When do banks listen to their analysts? Evidence from mergers ...

When do banks listen to their analysts? Evidence from mergers and acquisitions

David Haushalter

Penn State University E-mail: [email protected] Phone: (814) 865-7969

Michelle Lowry•

Penn State University E-mail: [email protected]

Phone: (814) 865-1483

July 13, 2009

• We thank Lubomir Petrasek for excellent research assistance. We thank Richard Bundro, Laura Field, Urs Peyer and seminar participants at Case Western Reserve University, Hong Kong University of Science and Technology, INSEAD, National University of Singapore, Singapore Management University, the University of Colorado, and the University of Lausanne for helpful comments and suggestions.

When do banks listen to their analysts?

Evidence from mergers and acquisitions

Abstract: We find that the asset management division of a bank ‘listens’ to its own analysts’ recommendations of client firms that have recently announced a merger. In contrast, we observe no evidence of such a relation in the pre-merger announcement period or among non-advisor banks. Empirical results suggest that increased information sharing across divisions regarding the value of the acquirer, rather than pressure on both analysts and asset management divisions to support the acquirer, explain this result. Among banks most sensitive to conflicts of interest from the investment banking division, consistency across divisions is weakest. Finally, banks’ strategy of only listening to their analyst recommendations in times when the analyst is most likely to have value-relevant information and in cases where analysts’ likely conflicts of interest are lowest appears to be a profitable one. In sum, the value of interactions between divisions within a financial institution varies with both conflicts of interest and the information environment.

1. Introduction

This study examines the interaction of divisions within financial conglomerates.

Although the divisions of a financial institution generally fall narrowly within the financial

industry, the fiduciary duties of these divisions vary substantially. An institution’s investment

banking division is often advising corporations that its analysts are recommending for outside

investors, and its asset managers are trading these same stocks for the bank’s own account or for

other outside investors. This setting has the potential to lead to conflicts of interest in which the

activities of one division are favored at a cost to another. It also creates potential for information

sharing across divisions in ways that would not be possible among stand alone entities.

Although these interactions may be optimal for the financial conglomerate, they are not

necessarily in the best interests of the bank’s clients. These concerns are heightened by the

importance of information in the transactions that financial institutions specialize, as noted by

Mehran and Stulz (2007).

Prior literature provides evidence of information transfers within investment banks, from

the investment banking division to analysts and from the investment banking division to asset

management. Michaely and Womack (1999) find that analysts at affiliated banks are pressured

into making upwardly biased recommendations on client firms, while Massa and Rehman (2007)

find that mutual funds of affiliated banks obtain inside information that informs their investments

in client firms, enabling them to earn higher returns. The combined findings of these papers

highlight both the conflicts of interest and information advantages that can arise within a

financial conglomerate. Although conflicts of interests can diminish the performance of non-

investment banking divisions, informational advantages gained from investment banking can

1

also benefit other divisions. The focus of this paper is to examine the relative importance of

these factors.

Our analysis on the association between the analyst and asset management divisions of

financial institutions focuses around two hypotheses. The ‘Division Consistency’ hypothesis

posits that investment banks invest in a manner consistent with the advice they are providing to

clients: they change their holdings of stocks in response to their own analysts’ upgrades and

downgrades.1 From the perspective of investors relying on such recommendations, this is

obviously what is expected. However, from the perspective of the agents within the bank, we

expect to only observe divisional consistency if analysts and asset managers both face similar

incentives and also act on similar information sets.

Alternatively, the ‘Division Inconsistency’ hypothesis proposes that divisions act

divergently: firms that are upgraded by an institution’s analysts are as likely to be bought as to

be sold by its asset managers. A finding of division inconsistency indicates that the analysts and

assets managers face different conflicts of interests and/or that the divisions rely on different

information sets.

We study these issues by following a financial institution’s analysts and asset managers

at the time that it advises an acquirer in a merger. Although potential conflicts of interest and

information sharing can be ongoing, they are arguably particularly large around mergers.

Mergers are a large source of revenues for investment banks, and the magnitude of these fees can

increase conflicts of interests within other divisions that could potentially support such deals.

1 While investment banks have buy-side analysts that directly report to asset management within the bank, Groysberg, Healy, Chapman, Shanthikumar, and Gui (2007) suggest that such analysts provide less informative recommendations, suggesting at a minimum that traders probably also pay attention to the recommendations of the sell-side analysts.

2

(see, e.g., Becher and Juergens, 2005, Kolanski and Kothari, 2009).2 In addition, mergers are

important information events. As highlighted by Moeller, Schlingemann, and Stulz (2005), the

value of companies can change dramatically around mergers. Therefore, the potential benefits of

information flow within the financial institution are particularly large.3

Our analysis of affiliated investment banks, i.e., banks advising the acquirer around a

merger event, provide support for the Division Inconsistency hypothesis in the period prior to the

merger announcement and the Division Consistency hypothesis in the period following the

merger announcement. More specifically, there is a strong positive relation between changes in

analyst recommendations and changes in bank holdings of the acquirer stock in the period

following the merger announcement, indicating that the banks, on average, are investing in a

manner consistent with recommendations made to clients. However, there is no significant

relation in the year leading up to this announcement, suggesting that the bank places a lower

value on this subset of recommendations. The increase in division consistency following the

merger announcement is not consistent with an increased bias in analyst recommendations

combined with an increased precision of asset management investments, as suggested by prior

literature.

The increase in division consistency indicates that there was a change in information

sharing across divisions at this time. We note that information sharing across divisions can take

different forms. Under one scenario, if the investment banking division shares information on

the quality of the merger with analysts, this should enable the analyst to make a better

2 In 2007 alone, the top 20 investment banks earned more than $42 billion in fees from underwriting mergers and acquisitions, about half of the total fees that they earned from all investment banking activities. In addition to these direct fees, mergers can also lead to revenues from follow-on business, including financing and underwriting. See http://www.bloomberg.com/news/marketsmag/mm_0408_story1.html. 3 As shown by Moeller, Schlingemann, and Stulz (2005), returns to acquirers around the announcement of mergers at the 5% and 95% level range from -6% to 7% between 1980 to 1997 and -19% to 13% between 1998 and 2001.

3

recommendation and one would expect the asset management divisions of the bank to be more

likely to listen to these better recommendations.4 However, under an alternative scenario, the

investment banking division may use their information on the deal to pressure both analysts and

asset management divisions to support the deal: analysts to make positive recommendations on

the acquirers that it has recently advised, and asset management divisions to increase

shareholdings of these acquirers (either on the bank’s own account or through mutual funds).

Although both are technically forms of information sharing, the first scenario results in better

quality advice to clients and better trading decisions. The second scenario, the result of conflicts

of interest, yields less useful advice and lower returns from asset management.

We conduct four tests to examine the ways in which different types of information

sharing across divisions may explain the increased consistency between analyst

recommendations and bank investment decisions following the merger announcement. Each test

sheds light on whether the increased division consistency reflects pressure on both analysts and

asset management to support the investment banking division, or whether it is consistent with

analysts having more information to make better recommendations and asset management

following these better recommendations. First, we compare the abnormal returns around

recommendation changes in the pre-merger announcement versus post-merger announcement

period. Second, we examine the patterns in bank investment decisions around upgrades versus

downgrades. Third, we develop proxies for the quality of the recommendation and the extent of

conflicts of interest, and examine whether banks are more or less likely to follow different

subsets of recommendations. Finally, we examine whether banks’ strategy of following their

analysts’ recommendations of client firms is profitable.

4 It is also possible that the asset management divisions received similar information themselves and therefore have a similar view as the analysts.

4

Empirical results across these four tests suggest that the increased division consistency

following the merger announcement primarily reflects increased information sharing from the

investment banking division, resulting in higher quality analyst forecasts that asset management

is more likely to follow. We find that the market reaction to analyst recommendation changes is

significantly greater in the post-announcement period, suggesting that these recommendations

contain more information. Moreover, the market reaction recommendation changes by affiliated

analysts is significantly greater than the reaction to unaffiliated analyst recommendation changes,

indicating that the market pe5rceives affiliated analysts to have more valuable information. In

addition, the consistency between analyst recommendations and asset management investments

is most pronounced for higher quality analysts, again consistent with the increase in division

consistency reflecting quality of information rather than common conflicts of interest across

asset management and analysts.

Across all tests, results provide no support for the idea that the increase in divisional

consistency is driven by pressure on both asset management divisions and analysts to support the

merger. Rather, we find that the extent of divisional consistency is mitigated by conflicts of

interest. For example, we find no significant relation between analyst recommendation changes

and bank investments within those banks where analysts likely face more severe conflicts of

interest. Also, the relation between bank investment decisions and analyst recommendation

changes is concentrated around analyst downgrades; there is less evidence that banks are more

likely to buy shares around analyst upgrades. To the extent that upgrades are more likely driven

by conflicts of interest, this suggests that the increase in division consistency caused by

information sharing is concentrated among those cases where conflicts of interest are least likely

to play a role.

5

Finally, stock returns following changes in holdings show the value of information

around mergers. A strategy of following a subset of analyst recommendations, recommendations

that are most likely to be based on valuable information and least likely to be biased by conflicts

of interest, appears to be a good one. Conditional on a recommendation change, those acquirers

in which the advisor bank increased holdings outperformed those in which the bank decreased

holdings by an average 1.36% per month over the subsequent three months. In contrast, we find

no evidence of excess returns following changes in analyst recommendations that are not

conditional on changes in holdings.

Our paper contributes to the literatures on both information sharing and conflicts of

interest within investment banks. Aggarwal, Prabhala, and Puri (2002), Schenone (2004),

Drucker and Puri (2005), Bodnaruk, Massa, and Simonov (2007), Massa and Rehman (2008),

Michaely and Womack (1999), and Ljungqvist, Marston, and Wilhelm (2005) all document

information sharing within investment banks. Information sharing affects the interactions

between analysts and the investment banking division, the relations between banks and affiliated

mutual funds, the quality of merger advice, and the loan terms between commercial and

underwriting banks. In some cases, such information sharing appears to benefit clients of the

bank, for example providing mutual fund investors with higher returns and causing IPO firms to

be less underpriced. However, in other cases, clients of the bank are harmed, for example by

investing in recent IPOs after observing a positive recommendation by an affiliated analyst.

Moreover, in several high profile cases, investment banks have been sharply criticized for

publicly supporting certain assets that informed players within the bank believed to be

overvalued. We examine the effects of information sharing in a setting in which both likely

downsides and upsides for clients are substantial. Moreover, we consider the effects of

6

information flows between multiple portions of the bank: investment banking, analysts, and

asset management.

Our paper proceeds as follows. Section 2 outlines the data and sample characteristics.

Section 3 tests the divisional consistency versus divisional inconsistency hypotheses, by

empirically examining the relation between analyst recommendations and institutional trades by

the advisor investment bank. Section 4 investigates the ways in which information sharing and

conflicts of interest contribute to divisional consistency. Section 5 provides an analysis of stock

returns, which quantifies the potential gains from considering both the presence of an analyst

recommendation change and the likely information set and incentives behind this change.

Section 6 examines whether non-advisor banks follow advisor analyst. Finally, Section 7

discusses several robustness checks, and Section 8 concludes.

2. Data

2.1 Sample Construction

Our data consists of mergers and acquisitions between 1995 and 2007, as obtained from

the Securities Data Company (SDC) database. To ensure that the merger is a material event for

the acquiring firm, we require the market value of the target to be at least 5% of the combined

market capitalization of the bidder and the target. Both targets and acquirers are public firms

traded in the U.S., and the acquirer must be publicly traded for at least three years prior to the

merger announcement. We require each bidder firm to be followed by at least one analyst, as

listed on the IBES recommendation database, and to be partially owned by at least one

institution, as listed in the Spectrum 13(f) filings, one year prior to the announcement of the

acquisition.

7

Our analysis necessitates merging the SDC merger data, the IBES recommendation data,

and the Spectrum institutional holdings data. For each merger, we identify the advisory

investment bank from SDC. We match by hand the identity of this bank with the IBES broker

code and with the Spectrum institutional name. In matching the institutions between the SDC,

IBES, and Spectrum databases, we are careful to account for both mergers between investment

banks and for banks reporting under different names (e.g., Smith Barney Inc. and Smith Barney

& Co). We attempt to match every investment bank that served as an advisor in at least 10 deals

over our sample period. The only advisors not matched were those such as Houlihan, Lokey,

Howard & Zukin and Greenhill & Co, LLC, neither of which have either a trading desk or

analysts. Mergers in which the advisor either did not have an advisory arm (i.e., wasn’t listed in

IBES), didn’t have a trading arm (i.e., wasn’t listed in Spectrum), or served as an advisor in less

than ten deals are omitted from our sample.

Institutional holdings data are reported in Spectrum quarterly, on March 31st, June 30th,

September 30th, and December 31st of each year. We calculate total shares held by each advisor

institution and each non-advisor institution over the period beginning five quarters prior to the

merger announcement and continuing through five quarters following the merger completion.

For our analysis of analyst recommendations, we obtain from IBES all analyst

recommendations on each acquirer firm.5 We identify the advisor firm recommendation

outstanding three days prior to each institutional trading date, and we aggregate all non-advisor

recommendations outstanding as of this same date into a non-advisor average consensus

recommendation. To remove confounding interests, we do not include advisors to the target

firms in this non-advisor consensus measure. We compute analyst upgrades as cases where an

5 Ljungqvist, Malloy, and Marston (2009) document data problems with IBES tapes. Consistent with the authors’ recommendations to future researchers, our analysis is based on the 2007 IBES download, which is less likely to contain biased data.

8

analyst revised its recommendation upwards, and analogously for downgrades. Kadan,

Madureira, Wang, and Zach (2008) note that many investment banks revised their

recommendations downward in the wake of the Global Settlement, in order to comply with

regulations and present a more balanced set of recommendations, i.e., more equal portions of

optimistic versus pessimistic ratings. The process of the banks reclassifying substantial numbers

of recommendations resulted in large numbers of recommendation changes that were not

information based. As Kadan et al show, banks generally reclassified their outstanding

recommendations within a very short period of time, and these changes did not result in

significant stock price reactions. Based on these findings, changes in recommendations related

to the Global Settlement are not classified in our sample as upgrades and downgrades.6

2.2 Sample Characteristics

As shown in Table 1, these requirements result in a sample of 1,197 mergers. Among

these 1,197, 154 were announced but never completed. Across the mergers, 555 are stock

acquisitions, 196 are cash, and 446 are mixed. Many of the mergers have more than one advisor.

Due to our interest in conflicts of interest at the investment bank level, many of our analyses

focus on advisor-level recommendations and stock ownership. Our sample includes 1,413

advisor-level observations. The sample is spread over time, with the largest number of

transactions occurring in the late 1990s. This concentration is consistent with the finding in prior

literature that M&A activity tends to be particularly high when the stock market is strong.

Looking at the industry distribution, the largest number of mergers is in the business equipment

and finance industries.

6 We thank Leonardo Madureira for providing the dates on which the banks revised recommendations in an effort to comply with the Global Settlement.

9

Table 2 provides descriptive statistics for the full sample. Of greatest interest for our

analysis is analyst coverage and stockholdings of the advisor. Therefore the sample is divided by

whether the advisor has an analyst covering the acquirer and by whether the advisor owns shares

in the acquirer, both measured one quarter prior to the merger announcement. Several

differences become apparent. The acquirers covered by the advisor’s analyst and owned by the

advisor are larger than other acquirers. This finding reflects the more general result that both

analyst coverage and institutional ownership are greater in larger firms, as shown by Gompers

and Metrick (2001) and Barth, Kasznik, and McNichols (2001). The acquirers in which their

advisors own shares and issue analyst recommendations have higher market-to-book ratios,

higher leverage ratios, higher profitability, and lower working capital as a fraction of total assets.

Finally, relative merger size is significantly lower among companies in which advisors provide

analyst coverage and own shares. This difference in relative merger size is potentially driven by

differences in firm size – companies in which the advisor bank owns shares are significantly

larger, meaning a given target size will be relatively smaller.

Table 3 examines the extent to which a bank’s tendency to issue analyst

recommendations or own shares in a firm is related to either expected or recent M&A advisory

business by the investment bank. The analysis begins five quarters prior to the merger

announcement and continues through five quarters following the merger completion (or through

the withdrawal date for non-completed mergers). In conducting this analysis, we assume that an

investment bank’s expectations regarding the acquirer can change substantially during this

period. A bank likely has a much better idea that there is an opportunity to advise a firm in a

merger one quarter prior to the merger announcement than five quarters prior to the

announcement.

10

Table 3 shows an increase in both the advisor’s analyst coverage of the acquirer and in

the advisor’s stockholdings of the acquirer in the period leading up to merger. These increases

are, however, comparable to those of other non-advisors.7 For example, the percent of advisors

with analyst coverage increases from 48% five quarters prior to the merger announcement to

57% one quarter before the merger announcement. While this increase is monotonic across

quarters, we observe no systematic pattern in advisor analysts as a percentage of all analysts

covering the acquirer, which varies between 11.4% and 12.0%. Evidently other analysts are also

picking up coverage of the acquirer during this time. Columns (3) and (4) show that advisor

analyst recommendations are consistently more optimistic than non-advisor recommendations,

where analyst recommendations are measured on a scale from one to five, with one being the

most optimistic.

Turning to the right-hand side of the table, the percent of advisors owning shares of the

acquirer increases from 54% to 58% during the pre-announcement period, but actually decreases

slightly as a fraction of shares owned by all institutions, from 0.94% to 0.89%. In sum, the

results provide little evidence of disproportionate changes in ownership or analyst coverage by

the advisor during this period. Rather, much of the increases in advisor analyst coverage and

advisor share ownership appear to be driven by increases in the size of the acquirer firm over the

quarters prior to the merger announcement. These increases in firm size cause an increase in

overall analyst coverage and institutional ownership in the acquirer firm.

7 As noted previously, non-advisors exclude advisors to both the acquirer and the target firms.

11

3. Are changes in advisors analyst recommendations and stockholdings correlated?

Our tests of the divisional consistency versus inconsistency hypotheses focus on the

relation between changes in the advisor’s analyst recommendations and changes in the advisor’s

stockholdings. If the banks invest in a manner consistent with the advice they provide to clients,

we would expect a positive relation. However, if one of the divisions faces conflicts of interest

different from the other division or if the asset management division has no confidence in the

quality of the information analysts are conveying in their recommendations, we would not expect

to observe divisional consistency.

3.1 Univariate Analysis of Changes in Recommendations and Stockholdings

Table 4 provides descriptive evidence on the relation between analyst recommendations

of the acquirer and stockholdings in the acquirer, by the advisor bank. The panels in this table

show the average change in stockholdings conditional on an analyst upgrade, downgrade, or zero

change in recommendation. We measure changes in the advisor stockholdings of the acquirer in

terms of both raw changes in shares held and percentage changes in shares held, where the

percent change is measured as number of shares held in quarter t minus number of shares held in

quarter t-1, all deflated by the number of shares outstanding in quarter t-1. The data underlying

the analyses represent a panel dataset, with one observation for each acquirer firm advisor in

each quarter.

Table 4 shows that the relation between recommendations changes and stockholdings

varies around the time of the merger. Results for the entire event period (five quarters pre-

announcement through five quarters post-completion) are shown in Panel A of Table 4.

Although it might not be surprising that advisors often upgrade acquirers around a merger,

downgrading is also common. During this period, there were 474 advisor bank downgrades of

12

acquirer firms, 777 upgrades, and 8,755 firm quarters with no change in advisor bank

recommendation. The frequency of downgrades is notable and contrasts strongly with

recommendation patterns following IPOs, where affiliated analysts almost always initiate with

very positive recommendations (see, e.g., Michaely and Womack, 1999). On average across the

474 downgrades, advisors increased their shareholdings by 44,279 shares, compared to an

increase of 84,621 shares in firm quarters with no recommendation change and 122,323 shares in

firm quarters with an analyst upgrade (by the advisor bank).8 The t-stat for the difference

between downgrades and upgrade quarters equals 1.66, significant at the 10% level. Similarly,

we observe a monotonic increase in the percentage change in shares held, as we move from

downgrades, to no recommendation change, to upgrades, however the difference in percentage

changes between downgrade quarters and upgrade quarters is not significant at conventional

levels.

Looking at Panel B, in the five quarters leading up to the merger announcement there are

almost twice as many upgrades of the acquirer by the advisor analysts as downgrades (389 to

199). Results, however, indicate that there is no relation during this pre-announcement period

between changes in these recommendations and changes in advisor stockholdings.

The relation between changes in analyst recommendations and stockholdings is strongest

following the announcement of a merger, defined as the period between the merger

announcement and 5 quarters following merger completion (or through the withdrawal date for

non-completed mergers). Regardless of the measure for change in stockholdings used, the

results indicate that advisors invest significantly more shares in acquirers that their analyst

upgraded than those that they downgraded. Moreover, there is a monotonic increase in both

8 Across the entire sample of firm quarters, both raw changes and percent changes in advisor shareholdings are positive because an increasing number of advisor banks own shares in the acquirer firm over time (as reported in Table 3). In addition, the size of the average position increases slightly over time (not tabulated).

13

measures (from downgrade, to no change, to upgrade). For example, percentage change in

shares held equals 0.13% across the 275 downgrades, 0.45% across the 5,001 firm quarters with

no recommendation change, and 0.88% across the 388 upgrade quarters.

3.2 Regression Analysis of Changes in Recommendations and Stockholdings

Table 5 examines this relation between changes in analyst recommendations and percent

changes in stockholdings in a regression framework. The dependent variable in each regression

equals the percentage change in shares held, as defined previously (number of shares held in

quarter t minus number of shares held in quarter t-1, all deflated by the number of shares

outstanding in quarter t-1). In columns 1 and 2, regression observations include the period

beginning five quarters prior to the merger announcement and extending through five quarters

following the merger completion (or through the withdrawal date for non-completed mergers).

In column 3, the sample is restricted to those quarters preceding the merger announcement, and

in column 4 the sample represents those quarters following the merger announcement.

Regressions are estimated with maximum likelihood, firm fixed effects, and standard errors

clustered by calendar year.

The independent variable of greatest interest in these regressions is the change in advisor

analyst recommendation, defined as the advisor recommendation outstanding immediately prior

to the quarter t holdings date minus the advisor recommendation outstanding immediately prior

to the quarter t-1 holdings date. Analyst recommendations range from one to five, with lower

numbers being more positive. The change in recommendation is multiplied by -1, so that a

positive recommendation change can be interpreted as an upgrade and a negative

recommendation change as a downgrade.

14

Control variables include dummies for the level of the advisor recommendation at the end

of quarter t-1. We only include dummies for strong buy, buy, and hold, because there are fewer

observations with lower recommendations (sells and strong sells). We control for the change in

the consensus recommendation across all non-advisor analysts and for the change in market

capitalization of the acquirer. We also include lagged percent of shares held by the advisor bank

in the acquirer firm, to account for the fact that a bank may be less likely to increase its holdings

if it already holds a substantial number of shares. Finally, we include an estimate of the number

of acquirer shares that the advisor bank would obtain automatically following completion of a

stock merger, as a result of shares previously held in the target. For stock mergers in the first

quarter following merger completion, we estimate this as the number of shares owned in the

target prior to merger completion times the ratio of target to acquirer price one day prior to

merger completion. This variable equals zero for all other firm quarters.

The results show the variation in the relation between changes in recommendations and

stock holdings around mergers. Column 1 shows a positive coefficient on change in advisor

recommendation, consistent with advisors changing their stock holdings in the same direction as

the change in their analyst recommendations. However, the coefficient is not significant at

conventional levels. The results in column 2, however, indicate that the lack of significance over

the entire event period actually combines two very different effects: a highly significant relation

over the post-announcement period and a lack of any significant relation in the pre-

announcement period. The interaction term, change in advisor analyst recommendation * post-

merger dummy, is significantly positive (t-statistic = 3.15). In contrast, the interaction term

advisor analyst recommendation * pre-merger dummy is negative and not significant at

conventional levels (t-statistic = -0.83). A Chi-squared tests indicates that these two coefficients

15

are significantly different (χ2 = 3.52, p-value = 0.06). Columns 3 and 4 yield similar inferences.

In column 3, where the sample only includes pre-announcement firm quarters, we observe no

significant relation between analyst recommendation changes and changes in stock positions.

However, the relation is positive and highly significant during the post-merger announcement

period (column 4).

Our evidence is somewhat inconsistent with the findings of Chan, Cheng, and Wang

(2009) who find a significant relation between analyst recommendations and in-house trading

throughout time. It is possible that their larger sample size (which they obtain by looking at all

firms across a ten-year sample period) gives them more power to find significant differences.

However, as a robustness check we also examine the possibility that our finding of a lack of

significance in the pre-merger announcement period is in some way related to the merger. For

example, if asset management knew of the merger ahead of time, they might be trading on inside

information during this period. In contrast, even if analysts knew of the information ahead of

time, they would be unlikely to convey it in public releases. To examine this, we re-estimate the

pre-merger announcement regression (column 3 of Table 5), using quarters -10 through -6,

relative to the merger announcement. No one within the bank is likely to foresee the merger this

far ahead of time, thereby lessening the probability that merger-related information flows are

affecting results. Results (untabulated) are qualitatively similar to those for quarters -5 through -

1: there is no evidence of a relation between analyst recommendation changes and investments

by asset management.

Results in Table 5 show different patterns in divisional consistency across the pre-merger

announcement versus post-merger announcement periods. In the pre-announcement period,

advisor banks’ investment decisions are completely unrelated to the advice being provided to

16

clients, suggesting the banks themselves do not believe in the information they are publishing.

However, following the merger announcement, advisor banks’ investment decisions are

positively related to the advice being provided to clients. Do analysts have higher quality

information in this post-announcement period, possibly as a result of information sharing from

the investment banking division? Alternatively, is pressure from the investment banking

division causing both analysts to upgrade the acquirer stocks and the asset management division

to purchase acquirer shares (either on its own account or through its mutual funds)? The

distinction is an important one: the first scenario implies a greater value in analyst

recommendations in the post-announcement period, while the second scenario actually implies

the reverse. The following section investigates in more depth the reasons for the greater

divisional consistency following the merger announcement.

4. Information Sharing versus Conflicts of Interest

4.1 Analyst upgrades versus downgrades

As a first step toward understanding the source of the greater divisional consistency in the

post-merger announcement period, we compare the extent to which asset management divisions

(of the advisor bank) invest in response to affiliated analyst upgrades versus downgrades. If the

investment banking division places pressure on both analysts and asset management divisions to

support the acquirer, this pressure would likely take the form of analyst upgrades and stock

purchases. Thus, if such pressure from investment banking explains the increased consistency

during the post-announcement period, we would expect this consistency to be greatest around

analyst upgrades. Alternatively, a finding of a stronger relation around downgrades would

suggest that conflicts of interest are stronger for the analysts, less for the asset management side.

17

The asset management side might be more likely to respond to downgrades because they are less

likely to be biased by conflicts of interest.

Table 6 shows two regressions, similar to those shown in table 5 except that the sample in

column 1 is limited to firm quarters with an analyst upgrade or no recommendation change, and

the sample in column 2 is limited to firm quarters with an analyst downgrade or no

recommendation change. The dependent variable in each is the percentage change in advisor

bank shareholdings in the acquirer, as defined earlier. Independent variables include the change

in advisor bank recommendation in the pre-announcement period, the change in advisor bank

recommendation in the post-announcement period, plus the same control variables as in Table 5.

Across both regressions, upgrades are denoted as a positive recommendation change and

downgrades as a negative recommendation change. Regressions are maximum likelihood, with

firm fixed effects and standard errors clustered by calendar year.

In column 1, we find no relation between analysts upgrades and changes in advisor

holdings. In contrast, column 2 shows a significant relation between downgrades and changes in

advisor holdings of the acquirer. When the advisor analyst downgrades the acquirer, the advisor

is significantly likely to sell more shares.

The finding that changes in the advisor’s shareholdings are only related to analyst

downgrades provides preliminary evidence against the idea that the increase in division

consistency following the merger announcement reflects pressure from the investment banking

division on both analysts and asset management to support the acquirer. Rather, results suggest

that the asset management divisions tend to place more weight on analyst downgrades, perhaps

because they are less likely affected by conflicts of interest.

18

4.2 Abnormal returns to analyst recommendations

If information sharing across divisions increases the accuracy of the advisor’s analyst

recommendations in the post-announcement period, we would expect a greater market reaction

to the recommendation changes in this period. Alternatively, if conflicts of interest decrease the

quality of advisor bank analyst recommendations in the post-announcement period, we would

expect less of a market reaction in this period, particularly for analyst upgrades.

Panel A of Table 7 shows the abnormal return around analyst recommendation changes

in the pre-announcement period and the post-announcement period, where abnormal returns are

defined as the cumulative firm return over days -1 through 1 net of the value-weighted market

return over this same period. Day 0 represents the day of the recommendation change. Row 1

shows the abnormal return across analyst upgrades, and row 2 across analyst downgrades.

Results indicate that the magnitude of the abnormal return is greater in the post-

announcement period, particularly with respect to downgrades. The abnormal return to upgrades

is insignificantly different between the pre-announcement and post-announcement periods.

However, the abnormal return to downgrades is -3.5% in the pre-announcement period versus

-5.9% in the post-announcement period, a difference that is significant at the 5% level (t-statistic

= -2.48).

Moreover, panel B shows that the market reaction to recommendation changes by the

advisor analyst is significantly greater than the market reaction to non-advisor recommendation

changes. Around upgrades, the average abnormal return to advisor analyst recommendation

changes is 2.4%, compared to 1.3% for non-advisor analysts, a difference that is significant at

the 1% level (t-statistic = 2.82). Analogously, around downgrades the average abnormal return

to advisor analyst recommendation changes is -5.9%, compared to -4.3% for non-advisor

19

analysts (t-statistic = 2.06).

The finding that the market reaction to advisor analyst recommendation changes is

stronger in the post-announcement period, combined with stronger market reactions to advisor

analyst recommendation changes than non-advisor analyst recommendations provides additional

evidence that the quality of information behind advisor analyst recommendations is particularly

high following the merger announcement, potentially as a result of information sharing from the

investment banking division. However, the greater market reactions to downgrades than

upgrades suggest that conflicts of interest are also important.

4.3 Level of conflicts of interest in banks, Quality of analysts and sources of revenue

To provide additional evidence into the ways in which information sharing versus

conflicts of interest affect interactions between divisions, this section strives to develop proxies

for both the extent of conflicts of interest within a bank as well as the ability of an analyst to

skillfully interpret information.

To examine the extent of conflicts of interest within investment banks, we classify

investment banks based on their sources of revenue. Following Agrawal and Chen (2008), we

posit that analysts working for institutions in which investment banking is a more important

source of revenue will face greater conflicts of interest, for example stronger pressures to

upgrade stocks of companies for which the bank has recently served as advisor on an acquisition.

If pressure from investment banking also extends to asset management divisions, then we would

expect the higher conflict of interest banks to be more likely to buy these same stocks.

For each publicly traded advisor investment bank, we obtain source of revenue data from

the bank’s 10K. Because not all banks in our sample are publicly traded, this limits us to 25 of

20

the investment banks. However, these 25 banks served as advisors in the majority of our

acquisitions. Investment banks are required to describe the source of their revenues, and the

banks generally break down the revenues into those from investment banking, as well as those

from various other activities on which we are not focusing. Thus, for each of these banks in each

year during which they were publicly traded, we are able to determine the fraction of revenues

from investment banking.

To examine analyst ability we follow Loh and Mian (2006) and use the average prior

forecast error of each advisor bank analyst in our sample, i.e., the analysts at the advisor bank

issuing recommendations on the acquirer firm. For each of these analysts, we collect data on

quarterly earnings forecasts he or she has made (on all firms he or she follows) over the prior

three years, where forecasts consist of the last forecast made prior to the end of the forecasted

firm’s fiscal quarter. We calculate the forecast error as the absolute value of the difference

between the forecast and actual earnings, deflated by the absolute value of earnings.

Although we expect percent of revenues from investment banking to be positively related

to the level of analyst recommendations, we do not expect any relation between analyst quality

and the level of analyst recommendations. Table 8 confirms both these predictions. The

dependent variable is the level of analyst recommendation, re-ordered such that a strong buy

receives the highest possible value (5), while strong sell receives the lowest possible value (1).

Analyst recommendations for each firm are measured at the end of the first quarter following the

merger announcement. Consistent with Agrawal and Chen (2008), we find that

recommendations are significantly more positive for firms that receive a greater portion of

revenues from investment banking (t-stat = 2.35). In contrast, there is no relation between the

analyst forecast error and the level of the recommendation. Control variables indicate that

21

investment banks tend to issue more positive recommendations about larger firms, and they tend

to issue more positive recommendations about firms that are making stock acquisitions.

Table 9 uses these proxies to categorize banks according to magnitude of conflicts of

interest and analyst quality. Specifically, for each year, we classify banks with above-median

(below-median) percent of revenues from investment banking as high (low) investment banking.

Similarly, analysts with above-median (below-median) forecast error are classified as low (high)

quality analysts. Table 9 is restricted to those observations for which we have the source of

revenues for the advisor bank and prior analyst forecasts to compute analyst quality. For each

firm, the sample consists of the first quarter following the merger announcement through five

quarters following the merger completion.

Similar to Table 5, Table 9 shows maximum likelihood regressions of the percent change

in advisor holdings of the acquirer on changes in their analysts’ recommendations, with firm

fixed effects and standard errors clustered by calendar year. The only difference between

column 1 in this table and column 4 in Table 5 is that the sample is restricted to those mergers

for which we have sources of revenues data for the advisor bank and prior forecasts for the

analyst. Similar to prior findings, we find a significant positive relation between changes in

advisor analyst recommendations and changes in advisor bank shareholdings of the acquirer

firm.

In columns 2 and 3, we restrict the sample to high quality and low quality analysts,

respectively, where high quality analysts are those with low forecast errors and vice versa for

low quality analysts. To the extent that information is shared between the investment banking

division and the analysts, we would expect to observe the highest divisional consistency among

high-quality analysts, who are better able to interpret the extra information and translate it into

22

meaningful recommendations. Alternatively, if conflicts of interest drive the increase in

divisional consistency, we might expect greater consistency within the sample of low quality

analysts, who are more easily pressured into issuing biased recommendations. Comparing

results in columns 2 and 3, we see results consistent with information sharing across divisions:

the relation between analyst recommendation changes and changes in advisor bank

shareholdings is only significant among the high quality analysts (t-statistic = 1.92).

Columns 4 and 5 similarly split the sample into two groups, but now the split is based on

percent of revenues from investment banking, with high IB banks in column 4 and low IB banks

in column 5. If both the analyst and the asset management divisions face pressure from the

investment banking division to support the acquirer, then we would expect the greatest divisional

consistency among high IB banks, where such conflicts are likely greatest. However, results

provide no support for the idea that the increase in division consistency reflects common

conflicts of interest across both asset management and analysts. Among banks most subject to

such conflicts (high IB banks), we observe no significant relation between analyst

recommendation changes and asset management investment decisions. Rather, the relation is

concentrated among low IB banks, where such conflicts are likely less severe.

In sum, results show that information sharing between the investment banking division

and analysts increases the value of analyst recommendations in the post-merger announcement

period, contributing to greater divisional consistency. They also, however, indicate that conflicts

of interest among analysts affect this relation. In particular, consistency is greater for analyst

downgrades than upgrades and greater when investment banking accounts for a small proportion

of the advisor’s revenues.

23

5. Returns from the advisors investment decisions following changes in recommendations

Results in Tables 4 through 9 suggest that the advisor bank selectively follows the subset

of analyst recommendations that is most likely to contain value-relevant information and least

likely to be driven by conflicts of interest. As an approximation of the bank’s level of success in

this strategy, we examine returns subsequent to changes in advisor shareholdings that are

conditional on analyst recommendation changes (i.e., returns from banks listening to their

analysts.) A finding that returns following the advisor’s purchases exceed returns following the

advisor’s sales suggests that the bank gained through its attention to not just the analyst

recommendations, but also the information set and incentives behind these recommendations.

As our focus is on banks’ response to recommendation changes, we define our sample as

firm quarters in which the advisor bank analysts changed their recommendation of the acquirer

firm. We estimate calendar time, four-factor regressions (Fama and French (1993) and Carhart

(1997)), where the dependent variable equals returns net of the risk free rate for acquirers who

have experienced a recommendation change by their advisor bank over the past 3 months (6 and

12 months). Specifically, a firm enters the sample on the first institutional reporting date (i.e.,

March 31, June 30, Sept. 30, Dec. 31) following the advisor analyst recommendation change,

and it stays in the sample for 3 months (6, 12 months). In Panel A of Table 10, we consider

portfolios of firms representing (1) a long position in acquirers in which the advisor bank

increased shares in the quarter of the recommendation change, (2) a long position in acquirers in

which the advisor bank decreased shares in the quarter of the recommendation change, and (3) a

long position in portfolio 1 combined with a short position in portfolio 2. The table reports

alphas (i.e., intercepts) from these four-factor models.

Looking first at Panel A, we see that acquirers in which the advisors increased positions

24

outperformed those in which the advisors decreased positions (following a recommendation

change by one of their analysts) by an average 136 basis points per month (significant at the 5%

level) over the first three months following the institutional reporting date. Return differences

are not statistically significant over longer intervals.

Panel B of Table 10 shows a similar analysis, except that firms are placed into terciles

based on the change in shareholdings of the advisor in the acquirer firm (rather than just

conditioning on buy versus sell). We consider portfolios of firms representing (1) a long position

in acquirers in the top tercile based on changes in shareholdings by the advisor bank, (2) a long

position in acquirers in the bottom tercile based on changes in shareholdings by the advisor bank,

and (3) a long position in portfolio 1 combined with a short position in portfolio 2. Similar to

inferences in Panel A, results indicate that this long-short portfolio produced significantly

positive returns of 154 basis points per month over the 3-month horizon.

Notably, the findings of excess returns are limited to recommendation changes that are

accompanied by changes in holdings, not just recommendation changes in general. In

untabulated tests we examine returns following all advisor analyst recommendation changes, and

find no evidence of significantly higher returns following advisor firm upgrades versus

downgrades. The finding that changes in the advisor’s analyst recommendations do not

unconditionally predict returns contrasts with the findings of Jegadeesh et al (2004), who find

that the quarterly change in consensus recommendations does predict returns. The results are

consistent with changes in recommendations reflecting an increase in both conflicts of interest

and information around mergers. Returns indicate that asset managers can sort the importance of

these factors when deciding whether to trade on these recommendations.

25

6. Do non-advisor institutions similarly rely on advisor bank recommendations?

Our results suggest that analysts to the advisor bank have more information regarding the

acquirer firm, on average. However, the extent of the greater reliability in the advisor analysts’

recommendations varies systematically across time, type of bank, and type of analyst. Empirical

tests that classify banks and analysts solely on readily available public information show that

banks only change their stock positions in response to analyst recommendations that are less

likely driven by conflicts of interest and more likely to contain value-relevant information,

suggesting that these are the only recommendations that have investment value. Following this

logic, if these less-biased recommendations do in fact have investment value, other institutions

might also be expected to base their trades on them. To examine this proposition, we re-estimate

the regressions from Table 5, using the percent change in advisor shares held net of the average

percent change in non-advisor shares held. If other institutions similarly consider the advisor

analysts’ recommendations in light of the likely information and likely conflicts of interest

behind such recommendations, the advisor analyst recommendation change should not be

statistically significant in these regressions.

The results from these regressions are shown in Table 11. Notably, the coefficients on

the analyst recommendation variables are comparable to those in Table 5. The relation between

analyst recommendation changes and raw changes in percent advisor shares held is qualitatively

similar to that between analyst recommendation changes and net percent changes in advisor

shares held. These findings suggest that the advisor bank changes its positions in the acquirer

stock significantly more than other institutional traders, in response to advisor analyst

recommendation changes.

This finding is consistent with the long-run returns evidence. There is no significant

26

difference in long-run returns for a strategy of buying all post-merger announcement advisor

analyst upgrades and shorting the analogous downgrades. Advisors primarily trade on the

subset of recommendations that is most likely to contain value-relevant information and is least

likely to be biased by conflicts of interest. The combination of findings presented in this paper

suggests that this determination is based on both private and public information, and is not

replicable by people outside of the bank. Moreover, the higher returns earned by advisor banks

only extend to a three-month horizon. By the time that advisor banks’ holdings are publicly

released (six weeks following the end of the quarter), the opportunity for higher returns no longer

exists.

7. Robustness Checks

7.1 Regulatory Changes

During our sample period, there were several important regulatory changes that altered

the structure of analyst recommendations and also the way that analysts could interact both with

outside investors and with other parts of the investment bank. In October 2000, Regulation Fair

Disclosure (Reg FD) required that all publicly traded companies disclose any material

information to all investors at the same time. Following the implementation of this rule, analysts

were no longer able to obtain information by calling companies directly; companies had to

provide any information to the entire public. In April 2003, the Global Settlement was reached,

which included a variety of provisions to address the conflicts of interest within investment

banks. For example, the requirement that investment banking departments and analysts be

separated via Chinese walls was strengthened, analysts were prohibited from going on IPO road

shows, and analyst compensation had to be independent of investment banking business. In

27

addition, in 2002 many brokerage houses refined their recommendation system, going from a

five-tier scale to a three-tier scale and making the ratio of optimistic to pessimistic

recommendations more balanced (see Kadan, Madureira, Wang, and Zach (2008) for a complete

discussion). These regulations should lessen the extent to information flow between the

investment banking division and analysts.

In an attempt to shed some light on the extent to which the dynamics observed in this

paper extended throughout our sample, we divide our sample into two parts: 1995 – 2000, and

2003 – 2007. While the smaller sample sizes weaken statistical significance, the magnitude of

the coefficient on analyst recommendation change*post-announcement period is approximately

equal in the two sub-samples (results not tabulated). Therefore, the importance of conflicts of

interest and information sharing persist after these regulatory changes.

We note that even in the presence of Chinese walls, those within an institution can have

greater insights into the factors affecting the actions of the institution’s other divisions, compared

to outsiders . Specifically, analysts may have better ability to interpret the actions of the

investment banking division and infer the quality of the merger, and asset management divisions

may have better ability to interpret the importance of conflicts of interest and information content

in analyst recommendations. Common knowledge regarding company practices, compensation

schedules, incentive schemes, etc. likely increase understanding of the true meaning of certain

actions.

7.2 Econometric Specifications

The sample underlying many of our regressions represents a panel dataset. As discussed

by Peterson (2008), the appropriate handling of both standard errors and fixed effects in such

28

samples is paramount. Following Peterson, we have included firm fixed effects and clustered

standard errors by calendar year. The firm fixed effects allow for the fact that there may be

company-specific factors (i.e., characteristics of the acquirer firm) that affect a bank’s tendency

to invest in the company, but which we have not controlled for. The clustering of standard errors

allows for the fact that both analyst recommendations and institutional investment vary over

time, for example due to regulatory changes as discussed in the prior subsection and

macroeconomic conditions. Results are also robust to specifying the regressions using calendar

year fixed effects and clustering standard errors on the deal level.

7.3 Advisors to the Target

If the advisor to the target firm gains an intimate knowledge of the acquirer through the

merger negotiation process, then it is possible that this information may be similarly shared

among other divisions of the target firm. Specifically, the investment banking division of the

target advisor may share information regarding the acquirer with the target advisor’s analysts

and/or asset management divisions. If such value-relevant information sharing takes place, then

we would expect to see similar relations between analyst recommendation changes and asset

management investments (of the acquirer) within the target advisor bank. Alternatively, if the

advisor bank learns more about the long-run acquirer value by advising the acquirer than the

target, then the target bank may learn substantially less about the future value of the acquirer.

Also, many banks have long-run relationships with clients, suggesting that the knowledge of the

acquirer advisor may not stem just from this one merger, but rather from a long association with

the company.

To examine whether the analysts working for the target advisor similarly have more

29

valuable information around the time of the merger, we examine the relation between changes in

analyst recommendations and asset management investments, by the target advisor in the

acquirer firm. Results show no significant relation, either in the pre-merger or post-merger

periods. The results support the notion that asset management divisions of the target advisor do

not consider their analysts to have abnormally valuable insights into the value of the acquirer.

Results suggest that acquirer advisors either learn more about the value of the merger (compared

to target advisors) or benefit from a longer-term relationship with the acquirer.

7. Conclusion

The potential for information sharing is pervasive in investment banks. Such information

sharing can be either beneficial or detrimental to various clientele of the bank, depending on the

extent to which it results in higher quality information and better investments, versus biased

information and suboptimal investments. We investigation this issue using a novel approach, by

examining the association between an investment bank’s own response to its analysts

recommendations.

Consistent with Ljungqvist, Marston, Starks, Wei, and Yan (2007) and with Fang and

Yasuda (2008), our results indicate the quality of analyst advice varies in predictable ways.

Specifically, we find recommendations to be more value-relevant following an investment

banking event as more value-relevant information flows from the investment banking division, in

banks that rely less on investment banking as a source of revenue and are therefore less subject

to conflicts of interest, and among higher quality analysts that are better able to distill the newly

available information into a meaningful recommendation.

We find that banks consider this variation in recommendation quality when making their

30

investment decisions. Changes in advisor firm stockholdings in the acquirer are based on both

the recommendation changes of its analysts and the likely information set and incentives behind

these recommendation changes. This attention to detail is rewarded: returns to firms that banks

purchased conditional on an analyst recommendation change are significantly higher than those

in that the banks sold.

Finally, the findings have implications for the literature on the diversification of activities

of financial conglomerates. Supporting arguments of agency problems from diversification,

prior work such as Delong (2001) and Laeven and Levine (2007) find that increased

diversification destroys – or at least does not create value. Our work shows that the information

benefits that an institution realizes from offering diverse activities, as suggested by Stein (2002)

and others, depend on divisional incentives and information environment. Specifically, the

benefits an asset management division realizes from the institution’s other activities are

decreasing in the conflicts of interest for analysts and increasing in the information generated

from investment banking.

31

References

Aggarwal, R., Prabhala, N., Puri, M., 2002. Institutional allocation in initial public offerings:

empirical evidence. Journal of Finance 57, 1421-1442.

Agrawal, A., Chen, M., 2008. Do analyst conflicts matter? Evidence from stock

recommendations. Journal of Law and Economics 51, 503-537.

Barth, M., Kasznik, R., McNichols, M., 2001. Analyst coverage and intangible assets. Journal

of Accounting Research 39, 1-34.

Becher, D., Juergens, J., 2005. Analyst recommendations and mergers: do analysts matter?

Unpublished working paper. Arizona State University.

Bodnaruk, A., Massa, M., Simonov, A., 2007. Investment banks as insiders and the market for

corporate control. Review of Financial Studies, forthcoming.

Carhart, M., 1997. On persistence in mutual fund performance. Journal of Finance 52, 57-82.

Chan, K., Chang, C., Wang, A., 2009. Put your money where your mouth is: do financial firms

follow their own recommendations? The Quarterly Review of Economics and Finance,

forthcoming.

Chen, X., Cheng, Q., 2002. Institutional holdings and analysts’ stock recommendations.

Unpublished working paper. University of British Columbia.

Delong, G., 2001. Stockholder gains from focusing versus diversifying bank mergers. Journal of

Financial Economics 59, 221-252.

Drucker, S., Puri, M., 2005. On the benefits of concurrent lending and underwriting. Journal of

Finance 60, 2763-2799.

Fama, G., French, K., 1993. Common risk factors in the returns on stocks and bonds. Journal of

Financial Economics 33, 3-56.

Fang, L., Yasuda, A., 2008. The effectiveness of reputation as a disciplinary mechanism in sell-

side research. Review of Financial Studies, forthcoming.

Gompers, P. and A. Metrick, 2001. Institutional investors and equity prices. Quarterly Journal

of Economics 116, 229-259.

Groysberg, B., Healy, P., Chapman, C., Shanthikumar, D., Gui, Y., 2007. Do buy-side analysts

outperform the sell-side? Unpublished working paper. Harvard Business School.

Jegadeesh, N., Kim, J., Krische, S., Lee, C., 2004. Analyzing the analysts: when do

recommendations add value? Journal of Finance 59, 1083-1124.

32

33

Kadan, O., Madureira, L., Wang, R., Zach, T., 2008. Conflicts of interest and stock

recommendations: the effects of the global settlement and related regulations. Review of

Financial Studies, forthcoming.

Kolanski, A., Kothari, S.P., 2009. Investment Banking and Analyst Objectivity: Evidence from

Analysts Affiliated with M&A Advisors. Journal of Financial and Quantitative Analysts,

forthcoming.

Laeven, L., Levine, R., 2007, Is there a diversification discount in financial conglomerates?

Journal of Financial Economics 85, 331-367.

Ljungqvist, A., Malloy, C., Marston, F., 2008. Rewriting history. Journal of Finance,

forthcoming.

Ljungqvist, A., Marston, F., Starks, L, Wei, K., Yan, H., 2007. Conflicts of interest in sell-side

research and the moderating role of institutional investors. Journal of Financial

Economics 85, 420-456.

Loh, R., Mian, G., 2006. Do accurate earnings forecasts facilitate superior investment

recommendations? Journal of Financial Economics 80, 455-483.

Massa, M., Rehman, Z., 2008. Information lows within financial conflomerates: evidence from

the banks-mutual funds relation. Journal of Financial Economics, forthcoming.

Mehran, H., Stulz, R., 2007. The economics of conflicts of interest in financial institutions.

Journal of Financial Economics 85, 267-296

Michaely, R., Womack, K., 1999. Conflict of interest and the credibility of underwriter analyst

recommendations. Review of Financial Studies 12, 653-686.

Moeller, S., Schlingemann, F., Stulz, R., 2005. Wealth destruction on a massive scale? a study

of acquiring-firm returns in the recent merger wave. Journal of Finance 60, 757-782.

Peterson, M., 2009. Estimating standard errors in finance panel data sets: comparing approaches.

Review of Financial Studies 22, 435-480.

Schenone, C., 2004. The effect of banking relationships on the firm’s IPO underpricing. Journal

of Finance 59, 2903-2958.

Stein, J., 2002. Information production and capital allocation: Decentralized versus hierarchical

firms. Journal of Finance 57, 1891 -1921.

Table 1: M&A Sample The sample consists of 1,197 mergers over the 1995 to 2007 period. For a merger to be included in the sample, the acquirer firm must be followed by at least one analyst, as listed in the IBES database, and be owned by at least one institutional investor, as listed in the Spectrum database, one year prior to the merger announcement. The target market capitalization must be at least 5% of the combined market capitalization of the target plus acquirer, where all market capitalizations are measured one month prior to the merger announcement. A merger with two advisors is treated as two advisor-level observations; there are 1,413 advisor observations across the 1,197 mergers. Mergers are classified into industries based on the Fama-French 12 industry groupings.

Relative Size > 5%

Number of advisor observations 1,413

Number of unique mergers 1,197

Withdrawn 154

Completed 1,043

Stock 555

Cash 196

Mixed 446

Year # Mergers Industry # Mergers 1995 99 Consumer Nondurables 25

1996 106 Consumer Durables 10

1997 168 Manufacturing 78

1998 166 Oil, gas, coal extraction 50

1999 125 Chemicals and allied products 24

2000 105 Business Equipment 197

2001 63 Telephone & TV transmission 34

2002 40 Utilities 36

2003 73 Wholesale, Retail 74

2004 74 Healthcare, Med. Eqpt, Drugs 105

2005 58 Finance 307

2006 68 Other 257

2007 52

34

Table 2: Descriptive Statistics Descriptive statistics are provided for the sample of 1,197 mergers over the 1995 – 2007 time period. All variables, with the exception of relative merger size, refer to the acquirer firm, and all statistics represent medians. Market capitalization (in millions) is measured one month prior to the announcement of the merger. All other financial variables are measured at the fiscal year end preceding the merger announcement. Market-to-book equals the equity market capitalization divided by the book value of equity. Book leverage equals the sum of short-term and long-term debt, divided by total assets. Market leverage equals the sum of short-term and long-term debt divided by the total firm market value, where total firm market value equals total assets plus market value of equity minus the book value of equity. Total assets, sales, sales/TA, EBIT/TA, and WC/TA are computed using the relevant Compustat data items. Relative merger size equals the target market capitalization divided by the combined market capitalization of the target plus acquirer, where all market capitalizations are measured one month prior to the merger announcement. Statistics are computed for the whole sample, conditional on whether or not the advisor bank to the acquirer firm has an analyst issuing recommendations on the acquirer one quarter prior to merger announcement, as listed on IBES, and conditional on whether or not the advisor bank to the acquirer firm owns shares in the acquirer firm one quarter prior to merger announcement, as reported on Spectrum. Asterisks denote whether the advisor analyst vs. no advisor analyst statistics are significantly different, and similarly whether the advisor institutional ownership vs. no advisor institutional ownership are significantly different (*, **, *** represent the 10, 5, and 1% levels of significance).

Whole Sample (n=1,197)

Advisor Analyst

Following (n=810)

No Advisor Analyst

Following (n=387)

Advisor Institutional Ownership

(n=778)

No Advisor Institutional Ownership

N=419) Market Cap (mil) 1,825 2,410 1,163*** 2,778 792***

Total Assets (mil) 1,759 1,928 1,464*** 2,565 881***

Sales (mil) 823 959 669*** 1,320 385***

Sales / TA 0.65 0.65 0.61 0.66 0.60

MB 2.28 2.34 2.12*** 2.35 2.17*

Book leverage 0.22 0.23 0.18** 0.23 0.17***

EBIT / TA 0.07 0.07 0.07* 0.08 0.06**

WC / TA 0.19 0.18 0.22* 0.17 0.25***

Relative Merger Size 0.28 0.25 0.33*** 0.25 0.32***

35

36

Table 3: Incidence of advisor recommendations and share ownership in the acquirer companies This table provides information on the incidence of advisor recommendations and advisor institutional ownership in the acquirer company, from five quarters prior to the announcement of the merger to five quarters following the completion of the merger (or through the withdrawal date for non-completed mergers). Percent of advisors represents the percentage of the 1,460 advisor-level observations in which the advisor bank to the acquirer had an analyst following the acquirer. Percent of total recs by advisor equals the number of advisors covering the firm divided by the total number of analysts following the firm, averaged across the 1,197 mergers. Average advisor rec equals the average advisor analyst recommendation in the acquirer, where recommendations vary from 1 to 5 with 1 being the most positive. Average non-advisor rec equals the average analyst recommendation in the acquirer, across all non-advisor analysts. Percent of advisors that own shares represents the percentage of the 1,460 advisor-level observations in which the advisor bank to the acquirer owned shares in the acquirer. Advisors as a % of total equals the number of advisors owning shares in the firm divided by the total number of institutions owning shares in the firm, averaged across the 1,197 mergers. Percent of advisors that issue recs and own shares equals the percent of the 1,460 advisors to the acquirer firms that both have an analyst following the acquirer and own shares in the acquirer. Company mkt cap equals the median market capitalization of the acquirer firm.

Issuance of Recommendations Ownership of Shares

% of Advisors

% of Total Recs that

are by Advisor

Avg Advisor

Rec

Average non-

Advisor Rec

% of Advisors that Own

Shares

Advisors as % of

total insts

% of Advisors that Issue Recs and

own Shares

Company Mkt Cap ($mil)

5 qtrs pre- ann’t 48% 11.4 2.07 2.11 54% 0.94% 33% $1,553 4 qtrs pre- ann’t 51% 12.0 2.09 2.11 55% 0.90% 34% $1,663 3 qtrs pre- ann’t 53% 12.1 2.09 2.11 56% 0.90% 36% $1,745 2 qtrs pre- ann’t 55% 11.9 2.05 2.12 56% 0.89% 37% $1,875 1 qtr pre- ann’t 57% 12.0 2.04 2.12 58% 0.89% 39% $2,020

1 qtr post-ann’t 57% 11.4 2.04 2.11 59% 0.93% 39% $2,220

1 qtr post-completion 61% 11.9 1.96 2.07 62% 0.75% 44% $2,633 2 qtrs post-completion 62% 12.3 1.96 2.08 64% 0.78% 46% $2,707 3 qtrs post-completion 63% 12.2 1.99 2.09 63% 0.80% 46% $2,817 4 qtrs post-completion 64% 12.0 2.01 2.13 64% 0.81% 47% $2,798 5 qtrs post-completion 62% 11.7 2.03 2.16 63% 0.79% 47% $2,769

Table 4: Relation between advisors’ recommendations changes and holdings changes Each panel tabulates the number of quarters in which the advisor bank upgraded, downgraded, and made no recommendation change to the acquirer firm. Panel A is based on the period beginning 5 quarters prior to the merger announcement and extending through 5 quarters after merger completion (or through the withdrawal date for non-completed mergers). Panel B focuses on the pre-announcement quarters and Panel C on the post-announcement quarters. Each panel shows the number of downgrades and two measures of changes in advisor ownership of the acquirer firm in the quarters of these downgrades (and similarly for quarters with no recommendation change and for quarters with upgrades). The first measure is the change in advisor shares held of the acquirer, from quarter t-1 to quarter t. The second measure is the percentage change in advisor shares held, defined as shares owned at the end of quarter t minus shares owned at the end of quarter t-1, all deflated by shares outstanding at the end of quarter t-1. T-tests are for differences between downgrades and upgrades. Panel A: Differences in Advisor positions from 5 qtrs pre-ann’t – 5 qtrs post-completion

Adv. Downgrade No Adv. Rec

Change Adv. Upgrade T-Test

Number Observations 474 8,755 777 ∆ Advisor Shares Held

44,279 84,621 122,323 1.66*

%∆ Advisor Shares Held 0.33% 0.46% 0.67% 1.49

Panel B: Differences in Advisor positions from 5 qtrs pre-ann’t through 1 qtr pre-annt

Adv. Downgrade No Adv. Rec

Change Adv. Upgrade T-Test

Number Observations 199 3,754 389 ∆ Advisor Shares Held 79,761 64,709 65,371 0.24

%∆ Advisor Shares Held 0.61% 0.48% 0.46% 0.42

Panel C: Differences in Advisor positions from 1 qtr post-annt through 5 qtrs post-completion

Adv. Downgrade No Adv. Rec

Change Adv. Upgrade T-Test

Number Observations 275 5,001 388 ∆ Advisor Shares Held

18,603 99,567 179,421 2.19**

%∆ Advisor Shares Held 0.13% 0.45% 0.88% 2.31**

37

Table 5: Determinants of change in acquirer shares held by the advisor investment bank This table shows firm fixed effects maximum likelihood regressions, where the dependent variable is percent change in advisor bank ownership in the acquirer, defined as shares owned at the end of quarter t minus shares owned at the end of quarter t-1, all deflated by shares outstanding at the end of quarter t-1. Regressions are estimated over the 1,460 advisor-level observations, for five quarters prior to the merger announcement to five quarters following the merger completion (or through the withdrawal date for non-completed mergers). The first independent variable is the change in the advisor analyst recommendation of the acquirer company (measured over the same quarter but observed prior to the measurement of institutional ownership), and the second independent variable is the change in average non-advisor recommendation. In column 2, these recommendation changes are interacted with both a pre-merger dummy and a post-merger dummy (equal in 1 in the quarters prior to and following the merger announcement, respectively, 0 otherwise). All recommendation changes are multiplied by negative 1, such that higher recommendations and increases in recommendations can be interpreted as more optimistic. Dummies for the level of the advisor recommendation at the beginning of the quarter are included (strong buy, buy, and hold). The change in market capitalization represents the change in the market capitalization of the acquirer company over the quarter. Pct held by Advt=-1 equals the percent of outstanding shares held by the acquirer advisor in the acquirer firm one quarter prior. For stock mergers in the first quarter following merger completion, Qtr t-1 Dum*Shares assumed in acquirert=+1equals the number of shares held by the advisor in the target firm in the previous quarter; for all other firm quarters this variable equals 0. Standard errors are clustered by calendar year, and Z-statistics are shown in parentheses.

Dep’t Var = % change in Advisor Shares Held (in acquirer)

5 qtrs pre-annt through 5 qtrs post-completion

Pre-ann’t period

Post-ann’t period

ΔRec (Advisor) 0.16

(1.36) -0.12 (-0.46)

0.35***

(2.67) ΔRec (Non-Advisor) -0.18

(-0.83) -0.27 (-0.86)

-0.03 (-0.13)

ΔRec * Pre- Merger (Advisor) -0.13

(-0.60)

ΔRec * Post Merger (Advisor) 0.40***

(3.15)

ΔRec * Pre Merger (Non-Adv) -0.31

(-0.97)

ΔRec * Post Merger (Non-Adv) -0.06

(-0.29)

Strong Buy Dummy (Advisor) 0.16 (0.64)

0.10 (0.38)

0.84*

(1.68) -0.14

(-0.43) Buy Dummy (Advisor) -0.14

(-0.92) -0.19

(-1.25) -0.07

(-0.34) -0.28

(-0.95) Hold Dummy (Advisor) -0.11

(-0.46) -0.15

(-0.61) 0.13

(0.41) -0.47

(-1.41) ΔMkt Cap 15.05***

(2.62) 15.11***

(2.60) 12.19 (0.79)

7.61*

(1.70) Pct held by Advt-1 -132.7***

(-7.44) -132.9***

(-7.49) -145.8***

(-8.76) -193.8***

(-6.02) Qtr t-1 Dum*Shares assumed in acquirert=+1

0.61***

(2.81) 0.61***

(2.81) 0.70***

(3.05)

N Obs 9,126 9,126 3,764 5,253

38

Table 6: Determinants of change in acquirer shares held by the advisor investment bank, Upgrades vs Downgrades This table shows firm fixed effects maximum likelihood regressions, where the dependent variable is percent change in advisor bank ownership in the acquirer, defined as shares owned at the end of quarter t minus shares owned at the end of quarter t-1, all deflated by shares outstanding at the end of quarter t-1. Across the 1,460 advisor-level observations and five quarters prior to merger announcement through five quarters post-merger completion sample (or through the withdrawal date for non-completed mergers), column 1 includes firm quarters with an analyst upgrade or no recommendation change (excluding cases where the previously outstanding advisor recommendation was a strong guy) and column 2 includes firm quarters with an analyst downgrade or no recommendation change. The first independent variable is the change in the advisor analyst recommendation of the acquirer company (measured over the same quarter but observed prior to the measurement of institutional ownership), and the second independent variable is the change in average non-advisor recommendation. In column 2, these recommendation changes are interacted with both a pre-merger dummy and a post-merger dummy (equal in 1 in the quarters prior to and following the merger announcement, respectively, 0 otherwise). All recommendation changes are multiplied by negative 1, such that higher recommendations and increases in recommendations can be interpreted as more optimistic. Dummies for the level of the advisor recommendation at the beginning of the quarter are included (strong buy, buy, and hold). The change in market capitalization represents the change in the market capitalization of the acquirer company over the quarter. Pct held by Advt=-1 equals the percent of outstanding shares held by the acquirer advisor in the acquirer firm one quarter prior. For stock mergers in the first quarter following merger completion, Qtr t-1 Dum*Shares assumed in acquirert=+1equals the number of shares held by the advisor in the target firm in the previous quarter; for all other firm quarters this variable equals 0. Standard errors are clustered by calendar year, and Z-statistics are shown in parentheses.

Dep’t Var = % change in Advisor Shares Held (in acquirer)

Upgrades Downgrades

ΔRec * Pre- Merger (Advisor) -0.24 (-0.98)

-0.09 (-0.28)

ΔRec * Post Merger (Advisor) 0.18 (0.86)

0.56**

(1.95) ΔRec (Non-Adv) -0.28

(-1.17) -0.30

(-1.25) Strong Buy Dummy (Advisor)

0.31 (1.05)

Buy Dummy (Advisor) -0.21 (-1.36)

-0.08 (-0.30)

Hold Dummy (Advisor) -0.15 (-0.66)

-0.08 (-0.21)

ΔMkt Cap 26.26*

(2.56) 17.91**

(3.16) Pct held by Advt-1 -166.50***

(-6.78) -137.34***

(-7.52) Qtr t-1 Dum*Shares assumed in acquirert=+1

0.88 (1.35)

0.44***

(2.55)

N Obs 6,069 8,429

39

Table 7: Abnormal Returns to Analyst Recommendation Changes This table shows the three day abnormal return to analyst recommendation changes. Abnormal returns are computed as the cumulative return to the stock around days -1 to 1, net of the cumulative return to the value-weighted index around the same days, where day 0 is the day of the recommendation change. The pre-announcement period consists of one to five quarters prior to the merger announcement, and the post-announcement period consists of one quarter following merger announcement through five quarters following merger completion (or through the withdrawal date for non-completed mergers). T-statistics in Panel A test the difference between the pre-announcement and post-announcement periods, and in Panel B between recommendation changes by advisor bank analysts versus non-advisor bank analysts. Panel A: Pre-Announcement Post-Announcement T-statistic

All Rec Changes (absolute value of abnormal return)

4.9% 6.4% 3.19***

Upgrades (abnormal return) 2.13% 2.42% 0.54

Downgrades (abnormal return) -3.5% -5.9% -2.48**

Panel B: Advisor Non-Advisor T-statistic

Upgrades (abnormal return) 2.4% 1.3% 2.82***

Downgrades (abnormal return) -5.9% -4.3% 2.06**

40

41

Table 8: Analyst Recommendations, conditional on conflicts of interest and analyst quality proxies

This table shows an OLS cross-sectional regression of advisor bank analyst recommendations on the percent of total bank revenue from investment banking, the average prior forecast error of the analyst, and control variables. Analyst recommendations in each firm are measured at the end of the first quarter following the merger announcement. Recommendations are ordered from 1 to 5, with 5 being the most positive. Control variables include: acquirer market capitalization, measured one quarter prior to merger completion; the number of shares held by the advisor bank in the acquirer company at the end of the previous quarter; abnormal market return to the merger announcement; stock and cash dummies denoting the method of payment in the merger; and the relative size of the merger, defined as the target market capitalization divided by the market capitalization of the new, combined company. T-statistics are reported in parentheses.

Dep’t Var Adv Rec

Constant -2.23***

(-24.11)

%Rev from IB 0.59** (2.35)

Analyst forecast error 0.08 (1.21)

Acquirer Mcap 4.01* (2.65)

Shrs Held Advt-1 1.03 (0.26)

M&A annt AR 0.25 (0.59)

Stock dummy 0.24*** (2.68)

Cash dummy -0.20 (-1.51)

Relative Size -0.03 (-0.37)

N Obs 418

Adj R-squared 5.1%

Table 9: Change in advisor shares of acquirer, conditional on conflicts of interest and analyst quality proxies This table shows firm fixed effects maximum likelihood regressions, where the dependent variable is percent change in advisor bank ownership in the acquirer, defined as shares owned at the end of quarter t minus shares owned at the end of quarter t-1, all deflated by shares outstanding at the end of quarter t-1. Regressions are estimated over the advisor-level observations that have data on both advisor bank source of revenue and prior analyst forecast accuracy, for one quarter following the merger announcement to five quarters following the merger completion (or through the withdrawal date for non-completed mergers). Column 1 includes all advisors with available data, column 2 (3) includes those advisor analysts with above median (below median) prior forecast accuracy. Column 4 (5) includes advisor analysts working at banks that receive an above median (below median) percent of revenues from investment banking. The change in the advisor analyst recommendation of the acquirer company is measured over the same quarter but observed prior to the measurement of institutional ownership. All recommendation changes are multiplied by negative 1, such that increases in recommendations can be interpreted as more optimistic. All other control variables are defined in Table 5. Standard errors are clustered by calendar year, and Z-statistics are shown in parentheses. Dep’t Var = % change in Advisor Shares Held (in acquirer)

Analysts at All Banks High quality analysts Low quality analysts Analysts at

High IB Banks Analysts at

Low IB Banks ΔRec (Advisor) 0.66***

(3.01) 0.67*

(1.92) 0.54

(1.16) 0.35

(0.90) 0.57**

(2.17) ΔRec * (Non-Adv) -0.20

(-0.60) 0.78

(1.37) -0.34

(-0.55) 0.37

(0.70) 0.01

(0.01) Strong Buy Dummy (Advisor) 0.61

(1.23) 0.75

(1.23) 0.27

(0.25) -1.45

(-1.52) 1.56

(1.68) Buy Dummy (Advisor) 0.54

(1.30) -0.23

(-0.69) 0.98

(1.16) -1.13

(-1.38) 1.71**

(2.09) Hold Dummy (Advisor) 0.09

(0.22) -0.69

(-1.50) 0.16

(0.18) -1.67**

(-2.26) 1.24

(1.52) ΔMkt Cap 6.55

(1.04) 17.33***

(3.88) 1.22

(0.12) 9.45

(0.98) 14.98 (0.82)

Pct held by Advt-1 -257.89*** (-8.52)

-317.32*** (-4.26)

-297.17*** (-5.91)

-372.41*** (-7.33)

-240.76*** (-3.69)

Qtr t-1 Dum*Shares assumed in acquirert=+1

1.03*** (3.45)

0.44*** (3.51)

1.43** (2.37)

0.78*** (3.85)

1.48*** (-3.69)

N Obs 2,924 1,063 1,092 1,115 1,040

42

Table 10: Returns to acquirers, conditional on changes in shareholdings of the advisor This table shows alphas of acquirers’ monthly returns computed from four factor regressions using Fama French and momentum factors. Returns are computed following the announcement of a merger if there is both a change in the advisor’s analyst recommendation and in the advisor’s holdings of the acquirer. Returns are sorted by the change in the advisors shareholdings in the quarter of the change in analyst recommendation. Panel A shows results in which acquires are sorted on whether the advisor bought or sold the stock. Panel B requires that a change in the advisor’s stock holdings be in the top or bottom tercile of the sample. T-stats are shown in parentheses.

Panel A: Alphas following increases versus decreases in holdings

Increase Decrease Increase – Decrease

3 months 0.27 (0.81)

-1.09** (-2.39)

1.36** (2.30)

6 months 0.28 (0.28)

0.34 (1.23)

-0.05 (0.92)

12 months 0.25 (1.20)

0.01 (0.05)

0.24 (0.89)

Panel B: Alphas following change in holdings in top versus bottom tercile

Highest tercile Lowest tercile Highest - Lowest

3 months 0.40 (0.92)

-1.14** (-2.27)

1.54** (2.32)

6 months 0.48 (1.06)

0.46 (1.55)

0.02 (0.04)

12 months 0.03 (0.15)

0.14 (0.61)

-0.10 (-0.33)

43

Table 11: Determinants of change in acquirer shares held by the advisor investment bank, net of change in shares held by other institutions This table shows firm fixed effects maximum likelihood regressions, where the dependent variable is percent change in advisor bank ownership in the acquirer net of average percent change in non-advisor bank ownership in the acquirer. Regressions are estimated over the 1,460 advisor-level observations, for five quarters prior to the merger announcement to five quarters following the merger completion (or through the withdrawal date for non-completed mergers). The first independent variable is the change in the advisor analyst recommendation of the acquirer company (measured over the same quarter but observed prior to the measurement of institutional ownership), and the second independent variable is the change in average non-advisor recommendation. In column 2, these recommendation changes are interacted with both a pre-merger dummy and a post-merger dummy (equal in 1 in the quarters prior to and following the merger announcement, respectively, 0 otherwise). All recommendation changes are multiplied by negative 1, such that higher recommendations and increases in recommendations can be interpreted as more optimistic. Dummies for the level of the advisor recommendation at the beginning of the quarter are included (strong buy, buy, and hold). The change in market capitalization represents the change in the market capitalization of the acquirer company over the quarter. Pct held by Advt=-1 equals the percent of outstanding shares held by the acquirer advisor in the acquirer firm one quarter prior. For stock mergers in the first quarter following merger completion, Qtr t-1 Dum*Shares assumed in acquirert=+1equals the number of shares held by the advisor in the target firm in the previous quarter; for all other firm quarters this variable equals 0. Standard errors are clustered by calendar year, and Z-statistics are shown in parentheses.

Dep’t Var = % change in Advisor Shares Held - % change in non-advisor shares held

5 qtrs pre-annt through 5 qtrs

post-completion

Pre-ann’t period

Post-ann’t period

ΔRec * Pre- Merger (Advisor) -0.19

(-0.99)

ΔRec * Post Merger (Advisor) 0.31**

(2.10)

ΔRec (Advisor) -0.21

(-0.82) 0.30**

(2.09) ΔRec (Non-Advisor) -0.02

(-0.08) -0.17

(-0.57) 0.14

(0.64) Strong Buy Dummy (Advisor) 0.21

(0.88) 0.93

(1.88) 0.08

(0.26) Buy Dummy (Advisor) -0.17

(-1.20) -0.13

(-0.52) -0.15

(-0.58) Hold Dummy (Advisor) -0.09

(-0.40) 0.13

(0.37) -0.33

(-1.04) ΔMkt Cap 15.25***

(2.63) 11.71 (0.77)

7.98 (1.78)

Pct held by Advt-1 -131.72***

(-7.44) -143.86

(-9.09) -189.03

(-6.14) Qtr t-1 Dum*Shares assumed in acquirert=+1

0.58***

(2.83) 0.67 (3.09)

N Obs 9,114 3,761 5,244

44