment. They also find that firms using an advisor are more likely to experience a negative stock price reaction at announcement of a deal and suggest that if managers want to make value-destroying investments to increase their private benefit, they may seek a fairness opinion as protection against shareholder lawsuits. A target firm s board should select an advisor to help negotiate a higher price for target shareholders, but the advisor may also suffer from misaligned incentives if the total investment-banking fee is tied to deal completion. Kale, Kini, and Ryan (1999) find some evidence that target firms employing reputable advisors are able to earn higher shareholder premiums at the expense of the acquirer s shareholders suggesting that M&A advisors can improve target bargaining. Firms may seek the validation of an investment bank advisor under well-founded legal concerns as well. The 1985 decision in Smith v. Van Gorkom by the Delaware Supreme Court created the obligation that, when evaluating a takeover proposal, the corporate boards of target firms must inform themselves of all reasonably available and relevant information to the decision. 5 In this case, the court ruled that the board of directors of Trans Union Corporation had been negligent in recommending a merger even though the offer provided shareholders with a 45% premium because the board had not made an informed decision. The court s ruling states that the board was not capable of evaluating the deal alone because: None of the directors, management or outside, were investment bankers or financial analysts. Bowers (2002) suggests that getting a fairness opinion from an experienced, independent advisor could demonstrate the board s attempt to become informed. 5 From the 1985 case 488 A.2d 858 before the Delaware Supreme Court. 5

7 In this study, we analyze the advice provided by M&A advisors. While a substantial literature exists that studies the advice investment bank-affiliated analysts provide about issues such as earnings forecasts and stock recommendations, our study evaluates the nature of the advice they provide in their capacity as M&A advisors. Previous M&A advisory research has studied indirect indicators of the advice provided by investment banks by focusing on variables such as advisory contract fee structures [McLaughlin (1990), (1992) and Rau (2000)], the relation between announcement and post-acquisition stock performance and the prestige of the investment bank advisor [Bowers and Miller (1990), Michel, Shaked, and Lee (1991), Servaes and Zenner (1996), and Rau (2000)], and deal completion rates [Rau (2000)]. Our study extends this literature by examining the specific advice provided by the advisors. As part of their advisory role, the investment banks in our sample are hired by the target or acquiring firm to determine a fair price for the target. This valuation information is presented to the board of directors and then distributed to the stockholders in the merger proxy statement that solicits their vote for the merger. We study a sample of 138 M&A advisors opinions, composed of 69 paired bidder and target valuations. If advisor opinions are unbiased, and both the acquirers and target advisors have the same information, we expect that acquirer and target advisor valuations would show no relation to each other: sometimes the acquirer advisor s would be higher, and sometimes the target advisor s would be higher. If advisor opinions are biased to improve negotiating positions, we expect that target firm advisors should generally have higher valuations relative to those of acquirer advisors. If advisor opinions are biased due to a conflict of interest that favors deal completion, we expect that acquirer firm advisors should generally have higher valuations than target advisors. We find that discounted cash flow valuations of the bidding firm s advisors are significantly higher than those of target advisors by an average of 21%. While target advi- 6

8 sors valuations are an average of 2% below the implied target offer price, acquirer advisors valuations are an average of 19% above the implied deal s value to the target shareholders. Based on these averages, such expert opinions would be interpreted by both target and acquirer stockholders as a recommendation to vote to approve the merger. While discounted cash flow valuations perhaps leave the greatest room for any bias to be manifested, since more variables are at the discretion of the advisor, it also may be the case that the acquirer advisor has different information. For example, different information may allow the acquirer s advisor to rationally price in synergies only recognized by the acquirer s management. Thus, we also study two common relative valuation methods often included in the advisors opinions: (1) comparable firm trading multiples, and (2) comparable industry transaction multiples. These techniques are described in Kaplan and Ruback (1995). Techniques relying on multiples should not suffer from the problem of the two advisors having different sets of information. Both advisors should observe a similar set of comparable firms or market transactions, calculate the multiple of interest (price-toearnings per share, for example), and multiply the multiple times the appropriate variable for the target (earnings per share, for example) to arrive at a share value. When we study the valuations provided by these alternate techniques, we find qualitatively similar results. Based on the comparable transactions approach, we find that target advisors valuations are an average of 6% below the implied target offer price, while acquirer advisors valuations are an average of 22% above offer price. The difference between the two averages is a statistically significant 28%. When the comparable firm method is used, we find that that target advisors valuations are an average of 20% below the implied target offer price, while acquirer advisors valuations are an average of 6% above offer price. The difference between the two averages, 26%, is also statistically significant. 7

9 We also examine the wealth effects of the merger announcements. We find that the stock market is less optimistic about the acquisitions than the acquirers advisors, as the average acquirer s stock falls by 6% at the merger announcement, reducing stockholders wealth by $600 million and implying that the stock market believes the average acquirer is overpaying by approximately 10%. However, the acquirers advisors estimate the value to be 6% to 22% above the merger price. The target s average stock price increases by 14%, or $495 million, suggesting a $100 million average total value decrease for both groups of stockholders. Overall, our results are consistent with what Rau calls the deal completion hypothesis in which the incentives in the contingent fee payments to the investment bank cause it to focus only on completing acquisitions as opposed to providing unbiased advice. However, we could also be documenting a manifestation of Roll s (1986) hubris hypothesis in which offers are only made when the valuation is too high; outcomes in the left tail of the distribution are never observed (page 199). However, we are unable to examine such unreported events. 2. Sample The sample is drawn from deals announced between 1998 and 2001 that had a total value of at least $500 million for which the Securities Data Corporation Domestic Mergers and Acquisitions database reports that the acquirer and target had at least one advisor. We focus on large deals to increase the probability that data will be available after locating and reading lengthy merger agreements. We also require that the proposed transaction includes the acquirer s stock as currency. This requirement was included because it is likely that such deals were required to file an SEC Form S-4, which often includes a merger agreement document with opinions from target and acquirer advisors. The form 8

10 S-4 is essentially a joint merger proxy statement and security registration statement. There were 509 firms meeting these criteria. From these 509 transactions, we proceed to locate and read the S-4s that were filed with the SEC, and we retain the 69 deals for which a target and acquirer advisor both provided at least a discounted cash flow valuation analysis reporting an estimate of the value of the target firm. There are several cases in which the opinion of only one firm s advisor is included or in which the investment banker s opinion discusses some results of a discounted cash flow valuation analysis but did not provide an estimate of the target stock price, so we did not include them in the sample. Table 1 shows that the transactions in our sample are large. The median transaction value is about $2.7 billion, and the distribution is highly skewed, ranging from $502 million to $63 billion. The transaction value is taken from SDC and includes the amount paid for all common stock, common stock equivalents, preferred stock, debt, options, assets, warrants, and stake purchases made within six months of the announcement date of the transaction. The median target value enterprise value, calculated by multiplying the number of target shares outstanding by the offer price and adding the cost to acquire convertible securities, plus short-term debt, straight debt, and preferred equity minus cash and marketable securities, is $3.7 billion. Target advisor fees are generally higher than acquirer advisor fees. Target advisory fees range from $150,000 to $47.3 million and have a median value of $12.5 million. Acquirer advisory fees have a median of $8.5 million and a range of $1 million to $33 million. 3. The content of fairness opinions There are several services that an investment bank can provide to firms considering a merger, from identifying candidates to performing due diligence and negotiating the final 9

11 terms. We study the fairness opinions that often are delivered in this process because they generally provide the only direct, quantitative evidence we are able to observe regarding the opinion offered by the advisor. A merger fairness opinion is a written and signed third-party assertion that certifies the price of the proposed deal. The opinions can be written by a number of parties including investment banks, certified public accounting firms, and consulting firms. Some authors, such as Bebchuk and Kahan (1989) and Sweeney (1999), hypothesize that it is virtually impossible to obtain an unbiased fairness opinion because the issuer faces a conflict of interest arising from existing and potential business relationships with the client, and there is substantial discretion as to how the advisors arrive at their valuation. Nevertheless, the advisors presumably realize that their reputation is on the line and claim to make unbiased fairness opinions. The following excerpt from an investment bank s marketing material asserts such a belief: Our reputation for independence and technical expertise in our fairness opinions enables boards of directors to make sound business decisions. Houlihan Lokey s fairness opinions are based upon strong analytical and empirical evidence combined with a fundamental understanding of the transaction process and deal environment. The fairness opinion regarding the deals in our sample are all from firms involved in investment banking. Most of the valuation opinions are provided by what Rau (2000) and Servaes and Zenner (1996) would likely consider first-tier investment banks, although we do not replicate their ranking methodology. Table 2 shows that of the 138 assignments, 116 are given to: Merrill Lynch (26), Goldman Sachs (22), Salomon Smith Barney (15), Credit Suisse First Boston (14), Morgan Stanley (14), Donaldson Lufkin & Jenrette (9), or Lehman Brothers (6). This is consistent with Rau (2000), who finds that the advisory of large mergers is dominated by prestigious investment banks, while smaller deals are dominated by lower-tier banks. When the advisors in our sample are called upon to provide an opinion, they provide an overall recommendation and a summary of a valuation analysis. The relationship be- 10

12 tween the advisor and the firm is exemplified in the following description concerning Salomon Smith Barney s opinion as Promus advisor in its proposed merger with Hilton: OPINION OF SALOMON SMITH BARNEY INC. Salomon Smith Barney was retained by Promus to act as its financial advisor in connection with the acquisition and the merger agreement. In connection with such engagement, Promus requested that Salomon Smith Barney evaluate the fairness, from a financial point of view, to holders of Promus common stock of the consideration to be received by such holders pursuant to the terms of the merger agreement. On September 3, 1999, at a meeting of the Promus board of directors held to approve the merger agreement, Salomon Smith Barney delivered an oral opinion, which was subsequently confirmed in writing, to the Promus board of directors to the effect that, as of the date of such opinion and based upon and subject to certain matters stated therein, the consideration to be received in the acquisition by holders of Promus common stock was fair from a financial point of view to the stockholders. In arriving at its opinion, Salomon Smith Barney reviewed the merger agreement and held discussions with certain officers and employees of Promus and Hilton concerning the business, operations, financial condition and prospects of Promus and Hilton. A substantively similar statement appears in all of the fairness opinions in the sample. Whether or not the advisor is hired by the acquiring or target firm, the advisors in our sample at least provide a valuation of the target. They generally relate the valuation to the implied offer price in the merger, which is often derived from the deal s exchange ratio. The exchange ratio refers to the number of shares of the acquirer s stock that one share of the target s stock can be converted into in stock transactions. Most mergers are structured as stock-for-stock exchanges to qualify as tax-free transactions under section 368 of the Internal Revenue Code [see Gaughan (2002)]. The results of each investment bank s analysis are presented to the board of directors and summarized in writing and included in the materials distributed with the merger proxy statement to stockholders. The three most popular target firm valuation techniques used by the advisors in our sample are: (1) discounted cash flow analysis, in which the advisors generally discount the free cash flows under an all-equity capital structure at the after tax-tax weighted-average cost of capital; (2) comparable firm trading multiples, usually referred to as selected 11

13 company analysis, in which valuation multiples of other publicly-traded firms in the target s industry are applied to the target firm; and (3) comparable transaction multiples, usually called selected transactions analysis, in which valuation multiples of recent industry acquisitions are used. These techniques correspond to those used by Kaplan and Ruback (1995) in their study of the accuracy of discounted cash flow valuation versus methods that use comparable multiples. The major difference from that used by Kaplan and Ruback (1995) is in the calculation of the terminal value in the discounted cash flow technique. The advisors in our sample generally estimate the terminal value as a comparable firm multiple, as opposed to Kaplan and Ruback s (1995) use of a growing perpetuity. We require that the advisor summarize their findings by providing a valuation range for each technique. In the two cases in which a firm hired two advisors, we used the first advisor s valuation analysis. 4. Empirical Results 4.1. Target firm valuations In this section, we present our findings from studying the advisors various valuation analyses. We also present in some detail the analysis of one representative transaction: Hilton s 1999 acquisition of Promus Hotel Corporation. The inclusion of parts of the reports of Morgan Stanley and Salomon Smith Barney provide a good indication of the language and procedure used in the investment bankers analyses for the sample as a whole Discounted cash flow In every transaction in the sample, both the acquirer and target advisors provide a range of target stock values based on their version of a discounted cash flow analysis. In a typical example from the sample, Hilton s acquisition of Promus, Hilton hired Morgan 12

14 Stanley to evaluate the fairness of the $38.50 offer price. Morgan Stanley provided the following summary of their analysis: Morgan Stanley performed a discounted cash flow analysis for Promus based upon publicly available information, equity research estimates and financial projections provided by the management of Hilton Morgan Stanley calculated unlevered free cash flows, defined as net income plus the aggregate of depreciation and amortization, other non-cash expenses and after-tax interest expense less the sum of capital expenditures and investment in non-cash working capital. Morgan Stanley calculated terminal values by applying a range of multiples from 8.5x to 9.5x to last 12 months EBITDA. The cash flow streams and terminal values were then discounted to the present using an estimated range of the weighted average cost of capital for Promus of 11.5% to 12.5%. Using this procedure, Morgan Stanley reported a range of stock values of between $45.25 and $ Similarly, Promus hired Salomon Smith Barney that provided their own discounted cash flow analysis, which they summarized as: Salomon Smith Barney performed a discounted cash flow analysis of Promus to estimate a range of values for the Promus common stock. The discounted cash flow analysis for Promus was based upon certain financial forecasts for the years 1999 through 2003 prepared by the management of Promus. Salomon Smith Barney performed a discounted cash flow analysis of Promus in order to determine the aggregate net present value of the unlevered free cash flows of Promus' businesses, net of the outstanding debt balances. A terminal value multiple range of 8.5x to 10.5x was applied to the projected EBITDA in year The unlevered free cash flows were then discounted to present value using discount rates ranging from 9.5% to 11.5%. Salomon s resulting valuation range was $35.03 to $ Morgan Stanley s average value of $48.75 is 27% above the implied offer price of $38.50, while Salomon Smith Barney s average value of $37.67 is 2% below the $38.50 price. Tables 2 and 3 summarize the results for the full sample and show that the Hilton-Promus deal is representative of the transactions studied. Column 1 of table 3 shows that the acquirers advisors tend to have higher estimates of target firm value than the targets advisors. While target advisors average target stock value is an average of 2.2% below the offer price, the acquirers advisors average value is 19.0% above the offer price. The 21.2% difference is statistically different from 0% at the 1% level. 13

15 In the analysis in table 4, we study the fraction of valuation ranges that exceed the implied transaction offer price. Panel A shows that the highest price in the acquirers advisors discounted cash flow range is significantly more likely to exceed the offer price. The acquirers advisors range extends above the offer price in 97% of the observations, versus 79% for the targets advisors. Panel B shows that the lowest price in the acquirers advisors discounted cash flow range is significantly more likely to exceed the offer price. The acquirers advisors range does not extend below the offer price in 39% of the observations, versus 11% for the targets advisors. At this point it could be argued that acquiring firm advisors are more able to price in factors such as synergies that would make the target firm worth more to the bidder. With this in mind, we next turn to an analysis of valuation techniques in which issues such as pricing in synergies seem to be less of a concern. However, it is important to note that it is the fiduciary responsibility of the target s board of directors, and thus their agents, the target firms advisors, to consider factors in a given potential sale, such as synergies, when arriving at their valuation. The Smith v. Van Gorkom ruling found error with the board s comparison of a third-party offer price by Marmon Group to a valuation based on an analysis of a leveraged buyout of the firm that concluded that the firm was worth $55. The ruling states: there was no call by the Board, either on September 20 or thereafter, for any valuation study or documentation of the $55 price per share as a measure of the fair value of the Company in a cash-out context by [Marmon]. 6 As evidence of the thoroughness of their analysis, most of the target and acquirer advisors mention that they met with the both firm s management to get an understanding of factors influencing the value of the firm. For example, in the Amgen-Immunex merger, Merril Lynch, the advisor for the target (Immunex), stated that it conducted discussions 6 From the 1985 case 488 A.2d 858 before the Delaware Supreme Court. 14

16 with members of senior management of Immunex and Amgen concerning their respective businesses and prospects before and after giving effect to the merger and the potential synergies expected. Similarly, Goldman Sachs, the advisor for the acquirer (Amgen), said that it held discussions with senior management of Amgen and with senior management of Immunex regarding the past and current business operations, financial condition and future prospects of Immunex Comparable firm trading multiples When using valuation methods involving comparable firm trading multiples, there is presumably less room for discretion on the part of the analyst performing the valuation relative to discounted cash flow analysis. The areas that require discretion are the selection of the comparable companies and the selection of the type of multiples to use (e.g. price-to-earnings). The remaining procedure is mechanical: multiply the ratio times the particular item being used. The advisors typically state the results of their analysis by providing a valuation range, often without explicitly discussing the precise method or methods they used. As with the discounted cash flow analysis, we study the valuation ranges provided. Once again, it is useful to study the Hilton-Promus transaction from the sample. Here, the advisor for Hilton (the acquirer), Morgan Stanley, provides the following summary: Using publicly available information, Morgan Stanley performed an analysis comparing Promus' current trading value and the implied multiples for a variety of operating statistics to those of selected publicly traded companies that share some of the same characteristics of Promus. In particular, Morgan Stanley focused on the following companies: - Choice Hotels International, Inc. - Hilton - Marriott International, Inc. - Starwood Hotels & Resorts Worldwide, Inc. - Wyndham International Inc. 15

17 Morgan Stanley reviewed financial information including the price to forecasted calendar year 1999 and calendar year 2000 earnings multiples and the aggregate value to forecasted calendar year 1999 and 2000 earnings before interest, taxes, depreciation and amortization ("EBITDA") multiples. The financial information was based on a compilation of earnings projections by securities research analysts. Based on this analysis, Morgan Stanley reported the implied value per share of Promus common stock ranging from $38.08 to $ The advisor for Promus (the target), Salomon Smith Barney, similarly provided a comparable firm multiple analysis, which they summarized as follows: Using publicly available information, Salomon Smith Barney analyzed the market values and trading multiples of the following five publicly traded companies in the lodging industry: - Choice Hotels - Marriott International, Inc. - Hilton - Starwood Hotels & Resorts Worldwide, Inc. - Wyndham International, Inc. Applying a range of multiples derived from the selected lodging companies of estimated calendar 1999 and 2000 EPS and estimated calendar 1999 and 2000 EBITDA to corresponding financial data of Promus resulted in an implied equity reference range for Promus of approximately $33.98 to $38.98 per share. It is noteworthy that even though both advisors used the exact same comparable firms and the same multiple, they arrived at very different values. Morgan Stanley s average value of $46.24 is 20% above the implied offer price of $38.50, while Salomon Smith Barney s average value of $36.48 is 5% below the $38.50 price. Table 3 shows that the results in the Hilton-Promus transaction are quite typical for the sample as a whole. The target advisors find that the target stock value is an average 21% below the offer price, while the acquirers advisors average value is 6% above the offer price. The 26% difference is statistically different from 0% at the 1% level. Panel A in 3 shows that the highest price in the acquirers advisors value range is significantly more likely to exceed the offer price. The acquirers advisors range extends above the offer price 16

18 in 67% of the observations, versus 31% for the targets advisors. Panel B shows that the lowest price in the acquirers advisors range is above the offer price in 7% of the observations, while it is never above the offer price for the target s advisors Comparable transaction multiples The final valuation technique we study is the comparable transaction method. In this technique, the investment bankers typically consult a database of recent transactions, identify the implied values as multiples of various accounting variables, and apply that multiple to the firm being studied. Once again, there is less discretion than the discounted cash flow method, as the only discretionary decisions seem to involve the selection of appropriate transactions and the choice of items to use in the multiples. From the Hilton- Promus transaction, Morgan Stanley, the advisor for the acquirer provides the following analysis: Using publicly available information, Morgan Stanley examined the terms of certain transactions involving acquisitions of companies in businesses that were similar in some characteristics to the business of Promus. These transactions included the following transactions: ACQUIREE Red Roof Inns, Inc Stakis plc Inter-Continental Hotels & Resorts Arcadian International plc La Quinta Inns, Inc Interstate Hotels Company Starwood Hotels and Resorts & Doubletree Corporation Wyndham International, Inc Renaissance Hotel Group N.V Red Lion Hotels, Inc ACQUIRER Accor, S.A. Hilton Group plc Bass plc Patriot American Hospitality, Inc. Meditrust Companies Corporation Patriot American Hospitality, Inc. Worldwide, Inc. Promus Patriot American Hospitality, Inc. Marriott International, Inc. Doubletree Corporation Based on this review, Morgan Stanley selected a range of times next 12 months EBITDA for the purposes of determining a valuation. Applying this range to Promus' next 12 months EBITDA produced an implied value per share of Promus common stock between $45.54 and $

19 Similarly, Promus advisor, Salomon Smith Barney, provided their version of a comparable transaction multiples analysis: Using publicly available information, Salomon Smith Barney reviewed the purchase prices and implied transaction value multiples paid or proposed to be paid in the following 12 selected transactions in the lodging industry: ACQUIRER Westbrook Partners Accor Hilton Plc Felcor CapStar Bass Plc Meditrust Patriot American Starwood Patriot American Marriott Doubletree TARGET Sunstone Red Roof Inns Stakis Bristol American General Inter-Continental La Quinta Inns Interstate ITT Wyndham Hotels Renaissance Red Lion Salomon Smith Barney compared purchase prices in the selected transactions as a multiple of one-year forward EBITDA. All multiples were based on publicly available financial information for the relevant transaction. Applying a range of multiples derived from the selected transactions of one-year forward EBITDA to estimated calendar year 2000 EBITDA of Promus resulted in an implied equity reference range for Promus of approximately $44.67 to $56.16 per share. In this technique, both advisors use somewhat different comparable transactions and arrive at similar valuations, with the average values within $1 of each other. However, the empirical results in tables 2 and 3 show that this is not typical of the sample as a whole. The target advisors find that the average target stock value is an average 6% below the offer price, while the acquirers advisors average value is 22% above the offer price. The 28% difference is statistically different from 0% at the 1% level. Panel A in table 4 shows that both advisors in all transactions present ranges that exceed the offer price. Panel B shows that the lowest price in the acquirers advisors value range is significantly more likely to exceed the offer price. The acquirers advisors range does not extend below the offer price in 21% of the observations, while 6% for the targets advisors ranges begin be- 18

20 low the offer price. The difference between the fractions is statistically different from 0% at the 10% level Value changes around the merger announcement In this section, we examine the wealth impact of the merger announcement by studying the target and acquirer stock price reactions to the news. Figure 1 shows the thirty-day window around the announcement beginning 10 days before the announcement date and ending 20 days after. Since it is apparent from the figure that the wealth impact is concentrated in the three-day period centered on the announcement, we study this window more closely in table 5. For the acquiring firms, the mean cumulative abnormal announcement return from the beginning of day 1 to the end of day 1 is 5.87%. The cumulative abnormal return is calculated by subtracting the return on the Center for Research in Security Prices value-weighted index return from each firm s return. The average wealth change, measured as the change in stock market capitalization, over the window is -$603 million. For the target firms, the mean cumulative abnormal announcement return from the beginning of day 1 to the end of day 1 is 14.4%. The average wealth change over the window is $495 million. Figure 2 shows that the gross value change, the sum of the wealth changes of the target and acquirer, is negative. The acquirer announcement returns are lower than those generally reported in the literature. For example, Bhagat, Hirshleifer, and Noah (2001) find a mean announcement abnormal return of 0.65% for 794 acquisitions between 1962 and The large negative announcement returns in our sample suggest that the acquisitions are considered value decreasing. Next, we investigate the hypothesis that firms hire investment banks with implicit assumption that the advisor will do what it takes to ensure deal completion regardless of shareholder wealth effects. 19

21 4.3. Acquirer advisor valuations and deal characteristics [This section is particularly preliminary and incomplete] Rau (2000) uses a ranking system based on market share league tables to classify investment banks as first-, second-, or third-tier. He finds that higher-tier banks advising in mergers and tender offers are more likely to charge a higher portion of their fees on the contingency that the deal is completed and are more likely to complete deals. Rau and Rodgers (2002) attempt to distinguish between the hypotheses that firms hire top-tier investment banks to independently certify the deal s value versus the hypothesis that they are hired to complete deals regardless of shareholder wealth effects. They study the relation between stock returns and variables proxying for the degree of potential agency problems and conclude that first-tier investment banks are hired by firms with a greater potential for significant agency problems and the firms experience worse post-acquiaition performance. In this section, we set out to investigate similar hypotheses with our data set. Our overriding hypothesis is that bidders with the potential for higher agency problems hire investment banks with the implicit assumption that the advisor will do what it takes to ensure deal completion regardless of shareholder wealth effects. Bowers (2000) claims that prior to rendering a fairness opinion, advisors are informed as to the terms of the agreement and the attitudes of the boards of the firms involved. She also finds that advisors are often retained subsequent to the acquisition to facilitate the merger. This suggests that a board with a desire to complete a transaction and an advisor with a desire to satisfy the board may provide biased analyses. Evidence consistent with the conflict of interest hypothesis would be that advisors working for firms with low officer and director ownership provide valuations that would maximize the likelihood of deal completion. In addition, 20

22 we include total advisory fees and total advisory fees as a fraction of deal size as proxies for the degree of the potential conflict of interest between the investment bank and the client. Table 6 shows the results of a regression analysis in which the dependent variable is the difference between the acquirer advisor s discounted cashflow average valuation and the implied offer price. If there is a conflict of interest in which advisors want to maximize the likelihood of acceptance, then this difference would be expected to tend to be large and positive. If there is tendency for firms with high degree of agency problems to hire investment banks with the duty of ensuring deal completion, then firms with a smaller fraction of inside ownership should be observed to have larger valuation differences. If investment banks are motivated to ensure deal completion to collect their fee, then the acquirer fee should be positively related to the valuation difference. We include both total advisor fee and total adviser fee as a fraction of transaction size to pick up this effect. Unfortunately, we do not have information regarding the amount of fees due on a contingency basis. Rau s (2000) sample was limited to deals completed before 1991, a period for which the Securities Data Corporation Mergers and Acquisitions Database reported the detailed fee breakdown. We have total fee data as reported by Securities Data Corporation for 55 of the transactions. Detailed descriptions of the fee breakdowns similar to the Morgan Stanley s fee structure charged to Household International in its merger with the Beneficial Corporation cited in the introduction of this paper are rare. Regressions 1, 5, and 6 in table 6 show that there is no relation between inside ownership and the valuation difference with and without controlling for transaction size, target profitability, the acquisition premium, acquirer fees, and the identity of the investment bank. Regression 4 provides some evidence that the adviser fee as a fraction of transaction size is positively related to the valuation difference, however the variable is insignifi- 21

23 cantrelated in regressions 5 and 6 which control for inside ownership, transaction size, target profitability, the acquisition premium, acquirer fees, and the identity of the investment bank. Overall, our cross-sectional anlaysis provides virtually no evidence of a conflict of interest on the part of the investment bank advisers or agency problem on the part of the managers. However, our test at this point is not very powerful, and the topic deserves further investigation. 5. Conclusions In this paper, we document the techniques used by investment banks in their capacity as advisors in mergers. While a substantial literature exists that studies the advice investment banks provide about issues such as earnings forecasts and stock recommendations, our study evaluates the nature of the advice provided in their capacity as M&A advisors. The investment banks in our sample are hired by the target or acquiring firm to determine a fair price for the target. This information is presented to the board of directors and then distributed to the stockholders in the merger proxy statement that solicits their vote for the merger. We find that the advisors use similar valuation techniques but tend to arrive at predictably different opinions. Advisors working for the acquiring firms provide estimates of target firm value that are significantly higher than those provided by the target advisors by 21% to 28%, depending on the valuation technique used. Regardless of the technique used, the average valuation of acquiring firm advisors is above the merger price, while the average valuation of the target firm is below the offer price. Based on these averages, such expert opinions would be interpreted by both target and acquirer stockholders as a recommendation to vote to approve the merger. Our results are consistent with a potential 22

24 conflict of interest between firms seeking fairness opinions about mergers and the investment banks providing the advice as noted by McLaughlin (1990). While these preliminary results can be interpreted as consistent with the existence of a conflict of interest by the investment bankers who favor deal completion, we can not rule out the possibility that we are observing a manifestation of Roll s (1986) well-known hubris hypothesis in which offers are only made when the valuation is too high; outcomes in the left tail of the distribution are never observed. 23

25 References Allen, A., and A. Saunders, 2001, The Role of Financial Advisors in Mergers and Acquisitions, NYU working paper. Bhagat, S., D. Hirshleifer, and R. Noah, 2001, The effect of takeovers on shareholder value, University of Colorado working paper. Bebchuk, L. and M. Kahan, 1989, Fairness Opinions: How Fair Are They And What Can Be Done About It?, Duke Law Journal 27, Bowers, H., 2002, Fairness Opinions and the Business Judgment Rule an Empirical Investigation of Target Firms' Use of Fairness Opinions, Northwestern University Law Review, Vol. 96, No. 2, Winter. Bowers, H., Miller, R., 1990, Choice of investment banker and shareholders' wealth of firms involved in acquisitions, Financial Management 19, Gaughan, Patrick, 2002, Mergers, Acquisitions, and Corporate Restructurings, Wiley. Kale, J.R., O. Kini, and E. E. Ryan, 1999, On the Participation and Reputation of Financial Advisors in Corporate Acquisitions, Emory University working paper. Kaplan, S. and R. Ruback, 1995, The Valuation of Cash Flow Forecasts, Journal of Finance 50, No McLaughlin, R., 1990, Investment-banking contracts in tender offers: an empirical analysis, Journal of Financial Economics 28, 209}232. McLaughlin, R., 1992, Does the form of compensation matter? Investment banker fee contracts in tender offers, Journal of Financial Economics 32, Michaely, R. and K.L. Womack, 1999, Conflict of Interest and the Credibility of Underwriter Analyst Recommendations, The Review of Financial Studies, 12 (4), Michel, A., Shaked, I., Lee, Y.-T., An evaluation of investment banker acquisition advice: the shareholders' perspective, Financial Management 20, 40}49. Rau, R., 2000, Investment bank market share, contingent fee payments, and the performance of acquiring firms, Journal Of Financial Economics 56, 2 pp Rau, R. and K. Rodgers, 2002, Do bidders hire top-tier investment banks to certify value?, Purdue working paper. Roll, R., 1986, The Hubris Hypothesis of Corporate Takeovers, Journal of Business 59 no. 2,

26 Servaes, H. and M. Zenner, 1996, The Role of Investment Banks in Acquisitions, Review of Financial Studies 9, Sweeney, P., 1999, Who says it's a fair deal?(fairness opinions in securities disclosures), Journal of Accountancy. 25

27 Figure 1 Cumulative abnormal returns around announcement of merger Bidder Target 20% 15% 10% 5% 0% -5% -10% Days relative to merger announcement 26

28 Figure 2 Cumulative average changes in market capitalization around the merger announcement 600 Target Bidder Total Millions of dollars Days relative to merger announcement 27

29 Table 1 Sample descriptive statistics Target firm characteristics and advisory fees in millions of dollars. Minimum First Quartile Median Third Quartile Maximum Value of transaction a 502 1,078 2,671 6,330 62,593 Enterprise value b 533 1,185 3,706 7,641 25,5776 Equity value c ,299 5,402 62,554 Net sales ,203 2,990 24,176 Total assets ,201 5, ,323 EBITDA ,365 Target advisor fees Acquirer advisor fees a Total value of consideration paid by the acquirer, excluding fees and expenses. The dollar value includes the amount paid for all common stock, common stock equivalents, preferred stock, debt, options, assets, warrants, and stake purchases made within six months of the announcement date of the transaction. Liabilities assumed are included in the value if they are publicly disclosed. Preferred stock is only included if it is being acquired as part of a 100% acquisition. If a portion of the consideration paid by the acquirer is common stock, the stock is valued using the closing price on the last full trading day prior to the announcement of the terms of the stock swap. If the exchange ratio of shares offered changes, the stock is valued based on its closing price on the last full trading date prior to the date of the exchange ratio change. b The enterprise value of a transaction is calculated by multiplying the number of target actual shares outstanding from the most recent source available by the offer price and then by adding the cost to acquire convertible securities, plus short-term debt, straight debt, and preferred equity minus cash and marketable securities, stated in millions. c Calculated by multiplying the actual number of target shares outstanding from its most recent balance sheet by the offer price per share plus the cost to acquire convertible securities, stated in millions. 28