An evaluation of investment banker acquisition advice: the shareholders' perspective.
* The recent wave of corporate acquisitions has brought renewed focus on the role of investment bankers in the acquisition process. While much has been written in the literature about the characteristics of mergers and acquisitions (see  and ), there is no available literature focusing on the quality of investment banker advice. This paper assesses both premiums paid by acquiring firms and rates of return accruing to acquirers as a function of their investment bankers.
While previous research has considered the effect of the choice of investment bankers on performance, such research has focused exclusively on the pricing of initial public offerings (IPOs). The results have generally supported the hypothesis that prestigious investment bankers underprice new issues less frequently than non-prestigious investment bankers. Since Logue  intially reported the finding, numerous other studies, most using different time frames, have reported similar results. These include Neuberger and Hammond , McDonald and Fisher , Neuberger and La Chapelle , and Block and Stanley . Johnson and Miller , however, provide evidence that the market for initial public offerings is mean-variance efficient. While they find that the level of investment banker prestige varies inversely with the degree of IPO underpricing, they show that this relationship disappears after adjusting for risk. Bowers and Miller  consider aggregate categories of investment bankers. They find that when either the bidding or target firm employs first-tier investment bankers, the total wealth gained from the transaction is greater than when neither firm employs a prestigious banker.
The research described in this paper differs from previous research in that it considers specific investment banker performance in providing merger advice. The advice considered relates to the price paid by an acquirer for an acquisition. Johnson and Miller's  categorization of investment banker prestige forms the basis of the classification used in this study. They refer to their most prestigious category as the "bulge bracket". It includes the First Boston Corporation, Goldman Sachs & Company, Merrill Lynch, Pierce, Fenner & Smith Inc., Morgan Stanley & Company, and Salomon Brothers. Their next most prestigious grouping is the "major bracket" including 15 firms such as Drexel Burnham Lambert Inc., Shearson/American Express Inc., Lehman Brothers, Kuhn Loeb Inc., and Dean Witter Reynolds Inc. Their third most prestigious list of investment bankers includes 15 firms such as Oppenheimer & Company, Alex Brown & Sons Inc., and Advent Inc.
Our study, like previous studies investigating initial public offerings, does not determine how investment banker prestige affects performance. Most studies have been unable to prove whether prestigious bankers specialize in high quality issues or whether they influence the perception of quality through their efforts. This study, however, casts doubt upon whether the prestige variable affects performance in the area of merger advice. It shows that Drexel Burnham Lambert, less prestigious than most of the other investment bankers included in the sample of firms studied, outperformed a more prestigious sample of investment bankers, in terms of advising clients' equity investment.
I. Methodology and Data
The event study methodology provides the frameworks used to compare the performance of investment bankers in acquisition situation. Acquiring firms are aggregated by investment banker. The premium paid by each acquirer to its target firm is measured using the capital asset pricing model (CAPM). Deviations between actual rates of return and expected rates of return are generated. These abnormal returns are cumulated and analyzed separately for each investment banker. Ceteris paribus, the lower the premiums paid to achieve an acquisition, the more satisfactory the performance of the investment banker. These premiums are observed by considering the cumulative abnormal returns (CARs) of the target firms classified by investment banker. If, around the announcement date, the CARs of the targets of firms advised by investment banker A are on average higher than those targets of firms advised by investment banker B, investment banker B is said to be providing comparatively advantageous advice to its clients, enabling them to purchase their targets at lower premiums.
Furthermore, the CARs of the acquiring firms are also investigated. The higher the CAR of an acquiring firm, the more beneficial the acquisition to the acquirer. Since the payment of high premiums to targets may be associated with the acquisition of more desirable firms, it is necessary to consider the resulting returns to the acquiring firms. Investment bankers whose clients generate higher cumulative abnormal returns on average are said to be providing
superior advice to their clients.
As a result, comparatively advantageous advice is said to be provided by the acquirer's investment banker which offers both the lowest premiums to its client's target firm and generates the highest abnormal returns to the acquiring firm being advised. Thus, performance is measured by considering abnormal returns to both acquiring firms and their targets around the announcement date.
The sample was drawn from the "100 Largest Transactions" listed each year in Mergers and Acquisitions, from 1981 through 1988.(1) It included mergers completed between 1981 and 1987. To obtain the final sample of companies, firms were eliminated for a variety of reasons. They included the following:
* In-House: Mergers and Acquisitions indicated that a number of transactions took place without an external advisor for the acquirer.
* Multiple Investment Bankers: Several transactions were listed as having multiple investment bankers for the acquirer.
* Undisclosed: In a number of the transactions, the investment banker was undisclosed by Mergers and Acquisitions.
* Partial Sale: If less than 50% of a firm was sold, it is not included in the database.
* LBOs: Post 1983, Mergers and Acquisitions listed separately those transactions which could be classified as LBOs. From 1981-1983, judgment was used to determine which transactions should be excluded as being LBOs.
* Divestitures: Similar to the methodology used for LBOs, Mergers and Acquisitions listed divestitures separately post 1983. Judgment was used to exclude such transactions between 1981 and 1983.
* International: Acquisitions in which the acquirer was a non-U.S.-based firm were not included in the sample. Using the aforementioned criteria, Exhibit 1 details the screening process used to obtain the final sample. [TABULAR DATA OMITTED]
Of the 203 firms used in the analysis, 122 are acquired firms and 81 are acquirers. Six major investment bankers and a category of "other" investment banks are considered in this study. The major investment banks were required to have advised in at least 10 situations where data was available for the acquired firm. They include Goldman Sachs & Co., Morgan Stanley & Co., First Boston Corp., Shearson Lehman Brothers, Salomon Brothers and Drexel Burnham Lambert. Initially, Lazard Freres, Dillon, Read, and Merrill Lynch were included, but were dropped out of the final list of investment bankers because although they each advised 10 or more firms, a lack of data availability caused fewer than 10 firms to be available in the final sample. As a result, the final sample included four firms from Johnson and Miller's  most prestigious "bulge" categorization. These included First Boston, Goldman Sachs, Morgan Stanley and Salomon Brothers. The final sample also included Drexel Burnham Lambert and Shearson Lehman Brothers (including its predecessor firms Lehman Brothers, Kuhn Loeb and Shearson/American Express). Drexel Burnham Lambert and both of Shearson's predecessor firms are included in the second of the three categories listed by Johnson and Miller. Because of Drexel's representation of small and midsized firms, its association with raiders and its prevalent use of junk bonds, Drexel's prestige ranking was considered markedly below that of most other firms in category 2 by numerous participants in the industry interviewed by the investigators.(2)
"Other" investment bankers are also included in this study. They are defined as those firms advising on fewer than 10 deals during the sampling period. Because of the limited number of deals per firm, the data for these investment banks are aggregated and included in one category which includes 97 deals advised by 30 different "other" investment bankers. After eliminating LBOs, divestitures, partial sales and those deals for which data was unavailable, 19 targets and 10 acquirers are included in the "other" investment bankers category.
Exhibit 2 indicates the number of acquiring companies and acquired companies used in the sample for each investment banker. [TABULAR DATA OMITTED]
This analysis considers abnormal returns around the announcement of the planned takeovers. The day prior to the formal announcement of the merger in the Wall Street Journal (WSJ) is used as the announcement date. In some cases, however, there was speculation of the pending announcement just prior to its publication. If more than a five-day gap exists between reported speculation of a pending announcement and the formal announcement, the formal announcement date is used. On the other hand, if five or fewer days exist between reported speculation of a pending merger and the formal announcement, the speculation date is used. Indeed, this is the date that one typically sees the jump in CARs.
In the 80-trading-day interval surrounding the announcement date (announcement date [+ or -] 40 days), the abnormal returns for portfolios of acquiring and acquired firms are attributed to announcement of proposed acquisitions. The average portfolio abnormal return for a specific investment banker on a given day is then analyzed to determine whether the excess return is statistically significant.
The abnormal returns (ARs) for each company were calculated by subtracting the expected returns from the actual returns for each day under investigation. Then, daily abnormal returns for each company were summed over the period of investigation to obtain cumulative abnormal returns (CARs).
Average abnormal returns (ARs) for each day were calculated across all acquiring firms being advised by a particular investment banker. This calculation is performed for each of the six investment bankers and the "other" investment bankers category in the study. Daily abnormal returns for acquired firms were also averaged over each of the investment bankers advising acquiring firms.
Finally, the cumulative abnormal returns (CARs) for each investment banker were obtained by summing the average abnormal returns over the period of investigation.
An analysis is thus performed for both acquired firms and acquiring firms. The Kruskal-Wallis test is used to determine whether the variations among investment bankers are chance variations or whether they represent significant population differences. The same test was used by Scott and Martin  in an industry study of mean common equity ratios and by Michel and Shaled  in a study of dividend policy across industries.
In addition to the Kruskal-Wallis test, the normality of the underlying CAR distributions is then tested to determine if a t-test is appropriate. In those cases where it is appropriate, a t-test is also used to assess differences between investment bankers.
The results for the sample of six investment banks and the category of "other" investment banks are based on a comparison of the CARs of the sample of investment bankers. Exhibits 3 and 4, respectively, illustrate the abnormal returns and cumulative abnormal; returns [+ or -] 3 days surrounding the announcement date. The ARs and CARs are presented for both acquired and acquiring firms and classified by investment banker. [TABULAR DATA OMITTED] [TABULAR DATA OMITTED]
As will be described, the results show statistically significant differences between some of the investment banks and nonsignificant differences between others. These results are evident for both the acquired firm and acquiring firm samples. The plot of acquired firm CARs for the 80-day period around the announcement date is shown in Exhibit 5.
The results indicate that the performance of Drexel dominates that of the other investment banks in minimizing premiums paid in an acquisition. The CARs generated by Drexel targets are less than those generated by targets of the other investment banks in the sample. This result holds during virtually the entire period tested. In particular, in the days immediately prior to the announcement date and in the entire post-announcement period large and significant differences between Drexel and other investment bankers were observed. For example, while the CARs of Drexel targets range from 0.025 on day -3 to 0.21 on day 40, First Boston targets range from 0.18 on day -3 to 0.41 on day 40. Using the Kruskal-Wallis test to compare pairs of investment bankers, Drexel significantly outperformed the poorest performer, First Boston, on each of days -3 through +2. Drexel also significantly outperformed Goldman Sachs on the days immediately prior to the announcement (days-3 through-1) and Morgan Stanley on day-1.
The underlying CAR distributions were tested and determined to be normal. As a result, results of the significance tests for the period surrounding the announcement date ([+ or -] 3 days) using both the Kruskal-Wallis and t-tests are presented in Exhibit 6.(3) [TABULAR DATA OMITTED]
The t-test confirms the results of the Kruskal-Wallis test and also indicates several additional dates on which differences between Drexel and other investment banks are significant. Comparisons between other pairs of investment banks not including Drexel were not significant.
Interestingly, when observing CARs of acquiring firms, the dominance of Drexel is only observed in the postannouncement period. However, in the preannoucement period, as can be observed in Exhibit 7, Morgan Stanley outperformed all other investment bankers included in the study. In this period Drexel performed the poorest, with First Boston dominated by all other investment bankers with the exception of Drexel.
Statistically significant differences can be observed using both the Kruskal-Wallis and t-tests prior to and following the announcement. In particular, as can be seen in Exhibit 8, Morgan dominates First Boston throughout the period, day -3 to day +3. [TABULAR DATA OMITTED]
Morgan also significantly outperforms Drexel and "other" bankers during the period from day -3 through the announcement date, day 0. Indeed, First Boston is statistically dominated by most of the investment banks in the sample of acquiring firms during both the three days prior to and post the announcement date. It is interesting to observe that in the three days prior to the announcement date Drexel is also dominated by all other investment banks in the study with the exception of First Boston. However, as seen in Exhibit 7, following the initial postannouncement period, Drexel dominates all other investment banks. By day 10 following the announcement, average Drexel client CARs have increased above the average CARs of each of the other investment banks in the study. Indeed by day 14, the performance of Drexel clients is statistically superior to the poorest performer, First Boston, at the 0.10 level. This statistical superiority of Drexel continues to increase over time during the investigation period.
Drexel's CAR increases from -0.07 at the announcement date to approximately 0.13 at day 40. This 20% improvement translates to a $100 million abnormal increase in equity value for a company with a stock market value of $500 million. It is interesting to note that while the Cars of Drexel clients are increasing following the announcement date, the CARs of clients of all other investment banks remain approximately level or decrease slightly.
While the results of the acquired firms shown in Exhibit 5 have the traditional S-curve obtained in most event study analyses, the results for the acquiring firms differ from typical acquiring firm data and are particularly interesting. Indeed, much of the change in CARs take place following the announcement date. This is contrary to results typically predicted by the efficient market hypothesis. The hypothesis suggests that all gains (losses) would be incurred prior to or at the time of the announcement, with abnormal returns following the announcement fluctuating randomly around zero. While the abnormal returns for Goldman Sachs, Morgan Stanley, Shearson Lehman and Salomon do fluctuate around zero with no significant trend, the postannouncement abnormal returns trend significantly upward for Drexel and trend slightly, but not significantly, downward for First Boston.
It is impossible to ascertain with certainty the reason for Drexel's superior performance. However, one plausible explanation is that market participants were not generally aware of Drexel's record in assisting and financing acquiring firms. While acquisitions are typically subject to the winner's curse (see ), Drexel clients investing in equity during the investigation period were able to avoid the curse on average.
The lack of recognition attributed to Drexel may have resulted from its position as a relative newcomer among major Wall Street merger advisers. It may also have resulted from the fact that the relatively small size of the monitored as many of the deals advised by the other investment bankers in the sample. The mean deal size included in the sample is presented in Exhibit 9.
Exhibit 9. Average Deal Price by Investment Banker Average Deal Price (in millions) Drexel Burnham Lambert $ 795 First Boston 1,951 Goldman Sachs 1,692 Morgan Stanley 1,354 Salomon Brothers 1,004 Shearson Lehman Brothers 734(*) (*) The average Shearson Lehman deal price was reduced by three deals under $180 million, while the firm was still Lehman Brothers, Kuhn Loeb, Inc.
Another factor which should be included in the evaluation of the results is the multiple bidding phenomenon. As Michel and Shaked  have shown, the CARs at the time of the announcement are not significantly different for multiple bidder and single bidder firms. Within 50 days after the announcement, the gap between single and multiple bidder CARs increases, but remains statistically insignificant. Because of the small size of the sample in the present study, it was impossible to eliminate all firms associated with multiple bidding. However, based on the Michel and Shaked results, it s unlikely that multiple bidding played a major role in the results. The following list indicates the fraction of firms targeted by each investment banker for which multiple bidding exists: Drexel Burnham Lambert (0.17), First Boston (0.37), Goldman Sachs(0.22), Morgan Stanley (0.14), Salomon Brothers (0.0), and Shearson Lehman Brothers (0.27). While the fraction of multiple bidding situations is greatest for First Boston and the acquired firm returns are greatest for First Boston, the converse result for the smallest fraction of multiple bidding situations does not hold. Salomon Brothers, at the end of the 40-day horizon, with zero multiple bidding situation, had the second highest level of CARs. Similarly, no other direct relationship in the ranking of the investment bankers can be associated with the number of multiple bidding situations related to each investment banker.
CAR-based premiums were used as the basis for this study, rather than premiums based on offer price. While the alternative definition is sometimes used in traditional merger studies, it has the disadvantage of not accounting for market movements or dividend payments during the investigation horizon. Furthermore, it implicitly assumes similar systematic risk level for each of the securities in the sample.
To ensure that the ex-post returns are not merely a result of downward biased betas caused by expectations of increased debt levels, a test comparing the pre- and post-betas was run. Daily returns from the CRSP tapes were converted to weekly returns for the year prior to the acquisition, yielding 52 observations for each acquired and acquiring firm for which data were available. The beta for each firm was calculated and weighted by its respective asset value to obtain a weighted average preacquisition beta. Also, for the period following the acquisition, the CRSP daily returns for 52 weeks were converted to weekly figures for the merged company and again a [Beta] was calculated. A t-test was then run comparing the weighted average preacquisition [Beta] with the merged firm's postacquisition [Beta] for each firm in the sample. The results showed no significant postacquisition shifts in [Beta] for each acquisition. The results were also run for each investment banker with similar results. For example, the mean preacquisition [Beta] for First Boston was 1.097, while the mean postacquisition [Beta] was an almost identical 1.112. Similarly, for Drexel the mean preacquisition [Beta] was 1.140 and the mean postacquisition [Beta] was 1.145. As a result of the individual firm results, it is unlikely that the ex-post positive excess returns for Drexel resulted from a downward bias in the posttransaction betas. In addition, results were calculated for the entire sample. The results were similar to those reported for each firm and each investment bank. The mean weighted average [Beta] for the entire sample prior to acquisition was 1.116 and the postacquisition [Beta] was 1.27, resulting in a statistically insignificant mean difference of 0.011 and a standard deviation of 0.074.
III. Summary and Conclusions
These results suggest that Drexel Burnham Lambert outperformed the other investment bankers in the sample in providing acquisition bid advice to client firms. Target firms of Drexel clients investing in equity generated the lowest CARs relative to those of other investment bankers in the sample. In other words, Drexel bid up their deals comparatively less than other investment bankers. The results were consistent over the entire investigation period. Although Drexel's CARs were consistently less than all other firms in the sample, significant results in both the preannouncement and postannouncement periods were obtained only in a pair-wise comparison with the performance of First Boston. To determine whether Drexel's advice regarding its low bid premiums was generated as a result of the purchase of unattractive targets, CARs of acquiring firms were also generated. Morgan statistically outperformed First Boston, Drexel and the "other" bankers throughout the three-day period immediately prior to the announcement. Immediately following the annoucement, only First Boston was statistically dominated. It was outperformed by Goldman, Salomon, Shearson and the "other" bankers. Following day 14, however, Drexel's results statistically dominated First Boston.
The structure of these results differs from that of the Bowers and Miller  study of investment banker choice in acquisitions and those studies investigating the role of investment banker prestige in underwriting initial public offerings. In these studies, the degree of prestige varied directly with performance. In this merger study, Drexel Burnham Lambert, which is in the least prestigious investment banker category according to the Johnson and Miller categorization, outperformed all others in the sample. Morever, First Boston, on of the Johnson and Miller "bulge" category firms, performed least effectively.
It must be noted that because of the limited data, these results must be considered preliminary. There are other possible explanations which, in future research, might be tested. These include a diverse set of reasons: inferior targets, the use of cash tender offers, specialization by Drexel in certain industries, etc. Further research investigating investment bank merger advice would be useful. One of the areas for such research is the observation based on our study that the reported performance of the bulk of the investment banks in the study clustered together, with only Drexel and First Boston being differentiated. Such research might investigate other parameters not included in this study which discriminate among the investment banks based on M&A performance. Research using a larger database of over-the-counter firms would also clarify the role of smaller firms in generating superior performance.
A further intriguing area for future research is the set of postmerge positive CARs of Drexel clients. Since this outcome is unusual, further investigation of whether this is a Drexel phenomenon or a representation of one or more other factors is warranted.
(1) For 1981, the "Top 25 Transactions" per quarter were used. The "100 Largest Transactions" were initiated in 1982.
(2) The initial version of this paper, including discussion with industry participants, was completed prior to Drexel's filing for protection under Chapter 11.
(3) The authors appreciate the suggestion by one of the reviewers to also run the t-test.
[1.] P. Asquith, R. Bruner and D. Mullins,Jr., "The Gains to Bidding Firms from Mergers," Journal of Financial Economics (March 1983), pp. 121-140. [2.] S. Block and M. Stanley, "The Financial Characterisitics and Price Movement Patterns of Companies Approaching the Unseasoned Securities Market in the Late 1970's," Financial Management (Winter 1980), pp. 30-36. [3.] H. Bowers and R. Miller, "Choice of Investment Banker and Shareholders' Wealth of Firms Involved in Acquisitions," Financial Management (Winter 1990), pp. 34-44. [4.] J.M. Johnson and R.E. Miller, "Investment Banker Prestige and the Underpricing of Initial Public Offerings," Financial Management (Summer 1988), pp. 19-29. [5.] D.E. Logue, "On the Pricing of Unseasoned Equity Issues; 1965-1969," Journal of Financial and Quantitative Analysis (January 1973), pp. 91-103. [6.] J.G. McDonald and A.K. Fisher, "New-Issue Stock Price Behavior," Journal of Finance (March 1972), pp. 97-102. [7.] A.J. Michel and I. Shaked, "Evaluating Merger Performance," California Management Review (Spring 1985), pp. 109-118. [8.] A.J. Michel and I.Shaked, "Country and Industry Influence on Dividend Policy: Evidence from Japan and the U.S.A.," Journal of Business Finance and Accounting (Autumn 1986), pp. 365-381. [9.] A.J. Michel, I.Shaked and R.Dickie, "The Winner's Curse in the Merger Game," Journal of General Management (Spring 1987), pp. 32-51. [10.] A.J. Michel and I. Shaked, "Corporate Takeovers: Excess Returns and the Multiple Bidding Phenomena," Journal of Business Finance and Accounting (Summer 1988), pp. 263-273. [11.] B.M. Neuberger and C.T. Hammond, "A Study of Underwriters' Experience with Unseasoned New Issues," Journal of Financial and Quantitative Analysis (March 1974), pp. 165-177. [12.] B.M. Neuberger and C.A. La Chapelle, "Unseasoned New Issue Price Performance on Three Tiers: 1975-1980," Financial Management (Autumn 1983), pp. 23-28. [13.] D.F. Scott, Jr., and J.D. Martin, "Industry Influence on Financial Structure," Financial Management (Spring 1975), pp. 67-73.
|Printer friendly Cite/link Email Feedback|
|Author:||Michel, Allen; Shaked, Israel; Lee, You-Tay|
|Date:||Jun 22, 1991|
|Previous Article:||Regulatory deterrence and registered insider trading: the case of tender offers.|
|Next Article:||Valuation of callable bonds under progressive personal taxes and interest rate uncertainty.|