Who Writes the News?Corporate Press Releases During Merger Negotiations

KENNETH R. AHERN and DENIS SOSYURA∗

ABSTRACT

Firms have an incentive to manage media coverage to influence their stock price during important

corporate events. Using comprehensive data on media coverage and merger negotiations, we find

that bidders in stock mergers originate substantially more news stories after the start of merger

negotiations, but before the public announcement. This strategy generates a short-lived run-up

in bidders’ stock prices during the period when the stock exchange ratio is determined, which

substantially impacts the takeover price. Our results demonstrate that the timing and content of

financial media coverage may be biased by firms seeking to manipulate their stock price.

Journal of Finance , forthcoming

∗Kenneth Ahern is at the Marshall School of Business, University of Southern California. Denis Sosyura is at theRoss School of Business, University of Michigan. We thank an anonymous referee, advisor, associate editor, andCam Harvey (editor), Anup Agrawal, Harry DeAngelo, Cesare Fracassi, Todd Gormley, Deborah Gregory, DirkJenter, Simi Kedia, Mark Mitchell, Keith Moore, Marian Moszoro, Dana Muir, Stefan Petry, Matt Rhodes-Kropf,Mark Seasholes, Nejat Seyhun, Geoff Tate, Paul Tetlock, and conference participants at the 2011 Utah WinterFinance Conference, the 2011 AFA Annual Meeting, the 2011 Napa Conference on Financial Markets Research, the2011 Financial Intermediation Research Society (FIRS) Conference, the 2011 Caesarea Center Academic Conference(IDC-Herzliya), the 2011 Finance Down Under Conference, the 2011 UCLA Economics Alumni Conference, the 2011Academy of Behavioral Finance and Economics Meetings, the 2010 ISB Summer Research Conference in Finance,and seminar participants at ESMT-Berlin, NYU, UCLA-Anderson, University of Alabama, University of Miami,University of Michigan, University of North Carolina-Chapel Hill, University of Southern California, and UVA-Darden. For research assistance we thank Taylor Begley, Jason Pang Jao, Nathish Gokuladas, Tiffany Huang, NabilIslam, Jaehyuk Jang, Dino Kaknjo, Adam Schubatis, Gregory Seraydarian, and Andrew Zeilbeck. This research wasconducted while Kenneth Ahern was at the University of Michigan.

1

The relation between information and stock prices is central to finance, dating back to at least

Fama, Fisher, Jensen, and Roll (1969). Key to this relation is the financial media, which serves

as the main channel through which information is disseminated to investors. For instance, recent

research shows that media coverage drives market trading (Barber and Odean, 2008; Engelberg

and Parsons, 2011) and affects prices of large and widely-followed stocks (Tetlock, 2007, 2010).

While this research documents that news coverage impacts the valuation of firms, relatively little

is known about the opposite: how firms actively manage their media coverage. In this paper,

we propose that firms originate and disseminate information to the media to influence their stock

prices during important corporate events, a hypothesis we label active media management. This

hypothesis suggests that the fundamental relation between information and stock prices can be

manipulated by firms seeking to advance their strategic interests through their media coverage.

To identify firms with incentives to manage their media coverage, we focus on large corporate

acquisitions that use the acquirer’s stock as payment. In fixed exchange ratio stock mergers, the

target and acquirer negotiate a fixed number of acquirer shares as payment for target stock. In

contrast, in floating exchange ratio stock mergers, the two firms negotiate a price per target share

in dollars, and at the close of the merger, the acquirer issues as many new shares as are needed to

match the price. The timing of when the exchange ratio is determined produces different incentives

for fixed and floating ratio acquirers. In particular, if an acquirer in a fixed exchange ratio stock

merger can raise its stock price during the merger negotiations, it can offer fewer of its shares

for each target share to achieve the same expected takeover price. In comparison, if a floating

exchange ratio acquirer can raise its price at the close of the merger when the floating exchange

ratio is determined, it can issue fewer shares to achieve the negotiated price.

We exploit this difference in the timing of acquirer incentives to identify whether and how firms

strategically manage their media coverage. We focus our attention on media coverage during the

private negotiation period when a fixed exchange ratio acquirer has an incentive to actively manage

its media to temporarily increase its stock price, but a floating exchange ratio acquirer does not. In

a sample of 507 acquisitions that use at least some stock as payment, we find strong causal evidence

that firms manipulate their stock prices through press releases precisely when they would benefit

the most from a temporary price increase.

2

Though the difference in the timing of acquirer incentives in stock mergers provides an arguably

clean setting to identify active media management, endogeneity is still a concern. In particular,

self selection and omitted variables that simultaneously affect media coverage and stock prices may

confound our tests. We address these issues in a number of ways.

First, we show that the main determinant of the choice between fixed and floating exchange

ratios is the volatility of the acquirer’s stock price. Across many other firm-level characteristics,

including valuation and size, we find that fixed and floating exchange ratio acquirers are economi-

cally and statistically indistinguishable. Thus, fixed and floating ratio deals are natural comparison

groups. In contrast, both fixed and floating exchange ratio acquirers differ significantly from all-cash

acquirers, the next closest comparison group.

Second, we run difference-in-differences regressions with firm-deal fixed effects, comparing changes

in media coverage that occur after the start of merger negotiations for fixed ratio bidders relative

to floating ratio bidders. The firm-deal fixed effects account for time invariant differences in the

acquirer’s stock price volatility, as well as all other observed and unobserved time-invariant dif-

ferences of both the firms (size, industry, opaqueness, historical media coverage, etc.), and of the

merger itself (industry relatedness, payment method, etc.). To account for time-series changes in

volatility, we control for five lags of the absolute value of bidders’ daily returns and turnover, thus

capturing the effect of volatility at different time horizons. Finally, using an option in Factiva, we

exclude all news articles related to stock or market data, thus imposing an additional filter for news

related to stock volatility or bidders’ valuation. In sum, this methodology identifies whether fixed

ratio bidders increase their media coverage, relative to floating ratio bidders, precisely when it is

in their best interest, while controlling for confounding variables.

Our empirical analysis relies on two unique datasets. First, we collect daily media coverage of

acquirers from over 400 media sources in Factiva, including news publications, electronic wires, and

press releases. To our knowledge, with over 600,000 articles, this is one of the most comprehensive

media datasets in financial research. A particular advantage of these data is that they allow us to

distinguish firm-originated press releases from overall media coverage. Though firms can dissemi-

nate information through many channels, corporate press releases have the advantage of being both

fast and widely followed by investors and journalists (Dyck and Zingales, 2003). Additionally, the

3

U.S. regulatory environment gives firms substantial flexibility in the content and timing of press

releases, as discussed in detail in the Internet Appendix. We combine these media data with a

novel hand-collected dataset on the details of merger negotiations reported in SEC filings, which

provides us with the dates when the merger negotiations begin. These key dates — which are

disclosed only after the official merger announcement — enable us to construct precise tests of the

role of financial media during the period when exchange ratios are determined in private.

Our first set of empirical results shows that bidders in fixed exchange ratio stock mergers increase

the number of press releases after they privately begin merger negotiations and before the public

announcement of the merger, during the period when the stock exchange ratio is established. In

contrast, floating exchange ratio acquirers show little difference in press release issuance during

this time. In differences-in-differences regressions, we find that fixed ratio acquirers issue nine

extra press releases during the average negotiation period, a 10% increase compared to baseline

averages.

The increase in media coverage has a significant and positive effect on stock returns, consistent

with prior studies (Barber and Odean, 2008; Da, Engelberg, and Gao, 2011), but also has real

implications for the merger. The abnormal increase in newswire coverage during negotiations is

associated with an estimated savings for an average fixed exchange ratio bidder of $230 million to

$558 million, or between 5% and 12% of the takeover price in an average deal, compared to floating

ratio bidders.

Next, we offer evidence on the tone of news articles issued during merger negotiations. Using

the classification of positive and negative words from Loughran and McDonald (2011), we find that

compared to floating exchange ratio bidders, fixed ratio bidders issue significantly fewer negative

press releases during the period when the exchange ratio is established. In contrast, there is no

change in the tone of newspaper articles, over which firms have much less control. We also evaluate

the changes in the spin of press releases by using the categories of overstated and understated words

from the Harvard IV-4 dictionary. The overstated category includes words that indicate confidence

or exaggeration, such as the adjectives ‘remarkable’, ‘confident’, or ‘strong’. In contrast, the set

of understated words indicates weakness or caution, such as the words ‘doubtful’, ‘uncertain’, or

‘weak’. We find that fixed exchange ratio bidders use significantly fewer understated words in their

4

press releases during the negotiation period, relative to floating ratio bidders. Overall, the evidence

on the changes in the tone of news is consistent with active media management rather than the

random arrival of good news.

Though the changes in tone and the run-up in media coverage and market equity during merger

negotiations are consistent with the hypothesis of active media management, other explanations are

possible. The first possibility is that fixed exchange ratio stock acquisitions coincide with periods

of strong operating performance, and that the media coverage during merger negotiations reflects

positive news about the bidders’ fundamentals. This hypothesis, which we call passive media

management, would explain the increase in news and stock valuation during merger negotiations.

To distinguish this view from active media management, we look at price reversals following the

abnormal increase in media coverage. The passive media management hypothesis predicts that the

increase in acquirer’s market equity prior to the merger announcement will persist if this increase

was driven by fundamentals. In contrast, the active media management view posits that market

equity is expected to correct to a lower level.

Consistent with active media management, we observe a correction in the stock prices of fixed ra-

tio bidders after the merger announcement. Over the three-day period around the merger announce-

ment, an average fixed exchange ratio bidder loses approximately 4.1% in market value, compared

to only 2.5% for floating ratio bidders, a significant difference. This downward trend continues

over the following two months, with fixed ratio bidders losing nearly 40% of the negotiation-period

abnormal stock price run-up. Using floating ratio deals as a benchmark, the reversal is associated

with a loss to an average target in a fixed ratio deal of about 8.5% of the takeover value. The

run-up and reversal in the valuation of fixed ratio bidders around merger negotiations are consis-

tent with strategic media management around these events and are difficult to explain by passive,

coincidental arrival of news.

A related, yet distinct, alternative scenario is that bidders with strong operating performance

systematically engage in bad acquisitions and choose fixed ratio deals because they expect a negative

market reaction. If the strong operating performance generates additional press releases, this

scenario could explain the run-up in media coverage, followed by a correction in the bidder’s stock

price at the announcement of a bad deal. To evaluate this alternative, we construct direct tests of

5

both predictions associated with this hypothesis: (1) bidders’ press releases reflect an improvement

in operating performance, and (2) fixed ratio deals underperform.

To test the first prediction, we use analysts’ forecasts of earnings per share as a proxy for the

changes in a firm’s operating performance. We find no significant change in analyst earnings

forecasts for acquirers from the pre-negotiation period to the negotiation period for both fixed and

floating deals, with no significant difference between the two groups. We also find that fixed ratio

bidders reverse the issuance of press releases after the merger announcement, relative to floating

exchange ratio bidders, contradicting the idea that the run-up in firm-originated news is unrelated

to the merger. To test the prediction that fixed ratio deals underperform, we compare long-run

stock returns of fixed and floating ratio bidders. Using holding periods of one to five years and

controlling for standard risk factors, we find that the performance of these deals is economically

and statistically indistinguishable. Overall, this evidence indicates that the observed pattern in

media coverage and bidders’ valuation is unlikely to be coincidental and cannot be explained by

the differences in deal quality, to the extent that deal quality is reflected in long-run stock returns.

Another alternative hypothesis, opportunistic acquisitions, refers to the endogenous decision to

initiate a takeover and pay in stock when the acquirer’s stock is favorably priced (Shleifer and

Vishny, 2003; Rhodes-Kropf and Viswanathan, 2004). This hypothesis is consistent with a run-up

in bidders’ stock prices before an acquisition and a subsequent correction in firm value. Furthermore,

if the strong stock performance itself generates news, this hypothesis would also explain the increase

in media coverage at the time of merger negotiations.

Several pieces of evidence suggest that our results are unlikely to be explained by opportunistic

acquisitions. To control for news triggered by stock returns, we exclude all stories tagged by Factiva

as recurring pricing and market data. As an additional filter for news related to abnormal stock

performance, we eliminate all articles with fewer than 50 words and all stories with the word

‘stock’ in the title. Our results remain unchanged after imposing these filters. Ultimately, it is

difficult to explain why firms pursuing opportunistic acquisitions without changes in fundamentals

would accidentally increase their press release issuance exactly at the time of negotiations without

a strategy of active media management.

6

Next, we investigate investors’ response to media management and find evidence consistent with

a rational updating of prices. Since the true extent of media management is unobservable outside of

the media-managing firm, there is always a non-zero probability that firm-originated news contains

value relevant information, and, consequently, affects prices. Before investors are aware of the

merger, media management is expected to be more effective, since it is more difficult to distinguish

it from regular variation in the quantity and tone of corporate news. When investors become aware

of the merger, they will revise the probability that news is strategically managed by the firm.

Consistent with this interpretation, we find a larger stock price correction for fixed ratio bidders

that have engaged in more active media management. After the merger is announced, we expect

that future firm-originated news is discounted by rational investors, but continues to influence

stock prices, unless the extent of media management can be perfectly detected. Consistent with

this prediction, we show that the effects of media coverage on stock prices are diminished, but

not completely erased, when investors can rationally anticipate media management, namely for

acquirers with a history of media management and for floating ratio bidders at the close of the

merger.

We also study how targets respond to media management. We find that target firms engage in

a similar strategy of active media management during merger negotiations and that the relative

intensity of media management between bidders and targets affects merger outcomes. In particular,

we estimate that a one standard deviation increase in an acquirer’s newswire coverage leads to an

11% increase in the acquirer’s gain relative to the target’s gain. This evidence is consistent with

the view that as long as the extent of media management is observed with noise, it can have a

significant effect on equity prices and corporate outcomes.

We complement our analysis with several robustness tests. First, we construct a set of tests unaf-

fected by selection issues by exploiting the within-group variation in media management incentives

for fixed exchange ratio bidders only, rather than across-group variation between fixed and float-

ing ratio bidders. We find that fixed ratio bidders for which additional news coverage is likely to

generate the strongest effect on returns (high market-to-book, greater analyst forecast dispersion,

and more retail investors, (Kumar, 2009)), show a larger increase in press releases than fixed ratio

bidders with weaker incentives. Similarly, we show that media management is attenuated in deals

7

where it is likely to be less effective, such as deals with protective collars and deals that use cash

as part of the payment. Second, we impose additional filters to control for possible information

leakage about the upcoming merger by excluding deals where a merger-related word appears in

the title of an article about a bidder in the largest newspapers during the negotiation period. Our

main findings are unchanged in these robustness tests.

The central contribution of this paper is to provide some of the first evidence that the fundamental

relationship between information and stock prices can be distorted by a firm’s strategic incentives to

control its news coverage. In contrast to the assumption of exogenous and random media coverage,

our results show that firms use the media to manipulate their stock prices in order to realize large

and long-lasting real effects. While our focus on mergers is motivated by a clean identification

setting, we believe that our findings extend to other corporate events where a short-run increase

in stock prices can have significant real effects, such as initial public offerings, secondary equity

issuances, and the expiration of executive stock options, among others.

Our findings also provide new explanations for two important empirical results in merger research:

the run-up in bidders’ stock prices and the subsequent correction at the merger announcement. In

contrast to assumptions of exogenous misvaluation (e.g., Shleifer and Vishny, 2003), our paper

provides evidence that the run-up is driven by firms’ strategic timing of news releases. Second, in

contrast to assumptions of managers’ irrationality (Roll, 1986) or governance failure (Jensen, 1986)

that have been proposed to justify negative announcement returns, our results are consistent with

the view that merger returns reflect a rational discounting of selectively-released news. Overall, we

provide a unified explanation for both patterns using a rational framework.

The rest of the article is organized as follows. Section I provides a brief overview of how this

paper relates to existing research. Section II discusses data and methods. Section III presents

empirical results on whether firms actively manage media. Section IV provides empirical evidence

on how investors and target firms respond to media management. Section V offers robustness

checks and reviews alternative hypotheses. Section VI concludes.

8

I. Related Literature

This paper contributes to the growing literature on media in finance and economics. Earlier re-

search shows that media outlets display political and economic biases (Groseclose and Milyo, 2005),

and that these biases tend to be slanted towards the customers of the media outlet (Mullainathan

and Shleifer, 2005; Gentzkow and Shapiro, 2006, 2010), the media firm’s advertisers (Reuter and

Zitzewitz, 2006; Gurun and Butler, 2012), or governments (Besley and Prat, 2006). More re-

cent research shows that media coverage has a substantial effect on both political and economic

outcomes, such as voting behavior (DellaVigna and Kaplan, 2007), stock trading (Engelberg and

Parsons, 2011), CEO compensation (Core, Guay, and Larcker, 2008), and corporate governance

(Dyck, Volchkova, and Zingales, 2008). These papers view the media as the active player, influenc-

ing outcomes through potentially biased editorial choices. In contrast, our paper views a firm as

the active player, using media coverage to influence its own outcome.

Another line of research on media in financial markets investigates the effect of media coverage

on stock returns and volume. One view is that news dissemination reduces information asymmetry,

resulting in quicker incorporation of new information, more efficient prices of financial assets, and

higher values (DellaVigna and Pollet, 2009; Fang and Peress, 2009). An alternative view is that

media coverage is subject to manipulation and can result in deviations of stock prices from their

fundamentals (Huberman and Regev, 2001). Consistent with this alternative view, a number

of papers find that media coverage produces short-term upward price pressure on stocks due to

increased attention by investors (Chan, 2003; Vega, 2006; Barber and Odean, 2008; Gaa, 2010).

Our paper provides evidence that the latter effect can dominate during important corporate events

and that companies can strategically use this channel to affect their stock price.

Two other papers have looked at the role of media in corporate finance, both in the setting of

IPOs. Cook, Kieschnick, and Ness (2006) investigate the relationship between marketing efforts,

including media coverage, and the success of an IPO. They find that news coverage significantly

affects IPO outcomes. Similarly, Liu, Sherman, and Zhang (2009) examine the role of media in

IPOs and show that offerings with greater media coverage have higher initial returns and greater

long-term value. We contribute to this literature by distinguishing between firm-originated news

9

and external media coverage and by demonstrating how firms actively manage their information

environment.

Last, our paper is related to the literature on investor relations. Bushee and Miller (2007)

find that companies that hire investor relations firms experience an increase in media coverage,

institutional ownership, and valuation. Solomon (2012) shows that investor relations firms spin

corporate news in a favorable way, resulting in a temporary increase in stock returns. Our paper

adds to this literature by highlighting the motives and channels of information management and

demonstrating the effects of this strategy on the outcomes of major corporate events.

II. Data and Methods

A. Background of the Merger Data

To construct our sample of mergers, we start with the largest 1,000 completed mergers of U.S.

publicly traded firms in the SDC database, as measured by deal value, announced between January

1, 2000 and December 31, 2008. Acquirers must purchase at least 20% of outstanding shares and

own more than 51% of the target firm shares following the merger. We begin our sample in 2000,

since Factiva’s news coverage in earlier years is scarcer and lacks intelligent indexing codes for many

merging firms (discussed in more detail below). We exclude withdrawn merger bids because we

require a merger agreement document to identify key dates in the negotiation process. Our focus on

larger deals is motivated by the substantial transfers of value in these transactions, which arguably

provide stronger incentives for managing stock valuation during the merger. We also obtain the

form of payment used in the merger from SDC and exclude deals paid only in cash because all-cash

offers are different from deals involving stock on many dimensions. We discuss these differences in

detail later in the paper when we explain our identification strategy. Therefore, our sample includes

only mergers in which some stock was used as payment.

For each deal in our sample, we retrieve information about the terms of the transaction and the

key dates in the merger process from Securities and Exchange Commission filings.1 This information

1We search through the following forms in order until we find the relevant data: DEFM14A, DEFA14A, DEFR14A,DEF14A, PREM14A, PRER14A, S-4/A, S-4, 424B3, 424B2, F-4, 497, 10-K, 8-K, N-148C/A, SC13E3, SC13E3/A,and SCTO-T/A.

10

typically appears in the section entitled “Background of the Merger” in the merger agreement,

which provides a narrative history of the merger process, though terms of the transaction are often

described in various other sections of the SEC filings. In particular, we collect the date when the

merging firms first discuss the potential merger, the date when the exchange ratio is first discussed

and the date when it is finalized, the type of the exchange ratio (fixed or floating) and the period

over which the price is determined for the floating ratio. These key dates (which are made public

only after the official merger announcement) enable us to construct sensitive tests of the active

media management hypothesis that rely on daily data from news sources. Hand-collecting these

data is also necessary to record whether a fixed or a floating exchange ratio is used since these data

are not reported in standard databases, such as SDC. The “Background of the Merger” section for

the merger of CVS and Caremark is provided as an example in the Internet Appendix.

Since the details on the identification of fixed or floating exchange ratios are critical for our

tests, we eliminate deals for which we cannot reliably establish this information. After collecting

the data, we found that the date that merger negotiations began was much more populated and

precise than the dates that pertain to the determination of the exchange ratio. In addition, for

the dates that we could obtain about the exchange ratio, we found that though the use of a fixed

exchange ratio was often decided relatively early, the exact date that the ratio was set was often

close to the public announcement date (within 10 days or less). Due to both the lack of dates and

the closeness to the announcement date, this information is a less reliable way to delineate time

periods by when we would expect to see more active media management. Therefore, we focus our

attention on the date that merger discussions begin.

Using these data on the timing of merger negotiations, we define the following time periods:

1. Pre-Negotiation Period: the 120 trading days that immediately precede the day that merger

discussions begin.

2. Negotiation Period: the period that begins on the date of the first discussion of the merger by

the two firms, until 16 days before the public announcement of the merger.

3. Announcement Period: the five days surrounding the first public announcement of the merger.

4. Transaction Period: the period that starts five days after the public announcement of the merger

until two days before the merger closes.

11

We restrict the negotiation period to end well before the public announcement to ensure that our

media coverage does not contain information about the merger. However, all of our results remain if

we change this requirement to include days up to ten or five days before the public announcement.

In all cases, it is important to remember that these dates are before there has been any public

acknowledgement of the merger and are only realized ex post by reading the SEC filings. Figure 1

illustrates the sequence of time periods in a typical merger.

B. Press Releases and Financial Media Data

News coverage is collected from the Factiva database. To collect articles and press releases for

each firm, we use the acquirer’s Intelligent Indexing Code assigned by Factiva. In particular, if a

news article discusses a firm in sufficient detail, Factiva matches this article to the firm’s intelligent

indexing code, enabling us to identify relevant articles and press releases based on firm identity,

rather than a less precise keyword match.

To study a firm’s media strategy around mergers, we collect daily data on the firm’s press releases

and news articles issued from the pre-negotiation period through the close of the merger. Our list

of possible news sources includes all English-language media sources included in Factiva’s category

of major news and business publications plus newswire services. Major newspapers and business

publications include a large number of publications, such as USA Today, The Wall Street Journal,

and The New York Times, among many others. Our counts of media articles include reprints or

highly similar articles. This means our media coverage variables measure breadth of coverage across

multiple media outlets, rather than unique news events.

The data on newswire articles are particularly crucial for understanding whether firms actively

manage media. These articles typically report firm press releases with no additional analysis.

As discussed in the introduction, a firm’s press releases may be particularly suitable for active

management because this channel is less regulated than accounting statements, affording greater

flexibility in content and providing an opportunity to spin news in a desired way (Dyck and Zingales,

2003). Using newswire articles to measure firm-originated news has an advantage over simple counts

of press releases as well, because newswires measure how widely circulated is the press release.

12

Finally, we eliminate deals for which the Factiva database provides an intelligent indexing code

only for the combined firm rather than a separate code for the acquirer and the target. As another

filter for article substance, we eliminate articles with fewer than 50 words and articles tagged by

Factiva as recurring pricing and market data. A representative sample of news articles is presented

in the Internet Appendix for the CVS-Caremark merger.

C. Summary Statistics

After imposing the media and merger data filters, we end up with 507 mergers, including 377

(74%) fixed exchange ratio deals and 130 (26%) floating exchange ratio deals. These percentages

are comparable to the sample used to study merger arbitrage in Mitchell, Pulvino, and Stafford

(2004). In a sample of 2,130 mergers from 1994 to 2000, they find that 78% of stock mergers

used a fixed exchange ratio and 22% used floating. Our final sample of articles includes 617,445

articles over the entire merger process across 421 sources, including local, national, international,

and foreign newspapers and newswires. The domestic newspaper with the most articles in the

sample is The Wall Street Journal, with 16,471 articles, or 2.7% of the total number of articles in

the sample. The next two newspapers with the most articles are The New York Times, and The

Washington Post, with 6,450 articles (1.0% of total) and 3,631 articles (0.6% of total).

The most common newswire and most common media outlet of any kind, by a large margin,

is Reuters News, with 139,789 articles, or 22.6% of the total sample of articles. Dow Jones News

Service and Business Wire account for the next two most frequent newswire sources with 98,373

articles (15.9%) and 38,540 articles (6.2%). Measured by the number of articles, newswires are

the predominant source of new information. The top 15 newswires account for more than 70% of

total media articles. Our media sources also include foreign newspaper articles that are written

in English. Not surprisingly, these media sources are in countries where English is the primary

or a common language, such as the U.K., India, Australia, and Canada. The most frequently

represented foreign newspaper is Financial Times of the U.K. with 8,045 articles, or 1.3% of total

articles. A detailed listing of the top 15 sources by type of media outlet is provided in Internet

Appendix Table I.

13

Throughout our analysis we use three different measures of media coverage. First, we record

all articles from any source in our sample. Second, since newswires and foreign newspapers likely

have a different audience than domestic newspapers, we separately measure the number of articles

in the top three most circulated domestic newspapers: The Wall Street Journal, USA Today, and

The New York Times. These three papers have a total combined circulation in 2009 of 4,852,236

newspapers per day, not adjusting for overlapping readers. Though USA Today is the second most

circulated newspaper, because it is a general interest national newspaper, it is only the 12th most

common domestic newspaper source for articles in our sample, with only 1,262 articles. However,

given its wide circulation, it is an important media outlet for firms. Last, we record the number

of articles in the top three newswires: Reuters News, Dow Jones News Service, and Business

Wire. As noted, we use this measure to capture firm-originated news, since the newswires typically

provide little analysis. One may be concerned that bidders are able to influence newspaper coverage

through public relations (PR) firms. However, recent research shows that the influence of PR firms

is primarily concentrated in the smaller, regional newspapers (Bushee and Miller, 2007; Solomon,

2012). In contrast, PR firms appear to have relatively little control over the largest national

newspapers, such as The Wall Street Journal or The New York Times.

Table I presents summary statistics of the merger and media data. An average (median) ac-

quirer appears in 73.4 (11) media articles per month (20 trading days) during the pre-negotiation

period. These figures represent normal media coverage unrelated to upcoming mergers. An av-

erage acquirer’s press releases are covered in 30.8 articles in the top three newswires, and only

3.0 articles in the top three domestic newspapers, during an average pre-negotiation month. These

cross-sectional distributions are right skewed with median newspaper articles much lower than mean

articles. Finally, there is substantial cross-sectional variation in the distributions of news articles

with the standard deviation exceeding the mean across the measures of media coverage.

Next, we show the within-deal change in media coverage from the pre-negotiation period to the

negotiation period. The skewness is substantially reduced for the within-deal count distributions.

The average bidder has an increase of 2.9 articles from all sources, and 1.5 articles in the top three

newswires, with medians close to zero for all three media source categories. Restricting attention to

the five days surrounding the merger announcement, the average (median) acquirer has 59.7 (30)

14

articles across all media sources in our sample, 29.9 (15) articles in the top three newswires, and 3.2

(2) articles in the top three domestic newspapers. This demonstrates that merger announcements

are large news days, as expected, with the announcement period media coverage equal to about

80% of an average acquirer’s monthly coverage in the pre-negotiation period. It also demonstrates

that firms issue multiple press releases on the same day. On the announcement days, the median

acquirer has 15 articles across just three newswires. It also demonstrates that coverage by one of

the top three newspapers is relatively rare. In the announcement period, the median acquirer only

has two articles in the top three newspapers and the average acquirer is covered in just 3.2 articles.

At the bottom of Table I, we present summary statistics of the timing of the merger process.

These dates are crucial for our study since we will compare the media coverage during each of the

three periods. From the start of the merger talks until the day of the public announcement is 64.7

days on average and 44 days at the median. The average (median) time from the announcement

to the completion of the merger is 64.5 (48) days. Since there are cases where the negotiation

period is very short, we restrict our analysis to only mergers where there are at least 20 days in

the negotiation period. The timeline illustrated in Figure 1 reports the average number of days in

each time period.

In untabulated statistics, the average (median) market equity of an acquirer in our sample is $26.8

billion ($6.17 billion), compared to $3.64 billion ($1.12 billion) for targets. The average relative

size of the deal, defined as the transaction size divided by the market equity of the acquirer, is

51.4% and the median is 27.2%. The average transaction value is $4.44 billion. These statistics

show that our sample is composed of large mergers, both in absolute and relative size. On average,

72.7% of the payment is made with stock (87% at the median), the average acquirer has 2.2 different

mergers in our sample, and 72.8% of acquirers and targets are in the same Fama French 49 industry

classification code. The top five industries of acquirers are Banking, Computer Software, Electronic

Equipment, Telecommunication, and Pharmaceuticals, comprising 50.9% of acquirers. The top five

industries of targets are the same top three as acquirers, plus Petroleum and Financials, with the

top five accounting for 52.5% of targets.

15

D. Identification Strategy

The central premise of this paper is that firms actively manage their media coverage during key

corporate events, namely, acquisitions. To identify the causal relationship between making a stock

acquisition and increased media coverage, we must address selection bias and endogeneity. First, the

act of acquiring is not randomly assigned to firms. Instead, there may be something characteristic

about acquirers that leads them to increase their media coverage, but that is unrelated to the

merger. For example, firms that are experiencing high growth may release more news and also

make more mergers at the same time. In this case, an omitted firm-level characteristic, such as a

high growth rate, may cause a spurious correlation between the timing of a merger and increased

media coverage.

More generally, it is likely that both observable and unobservable firm-level characteristics are

related to media coverage and mergers. To address this, we use a firm-deal fixed effects differences-

in-differences approach to control for any time invariant firm characteristics, as follows:

Mediai,t = α+ ηNegotiationi,t + φNegotiationi,t × Fixedi + γXi,t + δi + εi,t, (1)

where Mediai,t is the count of media articles for firm i on day t for one of the three categories

of media sources we defined previously, Negotiationi,t is a dummy variable that takes the value

of one if the observation occurs after merger negotiations have begun and zero if it occurs before

negotiations have begun, Fixedi is a dummy variable that equals one if the deal is structured using

a fixed exchange ratio and zero if it is structured using a floating exchange ratios, Xi,t is a set of

time-varying factors, and δi is a firm-deal fixed effect. The firm-deal fixed effect captures any time

invariant characteristic of the acquirer or the merger, including such observables as industry, firm

size, historical media coverage, financial ratios, as well as any unobservable firm characteristics that

do not change in the relatively short time of the merger process. In addition, this term captures

any time invariant aspect of the merger, such as the industry relatedness and the form of payment,

including whether the deal uses a fixed or floating exchange ratio. The coefficient on the interaction

of Negotiation and Fixed captures the differences-in-differences of fixed versus floating from the

period prior to negotiations to the negotiation period. Thus, this specification alleviates much of

the concern that omitted firm or deal characteristics may explain a firm’s level of media coverage,

16

and isolates the effect of a fixed or floating exchange ratio on differences in media coverage as

merger negotiations proceed.

We restrict our attention only to mergers that use stock to control for any differences between

all-cash and stock mergers. If all else were equal, acquirers in all-cash mergers would be a useful

benchmark since they do not have an incentive to use the media to affect the takeover price. How-

ever, prior research finds that all-cash deals and stock deals are systematically different: bidders’

returns vary significantly by cash and stock deals (Andrade, Mitchell, and Stafford, 2001), the

target’s form of legal organization has differential effects on announcement returns for cash and

stock deals (Chang, 1998; Fuller, Netter, and Stegemoller, 2002; Moeller, Schlingemann, and Stulz,

2004), and information asymmetry affects the choice of cash and stock as well as the announcement

returns of the merger (Hansen, 1987; Moeller, Schlingemann, and Stulz, 2007; Officer, Poulsen, and

Stegemoller, 2009; Cooney, Jr., Moeller, and Stegemoller, 2009). In addition, Loughran and Vijh

(1997) finds differences in long-run returns for all-cash and stock bidders, Faccio and Masulis (2005)

finds that financing constraints and corporate control threats affect the choice of cash or stock, and

Boone and Mulherin (2007) report that auctions are more likely paid in cash compared to negotiated

mergers.

In Table II, we empirically verify that cash and stock deals are substantially different in our

sample. All-cash bidders are significantly larger, have less dispersion in analyst estimates, greater

institutional ownership, greater intangibles and R&D, make acquisitions of relatively small targets,

and are less likely to make horizontal mergers, compared to stock deals, whether fixed or floating.

In contrast, fixed and floating stock deals are highly similar across a variety of measures, with

the exception of the acquirer’s stock price volatility. Additionally, in Internet Appendix Table II,

we run multinomial logit tests of the choice between fixed, floating, and all-cash deals and find

significant differences between the determinants of cash and fixed deals, but almost no difference

between fixed and floating deals.

The observed similarity between fixed and floating exchange ratio mergers makes them useful

comparison groups. In contrast, the many differences between cash and stock deals increase the

risk that they have different time trends in media coverage during merger negotiations, unrelated

to active media management. These unobserved trends would not be captured by firm-deal fixed

17

effects. For instance, previous work finds that all-cash acquisitions are often indicative of a free cash

flow problem and agency issues in the bidding firm (Harford, 1999). If cash deals are associated

with poorer governance, bidders’ management may prefer to “lead a quiet life” by keeping the

firm’s disclosure strategy constant, merely to avoid exerting additional effort. In this case, we

would observe no change in bidders’ media coverage during negotiations, but this outcome would

be driven by the differences in governance rather than by the lack of an incentive to manage media

in an all-cash bid. Additionally, the evidence in Table II shows that the size of the target relative

to the acquirer is nearly three times smaller in cash deals than in stock deals. The firm-deal fixed

effects would account for the differences in the relative size of the target, but they would not account

for the time-varying differences in the incentives for media management. For example, the acquirer

may choose not to alter the firm’s disclosure strategy for the smallest deals. In this case, we would

observe no change in bidder’s media coverage, but this outcome would be driven by the relative

size of the target rather than by the choice of the method of payment.

However, even within stock mergers, one may be concerned that the choice of fixed or floating is

not randomly assigned across acquirers. For this to confound our tests, it must be that there is an

omitted variable that both causes a firm to choose a fixed or floating exchange ratio and also leads

to greater media coverage for reasons unrelated to the merger itself. One potential variable is firm

over-valuation. If the management of the acquirer knows that its firm is overvalued at the start of

negotiations, it would have a preference for a fixed exchange ratio to lock in a favorable takeover

price. In addition, if media outlets have a preference for articles about highly valued firms, we may

see more media coverage during the merger negotiations caused by high valuations, rather than

active media management. Alternatively, fixed ratio bidders may be systematically undervalued

and their managers may release more information in order to help the market correct to true values.

In Table II, we find that the valuations of the two sets of stock acquirers, based on market-to-

book ratios or Tobin’s Q, are statistically and economically indistinguishable from each other. Since

the two types of firms have nearly identical valuations before negotiations begin, any subsequent

differences in media coverage during the negotiation period are unlikely to be driven by a spurious

correlation between the media’s preference for articles about highly valued firms and a highly valued

firm’s preference for a fixed versus floating exchange ratio.

18

We find that the one significant difference between fixed and floating mergers is the volatility

of the acquirer’s stock price. Volatility generates opposite preferences for fixed exchange ratios

between acquirers and bidders. If the payment method were established unilaterally by the target,

all else equal, a risk-averse target would likely prefer a floating exchange ratio to lock in its dollar

compensation. However, the bidder would have the opposite incentive and would prefer a fixed

exchange ratio. Specifically, bidders have strong incentives to fix the number of shares issued to

ensure that the merger is accretive to EPS and to manage post-merger voting control rights. Ulti-

mately, since the payment method is established bilaterally, the outcome is an empirical question.

In both the univariate results and the multinomial logit tests in Internet Appendix Table II, we find

that fixed ratio mergers are more likely when the acquirer’s stock price volatility is higher. 2 These

results imply that the choice of payment method is more sensitive to bidders’ preferences than to

targets’. More generally, this is consistent with the fact that the majority of mergers used fixed

exchange ratios, but a sizable minority use floating ratio bids, reflecting a negotiated outcome.3

Given that the choice of fixed or floating exchange ratios is driven primarily by the acquirer’s

historical stock price volatility, we do not expect our results to be affected by selection bias. First,

Brandt, Brav, Graham, and Kumar (2010) find that firm-level volatility is persistent over time.

Therefore, the cross-sectional differences in firm-level volatility between fixed and floating ratio

bidders are mostly captured by the firm-deal fixed effects we include in our tests. This is especially

true, considering the short time windows we study. Second, to account for any time-series variation

in volatility that could affect media coverage, we include controls for the magnitude of daily fluc-

tuations in the bidders’ stock returns in all of our regressions. These controls include five lags of

the absolute value of bidders’ daily returns, thus capturing the effect of volatility at different time

horizons and allowing for a decaying relation. Third, to control for the possibility that a fixed ratio

is chosen in response to a change in the bidder’s volatility, we also include the interaction term of

the bidder’s volatility (measured by the absolute value of daily returns) and the fixed ratio dummy.

We also remove any news articles that may be driven by stock volatility or bidders’ valuation. In

2Additional robustness checks for this result are presented in Internet Appendix Table III.3In untabulated results, we also find that more than 30% of repeat acquirers in our sample employ both fixed andfloating exchange ratio payments, consistent with the idea that payment method is not unilaterally chosen.

19

particular, using an option provided by Factiva, we exclude all media articles related to stock or

market data.

Finally, as another way to identify active media management from alternative hypotheses, we

exploit the variation in firm-generated news and news appearing in public media. If the increase in

news is explained by incentives to manage media coverage, we would expect a significantly greater

increase in firm-generated news (over which the company has full control) and a smaller increase

in news in other media.

D.1. The Incentives of Floating Exchange Ratio Bidders

Our identification relies on the assumption that bidders will want to increase media coverage

prior to the date when the exchange ratio is determined. For fixed exchange ratio bidders this is in

the negotiation period, while for floating exchange ratio bidders this is in the transaction period.

Though floating ratio bidders are not restricted from starting a media campaign before the merger

announcement, there are several potential costs to starting a media campaign early. First, a floating

ratio bidder may wish to avoid alerting the target that it may influence the takeover price through

media coverage. If media management is postponed until after the end of negotiations, it may be

more difficult for the target to estimate the extent of future media management and to fully adjust

for it when negotiating the takeover price. Second, the literature on financial disclosure shows that

a firm has an incentive to reduce disclosure when its product market rivals can make strategic use

of this information, such as in periods preceding important investments.4

On the other hand, a floating ratio bidder may start a media campaign early if it believes that it

will be harder to influence its stock price after the merger announcement or if there is no benefit to

delaying the news. If this is the case, we would expect to find smaller differences in media coverage

between fixed and floating ratio bidders. In other words, as the timing of incentives becomes more

similar across the two types of firms, the difference in media coverage between fixed and floating

ratio bidders will become understated.

4For example, this prediction is derived analytically in Darrough (1993) and Verrecchia (1983, 1990), and supportedempirically in Campbell (1979) and Ellis, Fee, and Thomas (2012), among others.

20

D.2. The Incentives of Target Firms

Target firms also have incentives to manage their media during the merger process to increase

the takeover price. However, in contrast to acquirers, targets will only benefit if they can increase

their stock price during the negotiation period, not the transaction period, regardless of whether

a fixed or floating exchange ratio is used. Once either the fixed exchange ratio or price per share

of the floating exchange ratio is settled in the negotiations, the target’s share price has no effect

on the final price that the target will receive, because the price is either already determined (in

floating deals) or fluctuating based on the acquirer’s stock price (in fixed deals), not the target’s.

Therefore, only the acquirer’s stock price could affect the price received. This means that we do

not have a clear benchmark of media coverage for targets. For this reason, we focus on acquirers

since we can run differences-in-differences tests that exploit the differences in timing of incentives

of fixed versus floating ratio bidders.

III. Evidence on Active Media Management

In this section, we first investigate the pattern of media coverage during negotiations for fixed

relative to floating exchange ratio acquirers. Second, we test whether increases in media coverage

affect market equity. Third, we test for reversals in media coverage and market equity. Finally, we

analyze changes in the tone of the media coverage over the merger timeline.

A. Do Fixed Ratio Bidders Increase Media Coverage During Negotiations?

The active media management hypothesis predicts that acquirers in fixed exchange ratio mergers

will have an incentive to try to increase their stock price through increased media coverage during

the merger negotiations. In contrast, acquirers in floating exchange ratio mergers would not have

the same incentive. Instead, they have an incentive to increase their stock price near the close of the

merger when the exchange ratio is determined. Thus we expect that acquirers in fixed ratio deals

will exhibit significantly higher media coverage during the negotiation period than will floating

ratio acquirers. For reference, we present these predictions in comparison to alternative hypotheses

in Table III.

21

Figure 2 presents a comparison of the cumulative number of daily abnormal newswire articles

for fixed and floating ratio bidders for the pre-negotiation period and the transaction period. We

compute daily abnormal articles as the ratio of the daily number of articles to the average daily

number of articles for the firm in the pre-negotiation time period. We then take the cumulative

sum over time. For comparison in the time trends, we normalize fixed and floating ratio bidders’

abnormal counts to be zero at 100 days prior to the public announcement in subfigure (a) and to

100 days prior to the close of the merger in subfigure (b). The figure reveals a clear distinction.

The firms involved in fixed exchange ratio mergers have a marked increase in the number of media

articles during the private negotiations, whereas the firms in floating exchange ratio mergers have

virtually no change in media coverage until right before the announcement. In subfigure (b), the

pattern is reversed in the transaction period, with a sharper increase in the number of media articles

for floating exchange ratio acquirers relative to fixed ratio acquirers.

Figure 3 presents analogous figures for market equity values during the negotiation and transac-

tion periods. Just as media coverage increases more rapidly for fixed exchange ratio bidders during

the negotiation period, so does market equity. Similarly, following the public announcement, there

is a strong reversal in market equity, with fixed exchange ratio bidders experiencing a marked

decline relative to floating ratio bidders.

These figures highlight two important phenomena. First, the timing of increased media coverage

is concentrated in a relatively short time span for fixed and floating exchange ratio bidders. Second,

the timing of increased media coverage and market equity corresponds directly with the time when

a firm has the most to gain from a temporary increase in its stock price: during the negotiations for

fixed acquirers and near the close of the merger for floating ratio acquirers. As mentioned previously,

we focus our analysis on media coverage of fixed exchange ratio acquirers during negotiations

because it provides a cleaner test than investigating media coverage of floating exchange ratio

acquirers at the close. Nevertheless, the patterns of media coverage are consistent for both types

of acquirers.

Though the differences in media attention by merger payment method revealed in Figure 2

are indicative, they are not statistical tests. In Table IV, we present univariate differences-in-

differences tests of media coverage in the pre-negotiation, negotiation, and transaction periods and

22

for fixed exchange versus floating exchange mergers. In Panel A, we find that fixed ratio acquirers

experience a significant increase in media coverage after the start of merger talks. In contrast,

floating exchange ratio acquirers receive significantly less media coverage after the start of merger

talks. The difference in the differences between fixed and floating is significant as well, consistent

with the pattern revealed in Figure 2. On average, fixed ratio acquirers have an increase of 2.3

more articles per day compared to floating exchange ratio acquirers.

In Panels B and C of Table IV, we repeat the analysis using the number of media articles in the top

three domestic newspapers and in the top three newswires. The results are qualitatively identical.

For domestic newspaper coverage, the difference between the pre-negotiation and negotiation period

media coverage in fixed exchange ratio mergers is a significant 0.07 articles higher per day than the

same difference for acquirers in floating exchange mergers. For newswire articles, the differences-

in-differences is 1.1 articles, also highly statistically significant.

The difference in media coverage by payment method is economically meaningful. If we multiply

the average daily difference in media coverage by twenty trading days to produce a monthly figure,

we have an additional 46.5 articles from all sources, 1.5 additional articles from the top three

domestic newspapers, and 26.2 additional articles from newswires, on average. Comparing these

figures to media coverage in the pre-negotiation period as presented in Table I, we find that the

additional number of articles in the negotiation period for fixed exchange acquirers compared to

floating exchange acquirers represents an increase of 50% for domestic news coverage and 85% of

newswire coverage, compared to baseline averages.

Table IV also shows significant differences in media coverage between the transaction period and

the pre-negotiation and negotiation periods. Consistent with Figure 2, fixed ratio bidders have

a decrease in media coverage during the transaction period, compared to the negotiation period,

while floating ratio bidders have an increase in media coverage. The differences-in-differences is

significant for each of the three measures of media coverage. While these results are consistent with

our main hypothesis, as stated before, we put more weight on evidence from the pre-negotiation

and negotiation periods because it is not contaminated by media coverage directly related to the

merger itself. In particular, the significantly greater media coverage in the transaction period,

relative to the pre-negotiation period for both fixed and floating acquirers, likely reflects media

23

coverage directly related to the merger. In contrast, the media coverage in the negotiation period

is before the public announcement of the merger and not driven by coverage of the merger itself.

Table V presents the coefficient estimates from multivariate fixed effects differences-in-differences

tests as in Equation 1. In column (1), the dependent variable is the number of media articles from all

sources, in column (2) it is daily media articles in the top three newspapers, and in column (3) it is

the number of articles in the top three newswires. Observations are over the pre-negotiation period

and the negotiation period, again up to 16 days before the public merger announcement. We control

for time-varying volatility, firm deal fixed, and account for heteroskedasticity and autocorrelation

in media coverage.

The differences-in-differences interaction terms are positive and significant when the dependent

variable is either all media or newswires. This indicates that fixed exchange ratio acquirers have

a greater increase in media coverage than floating exchange ratio acquirers during the merger

negotiation period, after controlling for all firm and deal-level characteristics.

For newswires, we find a negative coefficient for the negotiation period dummy, which indicates

that floating ratio bidders are decreasing their newswire coverage compared to the pre-negotiation

period. This is consistent with the argument that floating ratio bidders have incentive to withhold

news releases during the negotiation period.

The economic significance of the differences-in-differences estimates is substantial. Converting

the daily marginal effects into monthly effects by multiplying by 20 trading days, we find that

fixed exchange ratio acquirers realize an increase of more than 10% of the mean of newswire press

releases during the negotiation period, compared to floating exchange ratio acquirers. There is no

effect for newspapers. This is consistent with the active media management hypothesis, as firms

can choose to issue more press releases which will affect newswire coverage, but are likely to have

a harder time influencing newspapers.

Next, we expect that acquirers in all-stock deals have greater incentives to manage their media

than acquirers that use a smaller fraction of stock as payment. In tests reported in Internet

Appendix Table IV, we interact a dummy for all-stock deals with the fixed ratio and negotiation

period dummies. We find a positive and significant effect of all-stock deals on media coverage.

This implies that media coverage increases more as more stock is used as payment and provides

24

additional evidence that media coverage is driven by the incentives of acquirers to affect their stock

price precisely when it is in their best interest.

A.1. Alternative Methods to Account for Skewness in Count Data

In these previous tests, we use raw media counts as our dependent variable, rather than log-

transformed media counts. We do this for a number of reasons. First, by using firm-deal fixed

effects, much of the skewness in media counts is eliminated since we are effectively using within-

deal deviations from the acquirer’s average. Second, prior econometrics research argues that log-

transformations are convenient, but tend to produce biased estimates when applied to count data.

Abrevaya (1999) shows analytically that coefficient estimates in standard within-transformations of

panel data can be biased and inconsistent when the the transformation of the dependent variable is

unknown. Empirically, O’Hara and Kotze (2010) run Monte Carlo simulations of regressions based

on log-transformed count data and find that the log-transformed estimates are substantially biased.

To properly address count data in fixed effects models, Cameron and Trivedi (1998) recommend

using a generalized linear model assuming a Poisson distribution of the dependent variable. Start-

ing with Hausman, Hall, and Griliches (1984), Poisson regressions have been used in economics

and finance to study discrete count data in a wide range of settings, including patents (Lerner,

2006; Lerner, Sorensen, and Stromberg, 2011), airplane purchases (Pulvino, 1998), merger counts

(Huizinga and Voget, 2009), and, like our paper, the number of media articles (Core, Guay, and Lar-

cker, 2008). In the special case of count data that contains a large number of zeros (e.g., firm-days

with no media articles that are prevalent in our study), the literature recommends zero-inflated

Poisson and Negative Binomial regressions (Greene, 1994). Therefore, we re-estimate our main

regressions using each of these specifications suggested for count data: (1) fixed effects Poisson;

(2) zero-inflated Poisson; and (3) zero-inflated Negative Binomial regressions. For completeness,

we also run robustness tests using an OLS model with logged media counts and removing fixed

effects. Finally, we also verify the robustness of our results in a winsorized sample (at 1st and

99th percentiles) based on firms’ time-series of article counts. In all of these specifications, we find

consistent results that support our main findings. The results are presented in Internet Appendix

Table V.

25

In summary, acquirers in fixed exchange mergers have significantly greater media coverage during

the merger negotiation period than floating exchange acquirers, controlling for fixed effects, share

turnover, and returns. This is true for media coverage from all media sources and coverage in the

top three newswires. In addition, we find that the increase in media coverage is larger for newswire

articles than for newspaper articles.

The abnormal increase in newswire articles (which are mainly firm-originated news) following

the beginning of merger negotiations is inconsistent with the opportunistic acquisitions hypothesis

where acquirers respond to overvalued stock prices by using fixed exchange ratio stock payments.

Instead, these results imply that firms are actively generating media coverage by issuing additional

press releases. Next, we test the relation between media coverage and stock prices.

B. Does Increased Media Coverage Affect Firm Value?

A positive relation between media coverage and market equity is implicitly assumed in the

argument that fixed exchange ratio acquirers actively manage media coverage to improve the terms

of trade. We address this implied relationship in this section of the paper.

Table VI presents fixed effects differences-in-differences regressions of the log of market equity

on current and lagged daily media coverage, a negotiation period dummy, a payment method

dummy, and interactions between these variables. First, market equity is significantly higher in the

negotiation period compared to the period before the merger talks began. This is consistent with

the run-up in an acquirer’s stock price before a merger (Rhodes-Kropf, Robinson, and Viswanathan,

2005). Second, consistent with prior research, media coverage has a positive and significant effect

on stock prices independent of the timing of mergers or payment methods used, with newspaper

coverage having the greatest effect (Huberman and Regev, 2001).

The variables most important for the active management hypothesis are the interactions of media

and payment method and the triple interaction of media coverage, payment method, and the timing

dummy. We find that over the entire time span, only newspaper coverage has a differential effect on

market equity for fixed ratio acquirers compared to floating ratio acquirers, but that the marginal

effect for fixed ratio acquirers is greater in the negotiation period than in the pre-negotiation period

26

for all three types of media coverage. This is consistent with the incentive to attempt to influence

stock prices through media during a very specific time period.

The impact of strategic media coverage is economically significant. The increase in the acquirer’s

price will lead to fewer shares issued. If we use the acquirer price in the pre-negotiation period as

a benchmark, then multiplying this price times the number of shares that do not need to be issued

yields an estimate of the dollar savings from increased media coverage. If V is the deal size and Pt

is the acquirer stock price in period t, then the savings in the takeover price from an increase in

the bidder’s stock price is V (1 − P1

P2). In unreported tests, we find an average increase in market

equity of $3.53 billion from the pre-negotiation to the negotiation period for fixed ratio bidders.

Using the average bidder size of $24.58 billion and average deal size of $4.44 billion, this translates

into a savings of $558 million, or roughly 12.5%.

Some of this increase in market equity is not driven by media coverage, so we use the marginal

effects to calculate a more refined estimate of the cost savings. From Table IV, an average bidder has

an increase in newswire articles of 1.131 per day from the pre-negotiation to the negotiation period,

compared to floating ratio bidders. From Table VI, the marginal effect of an additional newswire

article for fixed ratio bidders in the negotiation period is 0.004 of ln(market equity). For the 65

days in the negotiation period, this implies a change in the ln(market equity) of fixed ratio bidders

compared to floating ratio bidders that is attributable to media coverage of 1.131 × 0.004 × 65 =

0.294, or $1.34 billion, on average. Using the average fixed ratio bidder size in the pre-negotiation

period and the average deal size, this generates 4.44(1 − 24.5825.92

) =$230 million in savings from

additional newswire coverage. This is roughly 5% of the takeover value.

C. Do Prices and Media Coverage Reverse Post-Announcement?

Though the above results imply that a firm can temporarily increase market equity values through

active media management, it is unlikely to be an effective long-run strategy. Eventually the market

will adjust its expectations about the value of new information for a firm, or a firm will run out

of relevant information and a reversal in the stock price will occur. In contrast, under the passive

media management hypothesis, if a firm simply times fixed exchange takeovers to coincide with

the release of relevant news that will boost its stock price, no reversal should be observed. To test

27

these hypotheses, in this section we analyze whether fixed exchange ratio acquirers have a different

pattern of stock price and media coverage reversal following the merger announcement than do

floating ratio acquirers.

In Table VII, we test for evidence of a reversal in market equity and media coverage. We use

the same fixed effects differences-in-differences approach again, though now the dependent variable

is logged market equity in the first two columns and newswire articles in the last two columns.

We use the level of market equity that immediately precedes the announcement as our baseline for

the stock price reversal test. For the reversal in media coverage, we use the negotiation period to

measure a baseline for the occurrence of media coverage.

First, we find that market equity falls in the announcement period for all acquirers. However, we

find a significant and negative differences-in-differences between fixed and floating exchange ratio

acquirers, with the fixed exchange ratio acquirers having a bigger decline in stock prices than floating

ratio acquirers. In untabulated results, we find that the cumulative abnormal stock return in the

three-day merger announcement window is −2.5% for floating exchange ratio bidders and −4.1%

for fixed exchange ratio bidders, a statistically significant and economically meaningful difference.

These results are consistent with the negative announcement returns for stock acquisitions of public

targets (Andrade, Mitchell, and Stafford, 2001; Fuller, Netter, and Stegemoller, 2002), but show that

fixed ratio acquirers are driving the large negative returns, even though both types of acquirers are

issuing new stock. Similarly, in column two, we find a reversal in market equity in the transaction

period. Market equity is significantly lower for fixed ratio acquirers than for floating ratio acquirers

in the transaction period, compared to the period immediately prior to the announcement.

Next, in columns three and four, we find that there are significantly more newswire articles in

both the announcement period and the transaction period, compared to the negotiation period, as

expected. We find no differences-in-differences for the announcement period, but we find a signifi-

cant decline in newswire articles for fixed versus floating ratio acquirers, relative to the negotiation

period. This provides validation of the reversal in media coverage in Figure 2 after controlling for

firm-deal fixed effects.

The economic magnitude of the reversals is substantial. In univariate tests, an average fixed

ratio acquirer experiences a larger negotiation period run-up in market value during the negotiation

28

period than does the average floating exchange ratio acquirer ($3.5 billion versus $1.3 billion in an

average deal) and also realizes a significantly larger decline in market value during the transaction

period (roughly a half a billion dollars). Aggregating the marginal impact on acquirer returns for

fixed ratio bidders in the announcement period and an average 64-day transaction period leads to

an overall decline of 8.5% in the acquirer’s stock price. For an average transaction value of $4.44

billion and an average acquirer size five days before the announcement of $26.7 billion, the decline

in acquirer value leads to a takeover price that is $413.7 million less, or roughly 9.3% of the average

transaction value.

These results indicate that following the announcement, fixed exchange ratio acquirers experience

a strong reversal of the increase in market value and newswire coverage experienced during the

negotiation stage. These results are inconsistent with the passive media management hypothesis

that fixed ratio stock acquirers are simply timing the merger to coincide with relevant news to boost

their stock prices and improve their terms of trade. Instead, the stock price reversal we document

in this section is consistent with the hypothesis that acquirers temporarily boost their stock price

through active media management.

D. The Information Content of Press Releases

In this section, we offer content analysis of the press releases issued by fixed ratio acquirers

during merger negotiations. Our analysis focuses on two dimensions: (1) tone of the text, and (2)

importance of news. To measure article tone, we use textual analysis based on the classification of

words as positive, negative, overstated, or understated. To evaluate the importance of news in the

press releases, we study the correlation between press releases and newspaper articles, relying on

the revealed preferences of journalists.

A number of recent studies have shown that the tone of media articles has a significant effect on

stock prices, even after controlling for the information disclosed in the article (Tetlock, 2007, 2010;

Demers and Vega, 2010; Solomon, 2012). In other words, the tone with which a particular piece

of news is reported has a significant effect on the market reaction to this news. Therefore, if fixed

ratio acquirers attempt to influence their stock prices during merger negotiations, they may do so

29

not only by increasing the volume of news, but also by presenting their news in a more positive

way or by withholding disclosure of negative news.

We take two approaches to measure the tone of news. In the first approach, we follow the

classification of positive and negative words in financial texts developed in Loughran and McDonald

(2011). One advantage of this classification is that it is developed for textual analysis in an economic

setting, thus matching the type of disclosures analyzed in our tests. The lists of positive and

negative words contain 353 and 2,337 words, respectively, and are downloaded from the web page

of Bill McDonald. To illustrate, some examples of positive words in our sample include ‘achieve’,

‘attractive’, ‘beneficial’, ‘excellent’, ‘favorable’, ‘improve’, ‘outstanding’, ‘regain’, ‘strengthen’, and

‘surpass’. Examples of negative words include ‘adverse’, ‘breach’, ‘detrimental’, ‘erode’, ‘penalties’,

‘terminate’, ‘threaten’, ‘unexpected’, and ‘unsuccessful’.

To develop a simple and replicable measure of article tone, we compute the fraction of positive

and negative words in the article and classify an article as positive (negative) if it has an above-

average fraction of positive (negative) words. To control for the variation in tone across news

sources, we compute the ratio of positive and negative words separately for each media outlet.

We estimate fixed effects differences-in-differences regressions where the dependent variable is

the number of positive or negative news stories about the acquirer. These tests are summarized

in Table VIII. The evidence in Panel A indicates that compared to floating ratio acquirers, fixed

ratio acquirers experience a significant decline in negative news stories in the negotiation period,

as shown by the negative and significant coefficient on the interaction term in column 1. Columns

3 and 5 indicate that the reduction in negative news stories is attributable to firm-originated news,

not newspaper articles. The economic magnitude of the change in article tone is nontrivial. A

fixed ratio acquirer with median characteristics experiences approximately a 10% reduction in the

number of days with negative newswire articles during merger negotiations than during the pre-

negotiation period. This evidence is consistent with acquirers’ strategy of withholding negative

news during merger negotiations or reporting this news in a less negative way, supporting the

active media management hypothesis.

The change in the tone of media articles during the negotiation period is driven primarily by the

reduction in negative news stories rather than an increase in positive news releases. The evidence in

30

columns 2, 4, and 6 of Table VIII shows no effect on positive news stories. One possible explanation

for this result is asymmetry in investors’ responses to positive and negative news documented in

the literature. For example, Tetlock (2007) shows that negative news coverage has a significantly

stronger absolute effect on stock prices. Under this interpretation, managing negative news is

particularly important for managing stock prices during merger negotiations.

To provide a comparison for the evidence from the negotiation period, we also repeat the tests

of the changes in media tone for the transaction period and present our evidence in Panel B of

Table VIII. As before, we use the pre-negotiation period as our benchmark for measuring the

regular tone of news. The evidence shows that newswire articles become more reserved in their

tone after the merger is announced, as indicated by a significant reduction in both positive and

negative newswire articles for fixed and floating ratio bidders. This evidence is consistent with the

idea that the content of the bidder’s press releases becomes more formal and the language becomes

less extreme when the merger news becomes public. We do not find any significant differences

in the tone of firm-originated news between fixed and floating ratio bidders, as indicated by the

insignificant coefficient on the interaction term of the transaction period and the fixed ratio dummy.

One possible explanation is that the news is dominated by required disclosures about the upcoming

merger, which contain more boilerplate language.

Our second approach to measuring the tone of the news relies on a set of word lists from the

Harvard-IV-4 and Lasswell dictionaries provided by General Inquirer. In particular, we use the

Overstated and Understated word categories, which capture a firm’s attempt to provide a positive

subjective interpretation of the news, or ‘spin.’ According to General Inquirer, these categories

include words, “indicating overstatement and understatement, often reflecting presence or lack of

emotional expressiveness.” To provide evidence on both the upward and downward spin, we offer

the analysis of each word category separately and then use the difference between the two categories

as a net measure of exaggeration.

The Overstated category includes 696 words, which indicate an “emphasis in realms of speed,

frequency, causality, inclusiveness, quantity or quasi-quantity, accuracy, validity, scope, size, clar-

ity, exceptionality, intensity, likelihood, certainty and extremity.” Examples of words from the

Overstated list include ‘colossal’, ‘fantastic’, ‘huge’, ‘reliable’, and ‘undoubted’. The Understated

31

category includes 319 words, which indicate “de-emphasis and caution.” These words can be viewed

as downward spin, often indicating weakness, doubt, or uncertainty. Examples of words on this list

include ‘ambiguous’, ‘minor’, ‘slight’, ‘unsatisfactory’, and ‘weak’.

Because these categories were not developed specifically for financial texts, we design our tests

with strong benchmarks to produce more statistical power and minimize noise from context-specific

word interpretations. In addition, we design our tests to capture firms’ spin above and over the

‘objective’ interpretation of news, as proxied by the top three national newspapers, on which firms

have less influence. The advantage of this methodology is that it controls for the time-varying

changes in the overall market sentiment and for the changes in firms’ fundamental news.

For overstated words, we calculate a variable at the daily level that takes the value of 1 if

the difference between the fraction of overstated words in the top three newswire articles and

the fraction of overstated words in the top three domestic newspapers is greater than the 75th

percentile, within the firm’s time-series of this difference from the pre-negotiation to negotiation

periods. The variable takes the value of −1 if the difference is less than the 25th percentile, and

takes the value of 0 otherwise. Using an analogous classification, we compute similar variables

for understated words and for the difference of the fraction of overstated minus the fraction of

understated words.

In Internet Appendix Table VI, we report the results from the tests of spin. Using the same

differences-in-differences firm-deal fixed effects tests as in the rest of the paper, we find that fixed

exchange ratio bidders use significantly fewer understated words in their press releases during the

negotiation period, relative to domestic newspapers and floating ratio bidders. We find similar

evidence that the net overstated-understated fraction is positive and significant for fixed ratio

bidders, relative to floating ratio bidders, during the negotiation period. These results are consistent

with active media management and indicate that firms tend to spin their news by increasing the

appearance of confidence in their press releases via reducing the indication of doubt, caution, or

uncertainty.

To complement the textual analysis of news, we study the extent to which firm-originated news

during merger negotiations is reported by newspaper journalists. Though newspapers derive much

of their news coverage from press releases, they must select the news that would be of greatest

32

importance to their readers. We posit that acquirers that are attempting to influence their stock

price through media management may be forced to issue press releases that contain less important

information for investors in the negotiation period compared to the pre-negotiation period. Since

newspapers selectively choose the most important information, the correlation of newswire to news-

paper articles will decrease if the additional press releases are less informative. Instead, if firms

are timing acquisitions to take advantage of positive new information, the content of the newswire

articles during the negotiation period should be at least as important, if not more important than

the newswire articles published during the pre-negotiation period. Therefore, under the alternative

hypothesis, the change in the correlation between the number of newspaper and newswire articles

should be non-negative.

We test this hypothesis by regressing the number of articles in newspapers on the number of

current and lagged newswire articles, controlling for firm-deal fixed effects, using three different

measures of newspaper articles. In unreported tests, we find that there is a positive and significant

correlation between newspaper articles and newswire articles, as would be expected, and that this

correlation is not significantly different for fixed and floating ratio bidders in the pre-negotiation

period. However, during the merger negotiation period, the correlation between firm-originated

news and newspaper articles drops by 9.1%age points, or approximately by two thirds for fixed-ratio

acquirers. This implies that the fixed exchange ratio acquirers are producing more press releases

during the merger negotiations than before merger talks begin, but newspapers are providing less

coverage of these press releases than normal, presumably because they contain less important

information.

Overall, the results in this section suggest that acquirers in fixed exchange ratio mergers actively

manage media coverage not only by increasing the number of press releases, but also by managing

the tone of reported news. In contrast, newspapers are less likely to report this news and show

no change in the tone of their articles. Taken together, this evidence indicates that the changes in

the amount and tone of firm-originated news reflect active media management by acquirers, rather

than the coincidence of merger negotiations with periods of good news.

33

IV. The Response to Media Management

The results in the prior sections provide strong evidence that firms actively manage their media

coverage during merger negotiations, which is associated with a significant run-up and reversal in

stock prices. The incentives for acquirers to manage their media are clear. It is less clear, however,

why investors respond to active media management. Second, how do target firms respond to active

media management? This section investigates these questions.

A. Investors’ Response to Media

We consider two possible explanations of investors’ response to media coverage during mergers.

On the one hand, rational investors may update prices in response to new information. On the other

hand, the run-up and reversal in acquirer stock prices could be explained by investors’ overreaction

to news.

First, the pattern of price run-up and reversal we document is consistent with a rational updating

of prices by investors. Since the true extent of media management is unlikely to be perfectly

observable to any party outside of the media-managing firm, there is always a non-zero probability

that firm-originated news contains value relevant information, and, consequently, affects prices.

Before investors are aware of the merger, media management is more effective, since it is more

difficult to distinguish active media management from regular variation in the quantity and tone

of fundamental news about the firm. When investors become aware of the upcoming merger at the

announcement, they are likely to revise upward the probability that news is strategically managed

by the firm. Therefore, the stock price of the bidder corrects at the merger announcement, reflecting

the reinterpreted view of the previously-issued news. In addition, future news issued by the firm

before the completion of the merger will be significantly discounted, but will continue to influence

stock prices, unless investors can perfectly detect the extent of media management in each disclosure

by the firm.

We investigate this explanation empirically. First, we examine the market reaction to news

issuance by repeat acquirers that have past histories of mergers and media coverage that are con-

sistent with active media management. When these firms make a subsequent acquisition, we expect

that, all else equal, rational investors will respond less to their news releases. At the same time,

34

these firms are likely to have greater capabilities for media management, since they have more

experience and since acquisitions play a greater role in their corporate strategy. Therefore, we

anticipate a higher increase in press release issuance by these bidders during merger negotiations.

To test the predictions about repeat acquirers, we create a dummy variable which identifies

deals where 1) the acquirer had a prior merger in our sample where it used a fixed exchange

ratio, 2) its media coverage from all sources and market equity increased from the pre-negotiation

to negotiation period during the prior merger, and 3) it had reversals in its stock price during

the announcement and transaction periods. We then interact this dummy variable with the fixed

exchange ratio dummy and the negotiation period dummy. In Internet Appendix Table VII, we find

that newswire coverage increases at a significantly greater rate for these acquirers in the negotiation

period, but the main effect of their media coverage on market equity is reduced, independent of

whether a fixed or floating ratio is used in the subsequent mergers. This evidence is consistent

with the view that investors discount news issued by repeat acquirers that are likely to be actively

managing their media. We also find evidence that investors rationally respond to media coverage of

floating ratio acquirers. Since the period when the number of shares to be issued by floating ratio

acquirers occurs after the announcement of the merger, we expect that the news issued during this

period will be discounted by the market. As shown previously, floating ratio bidders significantly

increase the issuance of press releases before the close of the merger (the period when the floating

exchange ratio is set). However, untabulated results show that this news generates a significantly

weaker market reaction, compared to news issued by fixed ratio bidders before the deal is publicly

announced.

While the evidence so far suggests that investors’ response to media management does not rely on

investor irrationality, the active media management hypothesis does not rule it out. For instance,

the evidence in previous research shows that investors exhibit behavioral biases that may result

in overreaction to news (Odean, 1999; Barber and Odean, 2000). In particular, previous studies

show that an increase in a stock’s media coverage may temporarily increase investor attention to

this stock, resulting in a price run-up followed by a subsequent correction. Much of the existing

research on media and finance is consistent with this idea (Huberman and Regev, 2001; Tetlock,

2007; Barber and Odean, 2008). This research on financial media shows that investors’ overreaction

35

to news is more pronounced for glamour stocks (e.g., high-market to book; high-tech industry) and

stocks with a larger retail ownership. If this is true, acquirers with such characteristics should be

more likely to attempt to use media coverage to affect their stock prices. In Section V.A., we review

cross-sectional patterns in media coverage and find evidence that supports these predictions.

Overall, it is likely that the market’s reaction to news management in our setting reflects a

combination of a rational response to value-relevant news, as well as a transient component driven

by a possible overreaction to firms’ spin.

B. Media and Merger Gains

Although our evidence so far indicates that firms manage their media coverage during merger

negotiations, it is unclear whether this strategy affects merger outcomes. On the one hand, a target

may be able to accurately infer the extent of media management and fully adjust its demands in

negotiations. In this case, media management will have no effect on merger outcomes. On the

other hand, if a target cannot perfectly observe the true extent of media management in acquirer’s

disclosures, there is always a positive probability that an acquirer’s news release reflects accurate,

value-relevant information. Under this assumption, media management can affect the terms of the

merger, even when the negotiating parties are informed. This section seeks to distinguish between

these alternatives.

To study the effect of media management on merger outcomes, we examine the distribution

of merger gains and losses between the target and the acquirer within the same merger, using

a measure of relative gains proposed in Ahern (2012). This measure is calculated as the dollar

abnormal announcement return of the target minus the dollar abnormal announcement return of

the acquirer, divided by the pre-merger sum of the acquirer and target market values. The dollar

abnormal announcement return is the percentage abnormal stock return multiplied by market

value. This measure captures the difference in dollar values realized at the announcement, under

the assumption that the announcement return is the best predictor of the merger gains. Using

dollar values accounts for the large difference in the sizes of acquirers and targets in a typical

merger. In addition, using the difference in dollar returns allows for negative returns. We expect

36

that cross-sectionally, those acquirers that have negotiated better terms will receive more gains

relative to the target, compared to acquirers with worse terms.5

As mentioned, we do not focus our attention on targets in this paper because targets’ incentives

do not change over time as do the acquirers’. However, we would like to measure the target’s media

management strategy, since it likely affects the target’s gain relative to the acquirer. Therefore,

for all targets in the 507 mergers in our sample that have Factiva intelligent indexing codes, we

collect the same media data as we collected for acquirers. Factiva codes are primarily assigned for

larger firms, which limits our sample size to 101 targets after accounting for all data filters. We also

collect data on other variables known to affect the relative merger gains, including market to book

(Rhodes-Kropf and Robinson, 2008), relative value, size, and industry relatedness (Ahern, 2012).

As reported earlier, targets in our sample tend to be smaller than their acquirers. Likewise,

targets have substantially less media coverage than acquirers. In the pre-negotiation period, the

median number of articles from all media sources about a target is roughly 26% of the median

number of articles about an acquirer. The median number of target newswire articles is about 46%

of the number of acquirer newswire articles. The difference in articles from all media sources and

newswires indicates that the larger acquirers normally issue moderately more press releases, but

that newspapers and other media sources are much more likely to cover the larger firms compared to

the smaller targets. Targets and acquirers are more similar along other dimensions. In particular,

the average (median) market-to-book of an acquirer is 3.46 (2.23), compared to 3.06 (2.29) for

targets.

We first provide evidence that targets increase their media coverage during negotiations. Like

fixed exchange ratio bidders, all targets have the incentive to increase their stock price during

the negotiation period only. Therefore, as we did for fixed ratio bidders, we compare targets’

media coverage to floating exchange ratio bidders in the negotiation period. In Internet Appendix

Table VIII, we report the results from differences-in-differences fixed effects regressions during

merger negotiations that are identical to our main specifications. We find that targets have a

significant increase in media coverage during the negotiation period, compared to floating exchange

5Because large stock price reversals for fixed exchange ratio bidders are consistent with active media management,negative announcement returns could indicate that the acquirer negotiated a favorable exchange ratio, a counter-intuitive result. However, if acquirers have successfully negotiated a favorable exchange ratio, we would also expectlow target announcement returns as the market revises the target value based on the expected offer price.

37

ratio bidders. This is consistent with the idea that targets manage their media coverage during the

period when it is most beneficial.

Next, we turn to our tests of the role of media coverage on the division of merger gains between

the target and the acquirer. Our variables of interest are the number of media articles of the

acquirer and the target in the negotiation period. Since our variable of interest, the relative gains

of the acquirer and target, is time-invariant for each merger, we cannot control for firm-deal effects

in this setting. Therefore, we normalize the media counts using the firms’ pre-negotiation averages.

Table IX presents coefficient estimates from OLS cross-sectional regressions, where the dependent

variable is the target’s dollar gain relative to the acquirer’s gain.

We find that targets that have greater media coverage during the negotiation period have larger

relative dollar gains compared to targets with less media coverage, whether measuring media cover-

age with newswire or newspaper articles. In addition, greater newswire coverage, but not newspaper

coverage, of the acquirer significantly reduces the share of gains captured by the target. This means

that media coverage during the negotiation period has an economically meaningful effect on the

bargaining outcome of targets and acquirers. A one standard deviation increase in the acquirer’s

newswire coverage leads to an additional acquirer gain of 0.07 dollars per every dollar in pre-merger

combined market equity. This compares to an average normalized acquirer gain relative to target

gain of −0.04. This means that an acquirer’s dollar gain relative to the target’s dollar gain is 11%

higher than average when it has a one standard deviation increase in newswire coverage.

Overall, the evidence in this section suggests that media management can have a significant

effect on merger outcomes. These results support our earlier evidence that media management can

be an effective strategy even when the agents are informed.

V. Robustness and Alternative Hypotheses

In this section, we provide a number of robustness checks and extensions of our main results.

A. The Determinants of Abnormal Media Coverage

In this subsection, we investigate the effect of media on merger negotiations using only the

within-group variation among fixed exchange ratio bidders. By only looking at fixed exchange

38

ratio bidders, this provides a robustness check on any selection bias in our prior results that may

be caused by the choice of fixed versus floating exchange ratios. Though all fixed exchange ratio

acquirers have an incentive to influence their stock price during the negotiation period, firms that

are hard to value are more likely to use this method than others. In addition, firms with more

retail investors, as opposed to institutional investors, may also be more likely to attempt to use

media coverage to affect their stock prices since retail investors may be more prone to respond to

media coverage.

To test these hypotheses, we regress the amount of abnormal media coverage in the negotiation

period on proxies for the difficulty of firm valuation and the presence of retail investors. First, we

use a dummy variable for above median market-to-book as a proxy for growth firms. The larger

intangible option value inherent in growth firms may make it easier for them to spin news to the

media. Second, we use a dummy variable for high dispersion in analysts’ earnings forecasts. We

calculate the coefficient of variation for the most recent analyst earnings forecasts before the an-

nouncement date of the merger using data from I/B/E/S. We then create a dummy variable equal

to one for firms with above-median coefficients of variation in analysts’ expectations. Third, we

use a dummy variable for high-tech industries to proxy for the difficulty of valuation. For firms

whose primary industry is in the Fama French 49 Industry Classifications of Computer Hardware

(35), Computer Software (36), or Electronic Equipment (37), this dummy variable equals one, for

all others it is zero. Our final proxy for the difficulty of valuation is R&D/Assets. To measure

ownership composition, we record the percentage of institutional investors in the firm in the most

recent reporting period before the merger announcement, using 13f filings from Thomson Finan-

cial. Since these are cross-sectional regressions, we cannot use firm-deal fixed effects. Instead, we

normalize media coverage by the average daily media coverage in the pre-negotiation period.

The regression results in Table X support the hypothesis that firms that are harder to value have

greater media coverage during the negotiation period. Firms with above-median market-to-book

ratios, firms with above-median dispersion in analysts’ forecasts, and firms in high-tech industries

have significantly greater abnormal newswire coverage. High R&D expenditures are negatively

related to newswire coverage. Institutional ownership is significantly negative, consistent with

the idea that retail investors are more likely to respond to information in the media than are

39

institutional investors. The effects for newspaper coverage are muted, though qualitatively similar

to the effects for newswires. This confirms our prior evidence that newspapers are less susceptible

to active media management strategies.

We also interact the explanatory variables with the negotiation period dummy. We find that

the same effects we listed above are heightened during the negotiation period, compared to the

pre-negotiation period. For example, firms in high-tech industries have 0.21 more newswire articles

per day than do non-high-tech industry firms in the pre-negotiation period, but 1.3 more articles

per day in the negotiation time period.

Overall, the results in this section are consistent with the notion that firms in fixed exchange

ratio mergers wish to influence stock prices before acquisitions, but those firms where media is

especially important for market prices, namely hard to value firms, experience the greatest effect.

Since more information should make the firm easier to value, reverse causation (more media creates

greater analyst forecast dispersion, for example) is unlikely to explain these results. Thus, these

results are additional evidence that firms actively manage media before a merger, controlling for

possible sample selection bias.

B. Insider Trading During Negotiations

If the management of an acquirer is actively trying to increase its stock price using the media,

managers who know their stock is overvalued may try to exploit their inside information by selling

shares. In this case, we could use insider selling to help identify overvalued shares. On the other

hand, our main hypothesis is that the management of acquiring firms strategically uses the media to

make acquisitions at reduced prices. Insider sales would send a negative signal to the market, which

would be counter-productive to the efforts to push up the stock price. Since insider trades are made

public almost immediately in the post-Sarbanes-Oxley period that covers most of our sample, the

market is likely to react to this information immediately. Second, insider sales may send a negative

signal to the target and threaten the deal’s completion. If the bidder’s management doesn’t want

to hold its own shares, why would target shareholders? If a deal does not materialize, the bidders’

insiders will likely lose significant bonuses that they receive for deal completion. Previous research

40

shows that such bonuses for completed acquisitions can be in the millions of dollars (Grinstein and

Hribar, 2004).

In Internet Appendix Table IX, we empirically test whether there is more insider trading during

the run-up period for acquirers. We collect data from the TFN Insider database for trades by

officers of acquirers. Following convention, we only include open market purchases or sales, delete

observations marked as inaccurate or incomplete (‘cleanse’ field of S or A), and only include ob-

servations that record all of the following information: the number of shares traded, the date, and

the price per share in the transaction. Using these data, we run firm-deal fixed effect differences-

in-differences regressions, following our main specification, on the number of net shares traded

(purchases minus sales). This specification isolates the change in the difference between fixed and

floating acquirers insider trading activity, from before negotiations to the negotiation period.

We find no significant effects in the bidders’ trading activity. We also run a second specification

where we only include non-zero trading days, since the large number of days with no insider activity

may bias the results. We again find no significant effects. These results suggest that though there

is a documented run-up in the stock price during the negotiations, insiders do not trade on this

information, likely because of a combination of strategic considerations and monetary incentives

associated with deal completion.

C. Merger Rumors as a News Driver

One of the key insights of this paper is that we identify abnormal media coverage before the

public announcement of the merger using ex-post data to construct the time period when merger

negotiations begin. This approach assumes that the media articles that occur before the public

merger announcement do not contain rumors about the merger. Though the acquirer does not have

any incentive to reveal its merger plans before the announcement, since it may attract additional

bidders, and we are conservative by restricting attention to the period at least 16 days before the

merger announcement, it may be possible that that the media coverage we identify is somehow

related to the upcoming merger.

To address this concern, we re-estimate our tests using a conservative filter to exclude deals where

the news about an upcoming merger may have leaked into the press, or there is speculation about a

41

merger. In particular, we exclude any deal where at least one of the following merger-related words

appears in the title of any article about a bidder in the top three domestic newspapers during

the negotiation period: ‘merge’, ‘merger’, ‘merges’, ‘bid’, ‘bids’, ‘acquire’, ‘acquires’, ‘acquirer’,

‘acquisition’, and ‘takeover’. This is a relatively strict filter, since some of the deals are excluded

even if an article focuses on a merger by a different acquirer and a different target, but discusses

the firm of interest for other reasons. This filter reduces the sample from 507 to 444 deals.

In Panel A of Internet Appendix Table X, we report the results from tests on the effect of

payment method on media coverage using this subsample. Our results are largely consistent with

our main results. In particular, we find a positive and significant differences-in-differences effect of

fixed ratio bidders in the negotiation period on media coverage, as in our main tests. While we

do not claim that there is no information leakage before the merger announcement, these results

provide strong evidence that our results are not driven by merger speculation in the media.

D. Collars

In our main tests, we do not separately identify mergers with price collars. A price collar creates a

hybrid between fixed and floating exchange ratios (Officer, 2006). For example, a collar on a floating

exchange ratio might be designed to change the exchange ratio to a fixed ratio if the acquirer’s

stock price is outside of a predetermined price range during the period when the exchange ratio is

determined. For the purposes of identification, collars will attenuate the differences in incentives

of acquirers in fixed and floating ratio deals. Thus, our main results likely understate the true

intensity of media management in an unconstrained setting. We address collars explicitly in two

robustness tests.

First, in Panel B of Internet Appendix Table X, we verify that our main results hold after

excluding deals that include collars, as reported in SDC. Compared to the full sample, when we

drop deals with collars from the analysis, the magnitude and statistical significance of the coefficient

estimate for newswire coverage increases. This result confirms the intuition that the use of collars

attenuates differences in the incentives of fixed and floating bidders and that removing these hybrid

deals strengthens the evidence on media management.

42

Second, in columns 2 and 3 of Internet Appendix Table IV, we exploit the terms of the collars to

identify variation in the incentives of firms. To measure the protectiveness of a collar, we measure

the width of the collar as the percentage point difference from the upside to the downside, relative

to the base price. Following our main specifications, in a sample of collared deals, we interact a

dummy variable for above median protectiveness of the collar with the negotiation period dummy

and the negotiation period × fixed ratio dummy. We find that in the deals with more protective

collars, the number of media articles for fixed ratio bidders is significantly smaller than that for fixed

ratio bidders with less protective collars. This is consistent with the idea that a more protective

collar reduces a fixed ratio bidder’s incentive to manipulate its stock price during the negotiations.

To provide more general evidence, we also estimate a model with three dummy variables: no collar,

weak collar, and strong collar, where strong collar is the most protective. We find that strong

collars have the greatest impact on media management. In contrast, the effect of a weak collar is

statistically indistinguishable from the effect of having no collar, likely because weak collars do not

bind a bidder’s media management strategy. In summary, after accounting for price collars’ impact

on bidders’ incentives, our main results are strengthened.

E. Analysts’ Forecasts

We have argued that press releases are a particularly attractive way for a firm to influence its

stock price. However, we acknowledge that firms may also use other channels to try to influence

their stock prices. Though we document patterns in media coverage consistent with active media

management, one may be concerned that the effects of media coverage on stock prices are actually

driven by an alternative channel. One alternative channel of influence is for bidders to influence

analysts’ recommendations.

To address this alternative, we study the changes in analysts’ earnings forecasts and recommen-

dations during merger negotiations, using the same identification strategy as in our tests of media

coverage. We collect analysts’ estimates and recommendations from I/B/E/S at the daily-level

based on the announcement date of the forecast or recommendation. In differences-in-differences

firm-deal fixed effects regressions, we find no significant change in analyst earnings forecasts or

stock recommendations for the acquirers from the pre-negotiation period to the negotiation period.

43

This conclusion holds for both fixed and floating acquirers (with no significant difference between

the two groups), and is robust to using various time windows for analysts’ forecasts: one quarter,

one year, and two years. These results show that the run-up in the stock price of fixed-ratio bidders

is not driven by the changes in analysts’ recommendations or forecasts. The results are presented

in Internet Appendix Table XI.6

Alternative channels are possible, but to the extent that firms use other strategies, these variables

must operate at a daily level, be robust to firm-deal fixed effects, and vary systematically between

fixed and floating ratio bidders precisely at the time of merger negotiations. Our evidence shows

that one of the most likely alternative channels, analysts’ recommendations, is not driving the effect

of media coverage on stock prices.

F. Good Firms That Make Bad Acquisitions

An alternative interpretation of our results is that the increase in media coverage during merger

negotiations reflects an improvement in operating performance, but the larger stock price correction

for fixed exchange ratio bidders at the announcement indicates that a bidder chose a fixed exchange

ratio because its management knew the merger would be viewed unfavorably by the market. To test

the first prediction, we use the changes in analysts’ forecasts and recommendations, detailed above,

as a proxy for the change in a firm’s operating performance. This proxy has several advantages.

First, analysts’ forecasts are forward looking, reflecting changes in firms’ operating performance

even if these changes do not have an immediate impact on reported financials. Second, analysts’

forecasts are updated frequently and allow for various time horizons of projections. As reported

above, we find no significant change in analyst earnings forecasts or stock recommendations for

acquirers from the pre-negotiation period to the negotiation period, nor any significant difference

between fixed and floating exchange ratio acquirers. In contrast, during the same time period,

our main results show that media coverage increases for fixed exchange ratio bidders, relative to

floating ratio bidders. These results imply that the increase in firm-originated news is unlikely to

reflect improvements in operating performance.

6The evidence on analyst forecasts is consistent with prior literature in finance and accounting, which shows thatfirms’ influence on stock analysts has been dramatically reduced or nearly eliminated by Regulation FD (Cohen,Frazzini, and Malloy, 2010), which covers the vast majority of our sample.

44

We also provide evidence that the increase in firm-originated news is unlikely to reflect improve-

ments in operating performance. If the increase in firm-originated news is unrelated to the merger,

we should expect to see 1) an equal increase in media coverage for fixed and floating ratio bidders

during negotiations, and 2) that the news continues after the announcement of the merger. In

contrast, we find that fixed and floating ratio bidders have significant differences in media cover-

age during the negotiations. Second, in Table VII, we show that the issuance of press releases by

fixed ratio bidders reverses in the transaction period, relative to floating exchange ratio bidders.

Given the short time periods we study, the quick reversal in the issuance of news releases appears

inconsistent with a sustainable improvement in operating performance.

The second testable prediction posits that the market reacts negatively to the merger announce-

ment because fixed ratio bidders are making bad acquisitions. To measure ex-post deal quality, in

Internet Appendix Table XII, we compare long-run stock returns from fixed ratio deals to those

from floating ratio deals. Using holding periods of one to five years for a portfolio formed by taking

long positions in bidders of floating ratio deals and short positions in bidders of fixed exchange ratio

deals, we run standard calendar-time regressions on the Fama-French factors plus the Carhart Mo-

mentum factor (Mitchell and Stafford, 2000). The coefficient on the intercept is economically small

and statistically insignificant for all holding periods. This evidence indicates that fixed exchange

ratio bidders do not make systematically worse deals, to the extent that deal quality is reflected

in long-run stock returns. In sum, neither of the two predictions of the alternative hypothesis is

supported by the data.

G. Reverse Causality

It is possible that the strong positive association between an increase in news coverage and

stock returns results from the coverage of extreme stock returns of the bidder. Consistent with

the opportunistic acquisition hypothesis, strong performance of the bidder’s stock may attract

the attention of news reporters or financial analysts, resulting in additional articles related to the

acquirer’s stock. This prediction matches the observed increase in media coverage and market

equity prior to the merger.

45

We address this concern in several ways. First, in our main tests, we explicitly control for lagged

stock returns and turnover to account for a delay in reporting. Second, in all of our tests, we

use media coverage that excludes stock pricing and market data, an option provided by Factiva.

We also eliminate articles with fewer than 50 words, since they are the most likely to contain

market content. Finally, to address the possibility that the bidder’s extreme stock performance

leads analysts or news reporters to write an article about the bidder, we eliminate all press releases

that contain the word ‘stock’ in the headline. After imposing previous filters, the number of such

articles is small (7.1%) in our sample and has no effect on our conclusions.

VI. Conclusion

Combining novel hand-collected data on the timing of merger negotiations with a comprehensive

dataset of media coverage, this article studies one of the main channels of active corporate com-

munication with investors — press releases — during some of the largest investments in the life

of the firm. Our results highlight an interesting pattern in a firm’s communication with investors

when management has strong incentives for favorable valuation. In particular, fixed exchange ratio

bidders dramatically increase the number of press releases disseminated to financial media during

the private negotiation of a stock merger, compared to floating exchange ratio bidders, who do not

have an incentive to manage their media during the merger negotiation. This effect is associated

with short-lived increases in both media coverage and bidder valuation.

We examine several hypotheses that may account for the observed pattern and find that our

evidence is most consistent with an active media management view. In particular, we argue that

firms issue press releases as a mechanism to raise their stock value temporarily by generating more

media coverage. While the volume of news stories increases, the number of articles with a negative

tone and downward spin decreases during merger negotiations. We estimate that active media

management reduces the takeover costs for an average acquirer by between $230 million to $558

million, or 5 to 12% of the takeover price. Evidence of subsequent stock price reversals and lower

correlation between firm-originated news and newspaper coverage contradicts the argument that

the firm is timing the merger to coincide with the release of good news. The dramatic increase

in firm-originated news after the start of merger negotiations contradicts an explanation based on

46

firms taking advantage of passively derived over-valuation. However, all of these empirical facts are

consistent with active media management in an attempt to improve the terms of the merger.

We also investigate how investors and target firms respond to an acquirer’s media management

strategy. We show that the efficacy of media management does not necessarily presume investor

irrationality. In particular, investors respond less to acquirer originated news when it is more likely

that the firm is pursuing a strategy of active media management, namely for repeat acquirers and

during the close of the merger. We also find that active media management has an effect on the

gains of the target relative to the acquirer, consistent with its purpose. A greater increase in media

coverage during merger negotiations increases a firm’s share of the merger gains.

The results of this paper suggest a new role for media in financial markets. In contrast to the

view that the information contained in media articles increases the efficiency of a market, we show

that the press can be strategically used by firms to advance their own interests. The strategic

use of media by firms is likely to affect many corporate actions beyond mergers, such as executive

compensation, stock issues and repurchases, proxy contests, and product market competition.

47

REFERENCES

Abrevaya, Jason, 1999, Leapfrog estimation of a fixed-effects model with unknown transformation

of the dependent variable, Journal of Econometrics 93, 203–228.

Ahern, Kenneth R., 2012, Bargaining power and industry dependence in mergers, Journal of Fi-

nancial Economics 103, 530–550.

Andrade, Gregor, Mark Mitchell, and Erik Stafford, 2001, New evidence and perspectives on merg-

ers, The Journal of Economic Perspectives 15, 103–120.

Barber, Brad, and Terrance Odean, 2000, Trading is hazardous to your wealth: The common stock

investment performance of individual investors, Journal of Finance 55, 773–806.

, 2008, All that glitters: The effect of attention and news on the buying behavior of individual

and institutional investors, Review of Financial Studies 21, 785–818.

Besley, Timothy, and Andrea Prat, 2006, Handcuffs for the grabbing hand? Media capture and

government accountability, The American Economic Review 96, 720–736.

Boone, Audra L., and J. Harold Mulherin, 2007, How are firms sold?, The Journal of Finance 62,

847–875.

Brandt, Michael W., Alon Brav, John R. Graham, and Alok Kumar, 2010, The idiosyncratic

volatility puzzle: Time trend or speculative episodes?, The Review of Financial Studies 23, 863–

899.

Bushee, Brian J., and Gregory S. Miller, 2007, Investor relations, firm visibility, and investor

following, Working Paper University of Pennsylvania and University of Michigan.

Cameron, A. Colin, and Pravin K. Trivedi, 1998, Regression Analysis of Count Data (Cambridge

University Press: Cambridge).

Campbell, Tim S., 1979, Optimal investment financing decisions and the value of confidentiality,

Journal of Financial and Quantitative Analysis 14, 913–924.

Chan, Wesley S., 2003, Stock price reaction to news and no-news: Drift and reversal after headlines,

Journal of Financial Economics 70, 223–260.

Chang, Saeyoung, 1998, Takeovers of privately held targets, methods of payment, and bidder

returns, Journal of Finance 53, 773–784.

48

Cohen, Lauren, Andrea Frazzini, and Christopher Malloy, 2010, Sell side school ties, Journal of

Finance 65, 1409–1437.

Cook, Douglas O., Robert Kieschnick, and Robert A. Van Ness, 2006, On the marketing of IPOs,

Journal of Financial Economics 82, 35–61.

Cooney, Jr., John W., Thomas Moeller, and Mike Stegemoller, 2009, The underpricing of private

targets, Journal of Financial Economics 93, 51–66.

Core, John E., Wayne Guay, and David F. Larcker, 2008, The power of the pen and executive

compensation, Journal of Financial Economics 88, 1–25.

Da, Zhi, Joseph Engelberg, and Pengjie Gao, 2011, In search of attention, Journal of Finance 66,

1461–1499.

Darrough, Masako N., 1993, Disclosure policy and competition: Cournot vs. Bertrand, The Ac-

counting Review 68, 534–561.

DellaVigna, Stefano, and Ethan Kaplan, 2007, The Fox News effect: Media bias and voting, The

Quarterly Journal of Economics 122, 1187–1234.

DellaVigna, Stefano, and Joshua Pollet, 2009, Investor inattention and Friday earnings announce-

ments, Journal of Finance 64, 709–749.

Demers, Elizabeth A., and Clara Vega, 2010, Soft information in earnings announcements: News

or noise?, INSEAD and Board of Governors Working Paper.

Dyck, Alexander, Natalya Volchkova, and Luigi Zingales, 2008, The corporate governance role of

the media: Evidence from Russia, The Journal of Finance 63, 1093–1135.

Dyck, Alexander, and Luigi Zingales, 2003, Corporate Governance and Capital Flows in a Global

Economy . chap. The Bubble and the Media, pp. 83–102 (Oxford University Press).

Ellis, Jesse A., C. Edward Fee, and Shawn E. Thomas, 2012, Proprietary costs and the disclosure

of information about customers, Journal of Accounting Research, forthcoming.

Engelberg, Joseph, and Christopher A. Parsons, 2011, The causal impact of media in financial

markets, Journal of Finance 66, 67–97.

Faccio, Mara, and Roland W. Masulis, 2005, The choice of payment method in European mergers

and acquisitions, The Journal of Finance 60, 1345–1388.

49

Fama, Eugene F., Lawrence Fisher, Michael C. Jensen, and Richard Roll, 1969, The adjustment of

stock prices to new information, International Economic Review 10, 1–21.

Fama, Eugene F., and Kenneth R. French, 1992, The cross-section of expected stock returns, The

Journal of Finance 47, 427–465.

Fang, Lily H., and Joel Peress, 2009, Media coverage and the cross-section of stock returns, Journal

of Finance 64, 2023–2052.

Fuller, Kathleen, Jeffry Netter, and Mike Stegemoller, 2002, What do returns to acquiring firms

tell us? Evidence from firms that make many acquisitions, Journal of Finance 57, 1763–1793.

Gaa, Charles, 2010, If it bleeds, it leads: Selective media coverage and the market’s reaction to

good and bad news, University of Oregon Working Paper.

Gentzkow, Matthew, and Jesse M. Shapiro, 2006, Media bias and reputation, Journal of Political

Economy 114, 2.

, 2010, What drives media slant? Evidence from U.S. daily newspapers, Econometrica 78,

35–71.

Greene, William H., 1994, Accounting for excess zeros and sample selection in Poisson and Negative

Binomial regression models, New York University Working Paper.

Grinstein, Yaniv, and Paul Hribar, 2004, CEO compensation and incentives: Evidence from M&A

bonuses, Journal of Financial Economics 73, 119–143.

Groseclose, Tim, and Jeffrey Milyo, 2005, A measure of media bias, The Quarterly Journal of

Economics 120, 1191–1237.

Gurun, Umit G., and Alexander W. Butler, 2012, Don’t believe the hype: Local media slant, local

advertising, and firm value, Journal of Finance 67, 561–598.

Hansen, Robert G., 1987, A theory for the choice of exchange medium in mergers and acquisitions,

The Journal of Business 60, 75–95.

Harford, Jarrad, 1999, Corporate cash reserves and acquisitions, Journal of Finance 54, 1969–1997.

Hausman, Jerry, Bronwyn H. Hall, and Zvi Griliches, 1984, Econometric models for count data

with an application to the patents-R&D relationship, Econometrica 52, 909–938.

Huberman, Gur, and Tomer Regev, 2001, Contagious speculation and a cure for cancer: A nonevent

that made stock prices soar, Journal of Finance 56, 387–396.

50

Huizinga, Harry P., and Johannes Voget, 2009, International taxation and the direction and volume

of cross-border M&As, The Journal of Finance 64, 1217–1249.

Jensen, Michael C., 1986, Agency costs of free cash flow, corporate finance, and takeovers, American

Economic Review 76, 323–329.

Kumar, Alok, 2009, Hard-to-value stocks, behavioral biases, and informed trading, Journal of

Financial and Quantitative Analysis 44, 1375–1401.

Lerner, Josh, 2006, The new new financial thing: The origins of financial innovations, Journal of

Financial Economics 79, 223–255.

, Morten Sorensen, and Per Stromberg, 2011, Private equity and long-run investment: The

case of innovation, The Journal of Finance 66, 445–477.

Liu, Laura, Ann Sherman, and Yong Zhang, 2009, The role of media in initial public offerings,

DePaul University and Hong Kong University of Science & Technology Working Paper.

Loughran, Tim, and Bill McDonald, 2011, When is a liability not a liability? Textual analysis,

dictionaries, and 10-Ks, The Journal of Finance 66, 35–65.

Loughran, Tim, and Anand M. Vijh, 1997, Do long-term shareholders benefit from corporate

acquisitions?, The Journal of Finance 52, 1765–1790.

Mitchell, Mark, Todd Pulvino, and Erik Stafford, 2004, Price pressure around mergers, Journal of

Finance 59, 31–63.

Mitchell, Mark L., and Erik Stafford, 2000, Managerial decisions and long-term stock price perfor-

mance, Journal of Business 73, 287–329.

Moeller, Sara B., Frederik P. Schlingemann, and Rene M. Stulz, 2004, Firm size and the gains from

acquisitions, Journal of Financial Economics 73, 201–228.

, 2007, How do diversity of opinion and information asymmetry affect acquirer returns?,

The Review of Financial Studies 20, 2047–2078.

Mullainathan, Sendhil, and Andrei Shleifer, 2005, The market for news, The American Economic

Review 95, 1031–1053.

Odean, Terrance, 1999, Do investors trade too much?, American Economic Review 89, 1279–1298.

Officer, Micah S., 2006, The market pricing of implicit options in merger collars, Journal of Business

79, 115–136.

51

, Annette B. Poulsen, and Mike Stegemoller, 2009, Target-firm information asymmetry and

acquirer returns, Review of Finance 13, 467–493.

O’Hara, Robert B., and D. Johan Kotze, 2010, Do not log-transform count data, Methods in Ecology

and Evolution 1, 118–122.

Pulvino, Todd C., 1998, Do asset fire sales exist? An empirical investigation of commerical aircraft

transactions, The Journal of Finance 53, 939–978.

Reuter, Jonathan, and Eric Zitzewitz, 2006, Do ads influence editors? Advertising and bias in the

financial media, The Quarterly Journal of Economics 121, 197–227.

Rhodes-Kropf, Matthew, and David Robinson, 2008, The market for mergers and the boundaries

of the firm, Journal of Finance 62, 1169–1211.

Rhodes-Kropf, Matthew, David T. Robinson, and S. Viswanathan, 2005, Valuation waves and

merger activity: The empirical evidence, Journal of Financial Economics 77, 561–603.

Rhodes-Kropf, Matthew, and S. Viswanathan, 2004, Market valuation and merger waves, The

Journal of Finance 59, 2685–2718.

Roll, Richard, 1986, The hubris hypothesis of corporate takeovers, The Journal of Business 59,

197–216.

Shleifer, Andrei, and Robert W. Vishny, 2003, Stock market driven acquisitions, Journal of Finan-

cial Economics 70, 295–311.

Solomon, David, 2012, Selective publicity and stock prices, Journal of Finance 67, 599–638.

Tetlock, Paul C., 2007, Giving content to investor sentiment: The role of media in the stock market,

Journal of Finance 62, 1139–1168.

, 2010, All the news that’s fit to reprint: Do investors react to stale information?, Review

of Financial Studies.

Vega, Clara, 2006, Stock price reaction to public and private information, Journal of Financial

Economics 82, 103–133.

Verrecchia, Robert, 1983, Discretionary disclosure, Journal of Accounting & Economics 5, 179–194.

, 1990, Endogenous proprietary costs through firm interdependence, Journal of Accounting

& Economics 12, 245–250.

52

-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 -70 -60 -50 -40 -30 -20 -10 0 -60 -50 -40 -30 -20 -10 0

Pre-negotiationPeriod

(120 days)

NegotiationPeriod

(65 days)

TransactionPeriod

(65 days)

AnnouncementPeriod(5 days)

NegotiationsBegin

PublicAnnouncement

MergerCompleted

Figure 1Timeline of an Average MergerThis figure shows the average length of time in each stage of the merger process for our sample of 507 acquisitions during 2000-2008.Numbers represent trading days relative to one of three key dates. In the Pre-Negotiation Period (120 trading days before negotiationsbegin until the day negotiations begin), days are numbered relative to the start of negotiations. In the Negotiation Period (the daynegotiations begin until 16 trading days before the public announcement), days are numbered relative to the public announcement ofthe merger. In the Transaction Period (five days after the public announcement until two days before the completion of the merger),days are numbered relative to the completion date of the merger. The announcement period denotes the five days centered on the dayof the public announcement.

53

-100 -90 -80 -70 -60 -50 -40 -30 -20 -10

Cumulative

Abnormal

Articles

Event Time Relative to Public Announcement

5

10

15

20

25

30

35Fixed exchange ratio

Floating exchange ratio

(a) The Negotiation Period

-100 -90 -80 -70 -60 -50 -40 -30 -20 -10

Cumulative

Abnormal

Articles

Event Time Relative to Merger Completion

30

60

90

120

150

180

210

Floating exchange ratio

Fixed exchange ratio

(b) The Transaction Period

Figure 2Abnormal Acquirer Newswire Coverage During MergersThis figure presents the average acquirer’s cumulative number of abnormal articles from newswiresin daily event time relative to the public announcement of the merger in subfigure (a) and relativeto the completion of the merger in subfigure (b). Abnormal articles are calculated as a firm’s dailynumber of articles divided by the average number of daily newswire articles in the pre-negotiationperiod (120 trading days prior to the beginning of merger negotiations). Time series are set to zero atevent date −100 for comparison. ‘Fixed exchange ratio’ denotes bidders in stock acquisitions wherethe number of bidder’s shares to be issued for each target share is fixed. ‘Floating exchange ratio’denotes bidders in stock acquisitions where the number of bidder’s shares to be issued for each targetshare floats to achieve a particular price per target share. Data are from 507 acquisitions during2000-2008.

54

-100 -90 -80 -70 -60 -50 -40 -30 -20 -10

Ln(M

arketEquity)

Event Time Relative to Public Announcement

0.02

0.04

0.06

0.08

0.10

0.12

0.14 Fixed exchange ratio

Floating exchange ratio

(a) The Negotiation Period

-100 -90 -80 -70 -60 -50 -40 -30 -20 -10

Ln(M

arketEquity)

Event Time Relative to Merger Completion

-0.30

-0.20

-0.10

0.00

0.10

0.20

Floating exchange ratio

Fixed exchange ratio

(b) The Transaction Period

Figure 3Acquirer Market Equity During MergersThis figure presents the average acquirer’s logged market equity in daily event time relative to thepublic announcement of the merger in subfigure (a) and relative to the completion of the merger insubfigure (b). Time series are set to zero at event date −100 for comparison. ‘Fixed exchange ratio’denotes bidders in stock acquisitions where the number of bidder’s shares to be issued for each targetshare is fixed. ‘Floating exchange ratio’ denotes bidders in stock acquisitions where the number ofbidder’s shares to be issued for each target share floats to achieve a particular price per target share.Data are from 507 acquisitions during 2000-2008.

55

Table ISummary Statistics of Media Coverage and Merger NegotiationsThis table presents summary statistics for 617,445 media articles from 421 sources reported in the Factiva database about acquirersin 507 mergers over 2000 to 2008 by different periods in the merger process. The pre-negotiation period is the period from 120trading days before the merger talks begin until the day before merger talks begin. The negotiation period starts on the day mergernegotiations begin and ends 16 trading days before the public announcement of the merger. The announcement period includes thefive days that surround the public announcement of the merger. The transaction period is the period five days after the announcementdate until two days before the merger closes. Monthly figures are aggregates of 20 trading days by firm-day observations. The WallStreet Journal, The New York Times, and USA Today are the top 3 domestic newspaper sources. Reuters News, Dow Jones NewsService, and Business Wire are the top 3 newswire sources.

Percentile

Mean Std. Dev. 25th 50th 75th Obs.

Monthly Media Coverage in Pre-Negotiation Period

Number of all articles per firm-month 73.42 195.97 0 11 60 60, 840Number of top 3 domestic news articles per firm-month 3.01 11.98 0 0 1 60, 840Number of top 3 news wire articles per firm-month 30.76 99.25 0 1 20 60, 840

Change in Monthly Media Coverage in Negotiation Period (Within-Firm-Deal)

Number of all articles per firm-month 2.85 50.78 −7.50 −0.09 7.50 27, 489Number of top 3 domestic newspaper articles per firm-month −0.12 2.93 −0.49 0.00 0.17 27, 489Number of top 3 newswire articles per firm-month 1.52 36.70 −3.42 −0.09 3.00 27, 489

Media Coverage During the Announcement Period

Number of all articles 59.70 86.93 17 30 67 2, 535Number of top 3 domestic news stories 3.23 5.18 1 2 3 2, 535Number of top 3 press releases 29.85 46.10 7 15 31 2, 535

Timing of the Merger Process

Days in Negotiation Period 64.68 63.93 17 44 92 507Days in Transaction Period 64.46 57.96 24 48 86 507

56

Table IIFirm Characteristics by Form of Merger PaymentColumns 1–3 present averages of acquirer characteristics for 377 fixed exchange ratio acquisitions, 130floating-exchange ratio acquisitions, and 139 all-cash acquisitions. Columns 4–6 present differences-in-means tests with p−values reported in parentheses. ‘Fixed exchange ratio’ denotes bidders instock acquisitions where the number of bidder’s shares to be issued for each target share is fixed.‘Floating exchange ratio’ denotes bidders in stock acquisitions where the number of bidder’s sharesto be issued for each target share floats to achieve a particular price per target share. ‘All cashoffer’ denotes bidders that offer only cash as consideration. ‘Stock return volatility’ is calculated overthe pre-negotiation period, defined in Table I. ‘Market equity’ is the value of market equity (in $billions) on the first day of negotiations. ‘Analyst dispersion’ is the coefficient of variation in analysts’quarterly earnings forecasts for the acquirer in the most recent forecasting period before the mergerannouncement. ‘Institutional ownership’ is the percent of shares owned by institutional investorsin the most recent reporting period before the merger announcement. ‘Intangibles/Assets’ is (TotalAssets - Net PPE - Current Assets)/Total Assets. ‘High-tech Industry’ is a dummy variable foracquirers in Fama French 49 Industries: Computers (35), Software (36), or Electronics (37). ‘Tobin’sQ’ is (Total assets - common equity + market equity)/Total assets. ‘Market-to-book’ is calculatedas in Fama and French (1992). ‘ROA’ is operating income before depreciation/total assets. ‘Rawmedia count’ is the number of daily media articles from all sources in the pre-negotiation period.‘Same industry’ is a dummy equal to one if the acquirer and target are in the same Fama French 49industry code. ‘Relative size’ is the merger transaction size divided by the acquirer’s market equity.Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

FixedExchangeRatio

FloatingExchangeRatio

AllCashOffer

Differences

(1) (2) (3) (1)− (2) (1) − (3) (2)− (3)

Stock return volatility 3.066 2.441 1.903 0.625∗∗∗ 1.164∗∗∗ 0.538∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Market equity 24.580 32.718 38.160 −8.138 −13.579∗∗ −5.441(0.202) (0.016) (0.464)

Analyst dispersion 0.053 0.043 0.023 0.010 0.030∗∗∗ 0.021∗∗

(0.462) (0.002) (0.040)

Institutional ownership 0.158 0.152 0.192 0.007 −0.034∗∗ −0.041∗∗

(0.624) (0.032) (0.019)

Intangibles/Assets 0.136 0.150 0.251 −0.014 −0.115∗∗∗ −0.101∗∗∗

(0.491) (< 0.001) (< 0.001)

R&D/Assets 0.030 0.024 0.032 0.006 −0.002 −0.008∗

(0.160) (0.653) (0.097)

High-Tech Industry 0.308 0.277 0.386 0.031 −0.078 −0.109∗

(0.505) (0.103) (0.058)

Tobin’s Q 2.331 2.240 2.196 0.091 0.135 0.043(0.648) (0.382) (0.834)

57

Table II – continued

FixedExchangeRatio

FloatingExchangeRatio

AllCashOffer

Differences

(1) (2) (3) (1)− (2) (1) − (3) (2)− (3)

Market-to-book 4.607 4.404 3.780 0.204 0.827∗∗ 0.624(0.718) (0.047) (0.262)

ROA 0.117 0.140 0.160 −0.024∗ −0.043∗∗∗ −0.019(0.054) (< 0.001) (0.115)

Raw media count 3.430 4.228 6.098 −0.797 −2.668∗∗∗ −1.871(0.384) (0.007) (0.109)

Same industry 0.735 0.708 0.479 0.027 0.256∗∗∗ 0.229∗∗∗

(0.558) (< 0.001) (< 0.001)

Relative size 0.497 0.565 0.188 −0.068 0.309∗∗∗ 0.377∗∗∗

(0.552) (< 0.001) (0.001)

58

Table IIIHypotheses and Testable PredictionsThis table presents a summary of the main hypotheses and their predictions defined in a difference-in-differences framework. The prediction shown is the expected change in the variable of interest forfixed ratio bidders relative to floating ratio bidders over the specified period in the merger process.The negotiation period starts on the day merger negotiations begin and ends sixteen trading daysbefore the public announcement of the merger. The transaction period starts five days after thepublic announcement of the merger and ends two trading days before the completion of the merger.

Variable of Interest

Firm-OriginatedMedia Articles

MarketEquity

Active Media Management

Negotiation Period Higher HigherTransaction Period Lower Lower

Passive Media Management

Negotiation Period Higher HigherTransaction Period Higher Higher

Opportunistic Acquisitions

Negotiation Period No difference HigherTransaction Period No difference Lower

59

Table IVUnivariate Differences-In-Differences Tests of Media CoverageThis table presents averages and univariate t−tests of media counts by media source, type of mergerpayment, and timing of merger negotiations for a sample of 507 mergers over 2000 to 2008. ‘Pre-negotiation,’ ‘Negotiation,’ and ‘Transaction’ denote time periods in the merger and are defined inTable I. ‘Fixed’ denotes bidders in fixed exchange ratio stock acquisitions. ‘Floating’ denotes biddersin floating exchange ratio stock acquisitions. Observations are at the firm-day level. The Wall StreetJournal, The New York Times, and USA Today are the top 3 domestic newspaper sources. ReutersNews, Dow Jones News Service, and Business Wire are the top 3 newswire sources. Significance atthe 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Pre-Negotiation Negotiation Transaction Differences

(1) (2) (3) (1)− (2) (1) − (3) (2)− (3)

Panel A: All daily media articles

Fixed 3.453 4.409 3.895 −0.955∗∗∗ −0.442∗∗∗ 0.513∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Floating 4.303 2.931 4.412 1.371∗∗∗ −0.110 −1.481∗∗∗

(< 0.001) (0.381) (< 0.001)

Difference −0.849∗∗∗ 1.477∗∗∗ −0.517∗∗∗ −2.327∗∗∗ −0.332∗∗ 1.994∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (0.019) (< 0.001)

Panel B: Top three daily domestic newspaper articles

Fixed 0.149 0.177 0.151 −0.029∗∗∗ −0.002 0.026∗∗∗

(< 0.001) (0.562) (0.006)

Floating 0.157 0.113 0.169 0.044∗ −0.012∗∗∗ −0.056∗∗∗

(< 0.001) (0.099) (< 0.001)

Difference −0.008 0.064∗∗∗ −0.018∗∗∗ −0.073∗∗∗ 0.010 0.082∗∗∗

(0.117) (< 0.001) (0.006) (< 0.001) (0.257) (< 0.001)

Panel C: Top three daily newswire articles

Fixed 1.441 1.900 1.665 −0.459∗∗∗ −0.223∗∗∗ 0.236∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Floating 1.818 1.146 1.926 0.673∗∗∗ −0.107∗ −0.780∗∗∗

(< 0.001) (0.096) (< 0.001)

Difference −0.377∗∗∗ 0.754∗∗∗ −0.261∗∗∗ −1.131∗∗∗ −0.116 1.016∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (0.113) (< 0.001)

60

Table VMultivariate Differences-In-Differences Tests of Media CoverageThis table presents coefficient estimates from firm-deal fixed effects regressions of media coverage.The dependent variable is the number of media articles. Observations are firm-days in the pre-negotiation and negotiation periods, defined in Table I. ‘Negotiation period dummy’ equals one forobservations in the negotiation period and zero for observations in the pre-negotiation period. ‘Fixedratio’ is a dummy variable equal to one for mergers that use a fixed stock exchange ratio and zerofor mergers that use a floating stock exchange ratio. The Wall Street Journal, The New York Times,and USA Today are the domestic newspaper sources. Reuters News, Dow Jones News Service, andBusiness Wire are the newswire sources. ‘Sum of turnovert−1,...,t−5’ is the sum of the coefficients ofturnover from each day t− 1 to t− 5. Turnover is daily volume divided by shares outstanding. ‘Sumof |returnt−1,...,t−5 |’ is computed analogously, where | returnst | are absolute values of daily returns.Heteroskedasticity- and autocorrelation-robust p−values are in parentheses. Significance at the 0.01,0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

All Media Domestic Newspapers Newswires(1) (2) (3)

Negotiation period dummy −0.144 −0.012 −0.129∗

(0.190) (0.114) (0.066)

Negotiaton period × Fixed ratio 0.277∗∗ −0.001 0.161∗

(0.044) (0.929) (0.062)

Sum of turnovert−1,...,t−5 26.668∗∗∗ 1.241∗∗∗ 14.765∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | 18.893∗∗∗ 1.324∗∗∗ 12.254∗∗∗

(0.002) (0.001) (0.005)

Sum of |returnt−1,...,t−5 | × Fixed ratio −1.452 −0.281 −2.032(0.833) (0.549) (0.666)

Firm-Deal fixed effects Yes Yes YesF−test 10.940 8.260 10.080p−value (< 0.001) (< 0.001) (< 0.001)Observations 85,808 85,808 85,808

61

Table VIThe Effect of Media Coverage on Market EquityThis table presents coefficient estimates from fixed effects regressions of ln(market equity). Observa-tions are firm-days in the pre-negotiation and negotiation periods, defined in Table I. ‘Negotiationperiod dummy’ equals one for observations in the negotiation period and zero for observations in thepre-negotiation period. ‘Fixed ratio’ is a dummy variable equal to one for mergers that use a fixedstock exchange ratio and zero for mergers that use a floating stock exchange ratio. ‘Media source’refers to one of the three categories of media sources as listed in the heading of the table. The WallStreet Journal, The New York Times, and USA Today are the domestic newspaper sources. ReutersNews, Dow Jones News Service, and Business Wire are the newswire sources. ‘Market equityt’ is theprice times shares outstanding on day t. ‘Mediat,...,t−5’ is the sum of the coefficients of media fromeach day t to t−5, where the media source is the same as the dependent variable. Heteroskedasticity-and autocorrelation-robust p−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levelsis indicated by ∗∗∗, ∗∗, and ∗.

Dependent Variable: Ln(Market Equityt)

Media Source All Media Domestic Newspapers Newswires(1) (2) (3)

Negotiation period dummy 0.105∗∗∗ 0.102∗∗∗ 0.106∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Negotiation period × Fixed ratio 0.036∗∗∗ 0.040∗∗∗ 0.034∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Mediat,...,t−5 0.002∗∗∗ 0.033∗∗∗ 0.003∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Mediat,...,t−5× Negotiation period −0.002∗∗∗ −0.041∗∗∗ −0.007∗∗∗

(< 0.001) (0.001) (< 0.001)

Mediat,...,t−5× Fixed ratio −0.001 −0.025∗∗∗ 0.000(0.100) (0.007) (0.680)

Mediat,...,t−5× Negotiation period × Fixed 0.003∗∗∗ 0.044∗∗∗ 0.008∗∗∗

(< 0.001) (0.001) (< 0.001)

Firm-Deal fixed effects Yes Yes YesF−test 49.500 49.320 48.910p−value (< 0.001) (< 0.001) (< 0.001)Observations 85,928 85,928 85,928

62

Table VIIReversals of Market Equity and Media CoverageThis table presents coefficient estimates from fixed effects regressions of ln(market equity) or thenumber of newswire articles on day t. Columns 1 and 2 test for changes in the level of market equity.The level of market equity that immediately precedes the merger announcement is the baseline fortesting the subsequent valuation adjustments that follow the announcement (dummy = zero for firm-day observations from the end of the negotiation period to three days before the announcement).Columns 3 and 4 test for changes in the daily flow of media coverage. The average daily number ofarticles in the negotiation period when the terms of the merger are being determined is the baseline(dummy = zero for firm-day observations in the negotiation period, as defined in Table I). In (1)and (3), ‘Announcement period dummy’ equals one for observations in the announcement period.In (2) and (4), ‘Transaction period dummy’ equals one for observations in the transaction period.Heteroskedasticity- and autocorrelation-robust p−values are in parentheses. Significance at the 0.01,0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Ln(Market Equityt) Newswire Articles

(1) (2) (3) (4)

Announcement period dummy −0.005∗ 3.670∗∗∗

(0.097) (< 0.001)

Announcement period × Fixed ratio −0.007∗ 0.704(0.084) (0.221)

Transaction period dummy −0.012∗∗∗ 0.304∗∗∗

(0.002) (< 0.001)

Transaction period × Fixed ratio −0.008∗ −0.189∗

(0.082) (0.061)

Sum of turnovert−1,...,t−5 −1.833 22.273∗∗∗

(0.681) (< 0.001)

Sum of |returnt−1,...,t−5 | 7.162 7.418(0.391) (0.170)

Sum of |returnt−1,...,t−5 | × Fixed ratio 5.735 1.250(0.572) (0.835)

Firm-deal fixed effects Yes Yes Yes YesF−test 13.900 31.320 12.700 7.620p−value (< 0.001) (< 0.001) (< 0.001) (< 0.001)Observations 8,826 55,277 29,723 76,178

63

Table VIIIThe Fraction of Positive and Negative ArticlesThis table presents coefficient estimates from firm-deal fixed effect regressions of the number of positive or negative articles by mediasource. An article is positive (negative) if it has more than the average fraction of positive (negative) words calculated separatelyfor each type of media outlet. Positive and negative words are categorized based on Loughran and McDonald (2010). Regressionsin Panel A include firm-day observations in the pre-negotiation and negotiation periods. Regressions in Panel B include firm-dayobservations in the negotiation and transaction periods. Time periods are defined in Table I. ‘Negotiation period dummy’ equalsone for observations in the negotiation period and zero for observations in the pre-negotiation period. ‘Transaction period dummy’equals one for observations in the transaction period and zero for observations in the negotiation period. ‘Fixed ratio’ is a dummyvariable equal to one for mergers that use a fixed stock exchange ratio and zero for mergers that use a floating stock exchange ratio.The Wall Street Journal, The New York Times, and USA Today are the domestic newspaper sources. Reuters News, Dow JonesNews Service, and Business Wire are the newswire sources. ‘Sum of turnovert−1,...,t−5’ and ‘Sum of |returnt−1,...,t−5 |’ are defined inTable V. Heteroskedasticity- and autocorrelation-robust p−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levels isindicated by ∗∗∗, ∗∗, and ∗.

All Media Domestic Newspapers Newswires

Negative Positive Negative Positive Negative Positive

(1) (2) (3) (4) (5) (6)

Panel A: Effects in the Negotiation Period

Negotiation period dummy 0.018 0.025∗∗ 0.007 0.007 0.009 −0.003(0.117) (0.027) (0.808) (0.817) (0.538) (0.827)

Negotiation period × Fixed ratio −0.030∗∗ −0.014 −0.029 0.006 −0.040∗∗ 0.012(0.031) (0.319) (0.388) (0.856) (0.019) (0.507)

Sum of turnovert−1,...,t−5 1.571∗∗∗ 0.039 0.869 0.902 0.412 0.031(< 0.001) (0.889) (0.240) (0.300) (0.290) (0.942)

Sum of |returnt−1,...,t−5 | 2.698∗∗∗ 1.193∗∗ 2.237∗ 2.657∗∗ 1.603∗∗ 1.087∗

(< 0.001) (0.021) (0.055) (0.020) (0.012) (0.096)

Sum of |returnt−1,...,t−5 | × Fixed −1.820 −0.688 0.182 −0.637 −0.727 −0.818(0.001) (0.224) (0.890) (0.620) (0.301) (0.261)

Firm-Deal fixed effects Yes Yes Yes Yes Yes YesF−test 8.740 2.290 3.870 4.690 3.040 1.070p−value (< 0.001) (0.002) (< 0.001) (< 0.001) (< 0.001) (0.379)Observations 40,753 40,753 7,671 7,671 24,865 24,865

64Table VIII – continued

All Media Domestic Newspapers Newswires

Negative Positive Negative Positive Negative Positive

(1) (2) (3) (4) (5) (6)

Panel B: Effects in the Transaction Period

Transaction period dummy 0.024∗ −0.043∗∗∗ 0.020 −0.052∗ −0.021∗ −0.048∗∗∗

(0.051) (< 0.001) (0.509) (0.074) (0.077) (0.001)

Transaction period × Fixed ratio 0.011 0.020 −0.050 0.024 −0.009 0.028(0.476) (0.167) (0.159) (0.489) (0.512) (0.115)

Sum of turnovert−1,...,t−5 2.286∗∗∗ −0.895∗∗ 2.215∗∗ −1.518 0.152 0.529(< 0.001) (0.028) (0.028) (0.111) (0.694) (0.278)

Sum of |returnt−1,...,t−5 | 1.655∗∗∗ 0.916∗ 0.560 1.955∗ 2.524∗∗∗ 1.177∗∗

(0.002) (0.056) (0.633) (0.063) (< 0.001) (0.049)

Sum of |returnt−1,...,t−5 | × Fixed −1.209 −0.026 0.242 −0.598 −2.280 −1.097(0.039) (0.962) (0.852) (0.615) (< 0.001) (0.107)

Firm-Deal fixed effects Yes Yes Yes Yes Yes YesF−test 9.220 3.430 2.070 2.700 3.900 2.650p−value (< 0.001) (< 0.001) (0.006) (< 0.001) (< 0.001) (< 0.001)Observations 40,398 40,398 7,274 7,274 25,947 25,947

65

Table IXMedia Coverage and the Division of Merger GainsThis table presents coefficient estimates from cross-sectional ordinary least squares regressions of themerger gains of targets relative to acquirers. The dependent variable is the difference of the dollarannouncement return of the target and the dollar announcement return of the acquirer, divided bythe aggregate market equity of the two firms two days prior to the announcement. Dollar abnormalreturns are firms’ returns minus the equally-weighted CRSP index return, multiplied by marketequity, and aggregated over the three days around the announcement. The Wall Street Journal, TheNew York Times, and USA Today are the domestic newspaper sources. Reuters News, Dow JonesNews Service, and Business Wire are the newswire sources. ‘Acquirer (target) media’ is the averagenumber of daily media articles of the acquirer (target) during the negotiation period, normalizedby the firms’ average number of media articles during the pre-negotiation period. ‘Fixed ratio’ is adummy variable equal to one for mergers that use a fixed stock exchange ratio and zero for mergersthat use a floating stock exchange ratio. ‘Acquirer (Target) M/B’ is the market-to-book ratio of theacquirer (target), calculated as in Fama and French (1992). ‘Relative value’ is the takeover pricedivided by the market value of the acquirer two days before the announcement. ‘Acquirer (Target)assets’ is the total assets of the acquirer (target). ‘Same industry’ is a dummy equal to one if bothfirms are in the same Fama French 49 industry code. ‘Percent cash’ is the fraction of the takeoverprice paid in cash. Robust firm-clustered p−values are in parentheses. Significance at the 0.01, 0.05,and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Dependent Variable = Target Gain Relative to Acquirer

Media = Newswires Media = Domestic Newspapers

(1) (2) (3) (4)

Acquirer media −0.008∗ −0.017∗∗∗ 0.038 −0.088(0.066) (0.004) (0.510) (0.522)

Target media 0.068∗ 0.158∗∗∗ 0.463∗∗∗ 0.468∗∗

(0.071) (0.001) (0.005) (0.017)

Acquirer media × Target media −0.018 −0.069∗ 0.355 −0.603(0.585) (0.053) (0.572) (0.637)

Fixed ratio 0.021 0.016(0.170) (0.284)

Fixed ratio × Acquirer media 0.006 0.141(0.516) (0.368)

Fixed ratio × Target media −0.144∗∗∗ −0.163(0.007) (0.480)

Fixed × Acquirer media × Target media 0.097∗∗ 0.969(0.020) (0.491)

Acquirer M/B 0.000 0.001 0.000 0.001(0.881) (0.660) (0.959) (0.721)

Target M/B −0.003 −0.003 −0.003 −0.003(0.328) (0.287) (0.409) (0.345)

66

Table IX – continued

Dependent Variable = Target Gain Relative to Acquirer

Media = Newswires Media = Domestic Newspapers

(1) (2) (3) (4)

Relative value 0.001 0.000 −0.002 −0.002(0.898) (0.964) (0.779) (0.801)

Acquirer assets −0.403 −0.560∗ −0.913∗∗∗ −0.807∗∗∗

(0.141) (0.063) (0.002) (0.003)

Target assets 0.000 0.000 0.000 0.000(0.961) (0.642) (0.320) (0.457)

Same industry −0.006 −0.003 −0.011 −0.009(0.682) (0.834) (0.451) (0.539)

Percent cash −0.051∗∗ −0.039∗ −0.047∗∗ −0.037∗

(0.011) (0.056) (0.017) (0.060)

Constant 0.093∗∗∗ 0.072∗∗∗ 0.096∗∗∗ 0.081∗∗∗

(< 0.001) (0.004) (< 0.001) (0.001)

Adjusted R2 0.073 0.126 0.117 0.133Observations 101 101 101 101

67

Table XCharacteristics of Firms with More Active Media ManagementThis table presents coefficient estimates from ordinary least squares regressions of media coverage ofacquirers from the top three domestic newspapers or top three newswires. Observations are firm-days in the pre-negotiation and negotiation periods, as defined in Table I for fixed exchange ratioacquisitions only. The Wall Street Journal, The New York Times, and USA Today are the top 3domestic newspaper sources. Reuters News, Dow Jones News Service, and Business Wire are the top3 newswire sources. ‘High market-to-book’ (denoted ‘Market-to-book’ in the interaction terms) is adummy variable equal to one for firms with above median market-to-book ratios, within the sample.‘High analyst dispersion’ (denoted ‘Analyst dispersion’ in the interaction terms) is a dummy variableequal to one if the coefficient of variation of analysts’ earnings forecasts for the acquirer in the mostrecent forecasting period before the merger announcement is above the median for the sample firms.‘High-tech industry’ (denoted ‘High-tech industry’ in the interaction terms) is a dummy variable equalto one for acquirers in Fama French 49 Industries Computers (35), Software (36), or Electronics (37).‘Institutional ownership’ is the fraction of shares owned by institutions on the most recent reportingdate before the merger announcement. ‘High R&D/Assets’ (denoted ‘R&D/Assets’ in the interactionterms)is a dummy variable equal to one for above-median R&D/Assets. ‘Negotiation period dummy’(denoted ‘Negotiation’ in the interaction terms) equals one for observations in the negotiation periodand zero for observations in the pre-negotiation period. ‘Sum of turnovert−1,...,t−5’ is the sum of thecoefficients of turnover from each day t − 1 to t − 5. Turnover is daily volume divided by sharesoutstanding. ‘Sum of |returnt−1,...,t−5 |’ is computed analogously, where | returnst | are absolutevalues of daily returns. Heteroskedasticity- and autocorrelation-robust p−values are in parentheses.Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Domestic Newspapers Newswires

(1) (2) (3) (4)

High market-to-book 0.020∗ 0.020∗∗ 0.444∗∗∗ 0.286∗∗∗

(0.061) (0.028) (< 0.001) (< 0.001)

High analyst dispersion 0.007 −0.003 0.467∗∗∗ 0.154∗∗

(0.572) (0.746) (< 0.001) (0.023)

High-tech industry 0.028 0.029∗ 0.873∗∗∗ 0.451∗∗∗

(0.210) (0.076) (< 0.001) (< 0.001)

Institutional ownership −0.032∗∗∗ −0.029∗∗∗ −0.424∗∗∗ −0.148∗

(0.002) (0.007) (< 0.001) (0.067)

High R&D/Assets −0.008 −0.011 −0.156∗∗∗ −0.107∗∗

(0.273) (0.150) (0.001) (0.044)

Negotiation period dummy 0.016 0.003 0.495∗∗∗ 0.231(0.261) (0.881) (< 0.001) (0.139)

Book assets 0.003∗∗∗ 0.003∗∗∗ 0.026∗∗∗ 0.025∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001)

Negotiation × Market-to-book −0.003 0.432∗∗

(0.883) (0.020)

Negotiation × Analyst dispersion 0.030 0.882∗∗∗

(0.311) (< 0.001)

68

Table X – continued

Domestic Newspapers Newswires

(1) (2) (3) (4)

Negotiation × High-tech industry −0.001 1.294∗∗∗

(0.986) (0.003)

Negotiation × Institutional ownership −0.009 −0.987∗∗∗

(0.804) (0.003)

Negotiation × R&D/Assets 0.006 −0.401∗

(0.773) (0.062)

Sum of turnovert−1,...,t−5 0.263 0.293 −7.769∗∗ −3.923(0.499) (0.402) (0.020) (0.217)

Sum of |returnt−1,...,t−5 | 0.187 0.183 −3.197 −2.507(0.634) (0.648) (0.221) (0.334)

Constant −0.014 −0.009 −0.345∗∗∗ −0.204∗∗

(0.404) (0.550) (0.003) (0.049)

F−test 5.340 5.140 3.790 4.080p−value (< 0.001) (< 0.001) (< 0.001) (< 0.001)Observations 20,974 20,974 20,974 20,974

Internet Appendix for

“Who Writes the News?

Corporate Press Releases During Merger Negotiations”

Kenneth R. Ahern and Denis Sosyura∗

This online appendix provides more details on the legal framework of corporate disclosure, ro-

bustness tests discussed in the paper, and an example of a merger background statement and

corresponding news articles for the CVS-Caremark merger.

I. Regulatory Mandates of Information Disclosure

A. Federal Law

The Securities and Exchange Act of 1934 regulates the timing and content of periodic filings

(annual and quarterly reports) and filings associated with specific corporate events (solicitation

of proxies, shares repurchases, changes in control, changes in outside auditor, and resignation of

directors, among others).1 The Act imposes an absolute duty to disclose the specific information

mandated by the filing within a regulated time window. Absent an absolute duty to disclose, U.S.

federal laws generally do not require firms to make public disclosure of all material corporate events

and allow for significant flexibility with respect to the content and timing of corporate news releases.

This is in contrast to regulations in other common law countries, such as Australia, which impose

a continuing duty on corporations to disclose all material information.

Although U.S. firms retain significant decision power with respect to their news releases, the

news that they do disclose must meet two main criteria: (1) it must be factually accurate and

(2) it must not be misleading by omission. For example, if a firm significantly increased its sales

by offering steep discounts, a press release that discusses the sales increase should not omit the

information about the discounts. However, a violation of these conditions by a public firm does not

∗Citation format: Kenneth R. Ahern and Denis Sosyura, YEAR: XXX, Internet Appendix to “Who Writes theNews? Corporate Press Releases During Merger Negotiations,” Journal of Finance [vol #XXX], [pages XXX],http://www.afajof.org/IA/[year XXX].asp. Please note: Wiley-Blackwell is not responsible for the content or func-tionality of any supporting information supplied by the authors. Any queries (other than missing material) shouldbe directed to the authors of the article.1See Securities and Exchange Act of 1934 sections 14 (a) and 13 (e) for proxy solicitation and share repurchases,respectively, and form 8-K, items (1), (3), and (6) for changes in control, changes in the auditor and outside directors.

2

necessarily constitute securities fraud. In particular, U.S. Supreme Court doctrine requires that a

plaintiff prove five conditions to establish the incidence of securities fraud under Section 10(b) of

the Exchange Act of 1934: (1) material misrepresentation or omission of fact; (2) scienter (i.e. the

intent of wrongdoing); (3) a connection with the purchase or sale of a security; (4) transaction and

loss causation; and (5) economic loss.

In practice, this array of requirements sets a high bar for proving securities fraud. A plaintiff

has to demonstrate that any misstatement or omission of information was material to cause an

economic loss and that the misstatement or omission was intended to deceive or defraud investors.

The presence of deceitful intent in corporate press releases is particularly difficult to prove. For

example, if a firm issues a news release with optimistic predictions about future events that do

not materialize, these predictions probably will be violations on Section 10(b) only if it can be

proven that the person making the statement did not believe it to be true at the time of press

release issuance or if this person had no reasonable basis for these predictions (see Va. Bankshares

v. Sandberg, 501 U.S. 1083 (1991) for a case involving similar facts). Moreover, firms’ forward

looking opinions are further protected from liability for securities fraud by the safe harbor provision

in the Private Securities Litigation Reform Act of 1995. Combined, these conditions establish a

high threshold for bringing cases of securities fraud based on corporate statements.

Despite the relatively lenient regulatory environment for corporate news releases, there are two

important considerations that limit managerial discretion over the timing of news disclosure. First,

corporate insiders may not trade in their company’s stock or its derivatives without disclosing

all material information in their possession, a requirement labeled the “disclose or abstain” rule.2

Second, firms may have an obligation to correct or update past statements that may have become

misleading or inaccurate, although this doctrine has met with controversy in case law.3

2For the courts’ interpretation of the rule in instances of alleged insider trading, see United States v. O’Hagan, 521U.S. 642, 643 (1997); Chiarella v. United States, 445 U.S. 222, 226-28 (1980); SEC v. Adler, 137 F.3d 1325, 1333(11th Cir. 1998).3For example, the Seventh Circuit has refused to recognize this rule (see Stransky v. Cummins Engine Co., 51 F. 3d1329, 1332 (7th Cir. 1995), although it was arguably applied in the First and Third Circuits (see Weiner vs. QuakerOats Co., 129 F.3d 310, 318 (3rd Cir, 1997) and Backman v. Polaroid Corp., 910 F.2d 10, 16-17 (1st Cir. 1990).

3

B. State Law

Delaware (and other state) corporation laws impose a fiduciary duty on corporate directors,

which can be broadly viewed as a duty to act in the best interest of the firm’s shareholders.4 This

duty is not necessarily inconsistent with information management if the information management

strategy does not harm the firm’s shareholders. For example, if strategic disclosure by bidders

during merger negotiations benefits the bidder’s shareholders by increasing the fraction of merger

gains captured by the bidder, it may not constitute a violation of fiduciary duty.

C. Listing Requirements

Listing standards of major U.S. stock exchanges require timely disclosure of material informa-

tion.5 However, these standards contain exceptions in order to allow firms to pursue their strategic

objectives and to retain their competitive position. Furthermore, the enforcement of these rules is

relatively weak, and violation of them does not give rise to a private cause of action.

D. Summary

U.S. firms have significant decision power over the content and timing of news releases, except

in cases of mandatory regulatory filings and filings associated with specific corporate events. This

notion is captured in a recent research study of securities regulation in the United States, which

concludes that “management retains substantial discretion over the nature and timing of the dis-

closure of the material facts about the company” and “all but the most egregious of the techniques

managers can use to manipulate disclosures to maximize the value of their stock options are either

legally permitted or not subject to effective legal control” (Yablon and Hill, 2000).

Reference

Yablon, Charles and Jennifer Hill, 2000, Timing Corporate Disclosures to Maximize Performance-

Based Remuneration: A Case of Misaligned Incentives? Wake Forest Law Review 35, 83–122.

4For example, in Delaware, the fiduciary duties include those of loyalty, care, and good faith. In Texas, the fiduciaryduty is defined as a duty of loyalty, care, good faith, and obedience (Gearhart Industries, Inc. v. Smith International,Inc., 741 F.2d 707, 719 (5th Cir. 1984).5NYSE Listed Company Manual P 202.05 and NASD Rules 4310 (c)(16), 4320 (e)(14), IM-4120-1.

4

II. Robustness Tests and Additional Tables

5

Internet Appendix Table IMost Frequent Media Sources by TypeThis table presents the fifteen most frequent media sources in our sample by type: domestic newspa-pers, newswires, and foreign newspapers written in English. Data is from 617,445 articles over 421sources reported in the Factiva database, excluding articles with fewer than 50 words and articlestagged by Factiva as recurring pricing and market data. Articles must also include a Factiva intelli-gent indexing code for one of the 507 acquirers in our merger sample from 2000 to 2008. Circulationdata are from the 2009 Audit Bureau of Circulation.

Rank Media Source Sample Articles Percent of Total Circulation

Domestic Newspapers

1 The Wall Street Journal 16,471 2.67 2,024,2692 The New York Times 6,450 1.04 927,8513 The Washington Post 3,631 0.59 582,8444 St. Louis Post-Dispatch 3,228 0.52 213,4725 Chicago Sun-Times 3,189 0.52 275,6416 Barron’s 2,387 0.39 303,0347 Seattle Post-Intelligencer 2,183 0.35 263,5888 Pittsburgh Post-Gazette 2,179 0.35 184,2329 The Boston Globe 1,762 0.29 264,105

10 The San Francisco Chronicle 1,697 0.27 251,78211 New York Daily News 1,524 0.25 544,16712 USA Today 1,262 0.20 1,900,11613 Times-Picayune 1,083 0.18 159,65514 Denver Post 1,050 0.17 340,94915 BusinessWeek 745 0.12 917,568

Sum of Top 15 48,841 7.91 9,153,273

Newswires

1 Reuters News 139,789 22.642 Dow Jones News Service 98,373 15.933 Business Wire 38,540 6.244 Associated Press Newswires 31,830 5.165 PR Newswire 25,707 4.166 Federal Filings Newswires 23,110 3.747 Dow Jones Business News 22,193 3.598 Dow Jones International News 20,032 3.249 PR Newswire (U.S.) 19,219 3.11

10 M2 Presswire 16,142 2.6111 Professional Investor Report 9,528 1.5412 Regulatory News Service 7,657 1.2413 Dow Jones Chinese Financial Wire 5,309 0.8614 Dow Jones Corporate Filings Alert 4,414 0.7115 Capital Markets Report 4,079 0.66

Sum of Top 15 461,843 74.80

Foreign English-Language Newspapers

1 Financial Times (U.K.) 8,045 1.30 426,6762 National Post (Canada) 6,066 0.98 150,8843 The Wall Street Journal Europe 5,738 0.93 74,9464 The Globe and Mail (Canada) 5,077 0.82 301,8205 The Asian Wall Street Journal 3,771 0.61 82,1866 The Australian 2,115 0.34 138,7657 The Times (U.K.) 1,999 0.32 617,4838 South China Morning Post 1,958 0.32 104,0009 The Guardian (U.K.) 1,812 0.29 358,844

10 The Economic Times (India) 1,726 0.28 620,00011 The Toronto Star 1,338 0.22 314,17312 Business Times Singapore 1,023 0.17 35,70013 Irish Times 969 0.16 106,92614 International Herald Tribune 945 0.15 242,07315 The New Straits Times (Malaysia) 838 0.14 241,000

Sum of Top 15 42,582 6.90 3,815,476

Total Articles from All Sources 617,445

6

Internet Appendix Table IIMultinomial Logit Tests of the Form of Merger PaymentThis table presents the coefficient estimates from a cross-sectional multinomial logit model where thedependent variable is a dummy variable indicating either a fixed exchange ratio, a floating exchangeratio, or an all-cash deal. Coefficients reflect the increase in the odds of observing a fixed ratio relativeto a floating ratio deal (first column), or a fixed ratio deal relative to an all-cash offer (second column).All variables are defined in Table II of the main paper. Estimates are calculated from one model, thusthe Pseudo R2 and observations are for the entire multinomial logit model. Heteroskedasticity-robustp−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and∗.

Fixed:Floating Fixed:Cash

Stock return volatility 0.193∗∗ 0.671∗∗∗

(0.040) (< 0.001)

ln(Market equity) 0.020 0.336∗∗

(0.869) (0.043)

Institutional ownership 0.542 −1.697∗

(0.525) (0.059)

Intangibles/Assets −0.293 −3.291∗∗∗

(0.672) (< 0.001)

Tobin’s Q −0.110 −0.045(0.157) (0.602)

ln(1+Raw media counts) −0.113 −0.588∗∗∗

(0.593) (0.005)

Same industry 0.189 0.844∗∗∗

(0.454) (0.001)

Relative size −0.124 3.152∗∗

(0.295) (0.013)

Constant 0.669 −1.100∗

(0.127) (0.078)

Pseudo R2 0.269 0.269Observations 228 228

7

Internet Appendix Table IIIVolatility Coefficients in Multinomial Logit Tests of the Form of Merger PaymentThis table presents coefficient estimates on acquirer stock price volatility from cross-sectional multi-nomial logit models where the dependent variable is a dummy variable indicating either a fixedexchange ratio, a floating exchange ratio, or an all-cash deal. Coefficients reflect the increase in theodds of observing a fixed ratio relative to a floating ratio deal, or a fixed ratio deal relative to anall-cash offer. All models include identical regressors as in Table II, except volatility is calculated insix different ways. ‘Volatility−3,0’ is calculated using daily data over the three month period that endsat the start of merger negotiations. ‘Volatility−12,0’ is calculated using daily data over the twelvemonth period that ends at the start of merger negotiations. ‘Volatility−12,−6’ is calculated usingdaily data over the six month period that ends six months before the start of merger negotiations.‘Total Volatility’ indicates that volatility is calculated using all daily returns. ‘Downward Volatility’indicates that volatility is calculated using only negative returns. Heteroskedasticity-robust p−valuesare in parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Fixed:Floating Fixed:Cash

Total Volatility Downward Volatility Total Volatility Downward Volatility

Volatility−3,0 0.171∗∗ 0.484∗∗ 0.591∗∗∗ 0.653∗∗

(0.049) (0.035) (< 0.001) (0.028)

Volatility−12,0 0.265∗∗∗ 0.524∗ 0.734∗∗∗ 1.539∗∗∗

(0.007) (0.052) (< 0.001) (< 0.001)

Volatility−12,−6 0.286∗∗∗ 0.521∗∗ 0.643∗∗∗ 1.269∗∗∗

(0.004) (0.044) (< 0.001) (< 0.001)

8

Internet Appendix Table IVDifferences-In-Differences Tests of Media Coverage: Cross-Sectional EffectsThis table presents coefficient estimates from firm-deal fixed effects regressions of media coveragefrom all sources. Observations are firm-days in the pre-negotiation and negotiation periods, definedin Table I of the main paper. The variable ‘All stock’ is a dummy variable that indicates that theform of payment only includes stock, with no cash portion. The variable ‘Weak collar’ (‘Strongcollar’) is a dummy variable that indicates the presence of a collar that has a percentage price rangethat is greater (less) than the median percentage price range. The percentage price range is theupside collar bound price in percentage above the base price minus the downside collar bound pricein percentage below the base price. Thus a weak collar is one that is less restrictive than the mediancollar. In column 2, only mergers that include a collar are included, where ‘Weak collar’ is theomitted category. In column 3, all mergers are included, including those with no collar, those withweak collars, and those with strong collars. The omitted category is no collar. ‘Turnover and |return|controls’ indicates the inclusion of the sum of the coefficients of turnover from each day t − 1 tot − 5, ‘Sum of |returnt−1,...,t−5 |’, and the interaction terms ‘Sum of |returnt−1,...,t−5 | ×Fixed’. SeeTable V of the main paper for definitions. Heteroskedasticity- and autocorrelation-robust p−valuesare in parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

(1) (2) (3)

Negotiation period dummy 0.008 −0.386 −0.215(0.948) (0.180) (0.235)

Negotiaton period × Fixed ratio 0.154 0.478 0.343∗

(0.284) (0.128) (0.085)

Negotiation period × All stock −0.725∗∗

(0.017)

Negotiation period × All stock × Fixed ratio 0.664∗

(0.055)

Negotiation period × Weak collar 0.071(0.758)

Negotiation period × Strong collar 0.476 0.316(0.147) (0.186)

Negotiation period × Fixed ratio × Weak collar 0.097(0.853)

Negotiation period × Fixed ratio × Strong collar −0.969∗∗ −0.772∗∗∗

(0.014) (0.009)

Turnovert−1,...,t−5 and |returnt−1,...,t−5 | controls Yes Yes YesFirm-deal fixed effects Yes Yes YesObservations 85,808 6,600 85,808F−test 10.180 1.790 11.990p−value (< 0.001) (0.019) (< 0.001)

9

Internet Appendix Table VDifferences-In-Differences Tests of Media Coverage: Alternative DistributionsThis table presents coefficient estimates from firm-deal fixed effects regressions of media coverage.Observations are firm-days in the pre-negotiation and negotiation periods, defined in Table I of themain paper. Media coverage includes all sources in Panel A, the top three domestic newspapers (TheWall Street Journal, The New York Times, and USA Today) in Panel B, and the top three newswiresources (Reuters News, Dow Jones News Service, and Business Wire) in Panel C. In column 1, mediacounts are winsorized at the 1st and 99th percentile, within each firm’s time-series of media coverageduring the pre-negotiation and negotiation periods and for each media source independently. Incolumn 2, the dependent variable is ln(1 + the number of media articles). Columns 1 and 2 presentordinary least squares (OLS) estimates. In columns 3 through 5, the dependent variable is theunadjusted count of media articles. Column 3 presents estimates from a generalized linear modelassuming a Poisson distribution and a log link. Column 4 presents estimates from a zero-inflatedPoisson regression. Column 5 presents estimates from a zero-inflated negative binomial regression.Ln(market equity) is the explanatory variable in the first stage of the zero-inflated models. Allvariables are defined in Table V of the main paper. Heteroskedasticity- and autocorrelation-robustp−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and∗.

WinsorizedOLS

LoggedMediaOLS

PoissonRegression

Zero-inflatedPoisson

Regression

Zero-inflatedNegative Binomial

Regression(1) (2) (3) (4) (5)

Panel A. All Media Sources

Negotiation period dummy −0.131 −0.018 −0.039∗ −0.271∗∗∗ −0.056∗∗

(0.211) (0.356) (0.100) (< 0.001) (0.029)

Negotiaton period × Fixed ratio 0.241∗ 0.030 0.072∗∗ 0.394∗∗∗ 0.109∗∗∗

(0.062) (0.189) (0.011) (< 0.001) (0.001)

Ln(Market equity) 0.381∗∗∗ 0.670∗∗∗ 0.646∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Sum of turnovert−1,...,t−5 26.379∗∗∗ 2.172∗∗∗ 7.412∗∗∗ 5.129∗∗∗ 4.872∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | 17.784∗∗∗ 5.354∗∗∗ 6.429∗∗∗ 7.437∗∗∗ 12.839∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | × Fixed −1.759 −0.294 −1.574 0.421 −1.163(0.783) (0.703) (0.262) (0.594) (0.122)

Firm-Deal fixed effects Yes No Yes No NoObservations 85,808 85,808 85,808 85,808 85,808

Continued on next page

10

Internet Appendix Table V – continued

WinsorizedOLS

LoggedMedia

PoissonRegression

Zero-inflatedPoisson

Regression

Zero-inflatedNegative Binomial

Regression(1) (2) (3) (4) (5)

Panel B. Domestic Newspapers

Negotiation period dummy −0.011 −0.012∗∗ −0.081 −0.185∗∗∗ −0.201∗∗∗

(0.113) (0.014) (0.103) (< 0.001) (< 0.001)

Negotiaton period × Fixed ratio −0.002 0.024∗∗∗ −0.002 0.193∗∗∗ 0.192∗∗∗

(0.787) (< 0.001) (0.968) (< 0.001) (< 0.001)

Ln(Market equity) 0.068∗∗∗ 0.584∗∗∗ 0.631∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Sum of turnovert−1,...,t−5 1.126∗∗∗ −0.407∗∗∗ 12.722∗∗∗ 5.133∗∗∗ 5.556∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | 1.187∗∗∗ 1.876∗∗∗ 8.448∗∗∗ 5.327∗∗∗ 8.894∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | × Fixed −0.211 −0.974∗∗∗ −2.335 4.454∗∗∗ 1.729(0.608) (< 0.001) (0.369) (0.001) (0.226)

Firm-Deal fixed effects Yes No Yes No NoObservations 85,808 85,808 85,808 85,808 85,808

Panel C. Newswires

Negotiation period dummy −0.119∗ −0.060∗∗∗ −0.089∗∗ −0.264∗∗∗ −0.097∗∗∗

(0.070) (< 0.001) (0.017) (< 0.001) (0.003)

Negotiaton period × Fixed ratio 0.134∗ 0.080∗∗∗ 0.107∗∗ 0.462∗∗∗ 0.215∗∗∗

(0.090) (< 0.001) (0.017) (< 0.001) (< 0.001)

Ln(Market equity) 0.258∗∗∗ 0.537∗∗∗ 0.556∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Sum of turnovert−1,...,t−5 14.609∗∗∗ 1.428∗∗∗ 7.738∗∗∗ 5.044∗∗∗ 6.624∗∗∗

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | 10.983∗∗∗ 5.094∗∗∗ 8.934∗∗∗ 12.392∗∗∗ 18.481∗∗∗

(0.006) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | × Fixed −2.018 −1.671∗∗∗ −2.266 −3.368∗∗∗ −6.302∗∗∗

(0.629) (0.005) (0.282) (0.001) (< 0.001)

Firm-Deal fixed effects Yes No Yes No NoObservations 85,808 85,808 85,808 85,808 85,808

11

Internet Appendix Table VIThe Fraction of Overstated and Understated ArticlesThis table presents coefficient estimates from firm-deal fixed effect regressions of the number ofoverstated or understated newswire articles, relative to domestic newspapers. Observations are firm-days in the pre-negotiation and negotiation periods, defined in Table I of the main paper. Thedependent variable in column 1 is computed at the firm-day level, and takes the value of one if thedifference between the fraction of overstated words in the top three newswires and the fraction ofoverstated words in the top three domestic newspapers is greater than the 75th percentile, within thefirm’s time-series of the difference in the pre-negotiation and negotiation periods. The variable takesthe value of −1 if the difference is less than the 25th percentile, and takes the value of zero otherwise.The dependent variable in column 2 is constructed analagously for the fraction of words that areunderstated. The dependent variable in column 3 is computed analagously using the difference inthe fraction of overstated minus the fraction of understated words, on the same day. Overstatedand understated words are categorized based on the classification of words in the Harvard-IV-4 andLasswell dictionaries provided by General Inquirer. The Wall Street Journal, The New York Times,and USA Today are the domestic newspaper sources. Reuters News, Dow Jones News Service, andBusiness Wire are the newswire sources. ‘Turnover and |return| controls’ indicates the inclusion ofthe sum of the coefficients of turnover from each day t−1 to t−5, ‘Sum of |returnt−1,...,t−5 |’, and theinteraction terms ‘Sum of |returnt−1,...,t−5 | ×Fixed’. See Table V of the main paper for definitions.Heteroskedasticity- and autocorrelation-robust p−values are in parentheses. Significance at the 0.01,0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Overstated UnderstatedOverstated −Understated

(1) (2) (3)

Negotiation period dummy 0.040 0.095∗∗ −0.051(0.375) (0.029) (0.228)

Negotiaton period × Fixed ratio 0.061 −0.114∗∗ 0.125∗∗

(0.252) (0.027) (0.014)

Turnover and |returns| controls Yes Yes YesFirm-Deal fixed effects Yes Yes YesF−test 2.090 1.720 1.590p−value (0.005) (0.033) (0.058)Observations 6,100 6,100 6,100

12

Internet Appendix Table VIIThe Effect of Acquisition Experience on Media and Market EquityThis table presents coefficient estimates from firm-deal fixed effects regressions of media articles inPanel A and ln(market equity) in Panel B. Observations are firm-days in the pre-negotiation andnegotiation periods, defined in Table I of the main paper. The dummy variable ‘Active media’ equalsone for deals where the acquirer had a prior merger in the sample where it used a fixed exchangeratio, its media coverage from all sources and market equity increased from the pre-negotiation tonegotiation period, and it had reversals in its stock price during the announcement and transactionperiods. The Wall Street Journal, The New York Times, and USA Today are the domestic newspapersources. Reuters News, Dow Jones News Service, and Business Wire are the newswire sources. SeeTable V and Table VI of the main paper for definitions. Heteroskedasticity- and autocorrelation-robust p−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗,∗∗, and ∗.

Media Source All MediaDomestic

NewspapersNewswires

(1) (2) (3)

Panel A. Dependent Variable: Media Coveraget

Negotiation period dummy 0.072 0.000 0.009(0.466) (0.971) (0.887)

Negotiaton period × Fixed ratio 0.049 −0.014 0.023(0.705) (0.127) (0.775)

Negotiaton period × Active media −2.817∗∗∗ −0.129∗∗∗ −1.790∗∗∗

(0.001) (0.008) (0.002)

Negotiaton period × Active media × Fixed ratio 2.900∗∗∗ 0.165∗∗∗ 1.734∗∗∗

(0.001) (0.001) (0.003)

Sum of turnovert−1,...,t−5 26.157∗∗∗ 1.240∗∗∗ 14.434∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | 19.340∗∗∗ 1.295∗∗∗ 12.399∗∗∗

(< 0.001) (0.001) (< 0.001)

Sum of |returnt−1,...,t−5 | × Fixed ratio −1.832 −0.245 −2.053(0.792) (0.605) (0.666)

F−test 10.190 9.350 7.580p−value (< 0.001) (< 0.001) (< 0.001)Observations 85,808 85,808 85,808

continued on next page

13

Panel B. Dependent Variable: Ln(Market Equityt)

Negotiation period dummy 0.108∗∗∗ 0.106∗∗∗ 0.111∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Negotiaton period × Fixed ratio 0.027∗∗∗ 0.030∗∗∗ 0.024∗∗∗

(0.002) (< 0.001) (0.005)

Negotiaton period × Active media −0.079∗∗∗ −0.083∗∗∗ −0.091∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Negotiaton period × Active media × Fixed ratio 0.214∗∗∗ 0.211∗∗∗ 0.228∗∗∗

(< 0.001) (< 0.001) (< 0.001)

Mediat,...,t−5 0.002∗∗∗ 0.034∗∗∗ 0.002∗∗∗

(< 0.001) (< 0.001) (0.007)

Mediat,...,t−5 × Negotiation period −0.002∗∗∗ −0.038∗∗∗ −0.007∗∗∗

(0.001) (0.005) (< 0.001)

Mediat,...,t−5 × Fixed ratio −0.001 −0.022∗∗ 0.000(0.295) (0.022) (0.782)

Mediat,...,t−5 × Active media 0.000 −0.025∗ 0.001(0.958) (0.088) (0.625)

Mediat,...,t−5 × Negotiation period × Fixed ratio 0.002∗∗∗ 0.038∗∗∗ 0.008∗∗∗

(0.001) (0.009) (< 0.001)

Mediat,...,t−5 × Negotiation period × Active media 0.002 0.026 0.009∗∗∗

(0.138) (0.323) (< 0.001)

Mediat,...,t−5 × Negotiation × Active media × Fixed −0.003∗ −0.035 −0.012∗∗∗

(0.054) (0.241) (< 0.001)

F−test 30.360 30.270 30.090p−value (< 0.001) (< 0.001) (< 0.001)Observations 85,928 85,928 85,928

14

Internet Appendix Table VIIIMedia Coverage of Targets and Floating Ratio Bidders During NegotiationsThis table presents coefficient estimates from firm-deal fixed effects regressions of media coverage. Thedependent variable is the number of media articles. Observations are firm-days in the pre-negotiationand negotiation periods, defined in Table I of the main paper. The Wall Street Journal, The NewYork Times, and USA Today are the domestic newspaper sources. Reuters News, Dow Jones NewsService, and Business Wire are the newswire sources. ‘Target dummy’ equals one for observations oftargets’ media coverage and zero for observations of floating exchange ratio bidders’ media coverage.These regressions correspond to identical tests presented in Table V of the main paper. See thecaption of Table V for additional details. Heteroskedasticity- and autocorrelation-robust p−valuesare in parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

All Media Domestic Newspapers Newswires(1) (2) (3)

Negotiation period dummy −0.145 −0.012 −0.130∗

(0.189) (0.115) (0.065)

Negotiaton period × Target dummy 0.203∗ 0.016∗∗ 0.156∗∗

(0.084) (0.049) (0.037)

Turnover and |returns| controls Yes Yes YesFirm-Deal fixed effects Yes Yes YesF−test 3.710 5.250 2.980p−value (< 0.001) (< 0.001) (< 0.001)Observations 48,581 48,581 48,581

15

Internet Appendix Table IXDifferences-In-Differences Tests of Insider TradingThis table presents coefficient estimates from firm-deal fixed effects regressions of net share purchases(purchases minus sales) by insiders. Open-market transactions by insiders in the TFN Insider FilingDatabase are included. Observations must not be missing the transaction price, number of shares,or transaction date, and cannot have CLEANSE code S or A. Observations are firm-days in thepre-negotiation and negotiation periods, defined in Table I of the main paper. ‘Negotiation perioddummy’ equals one for observations in the negotiation period and zero for observations in the pre-negotiation period. ‘Fixed ratio’ is a dummy variable equal to one for mergers that use a fixed stockexchange ratio and zero for mergers that use a floating stock exchange ratio. Column 1 includesall firm-day observations. Column 2 only includes non-zero observations. Heteroskedasticity- andautocorrelation-robust p−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levels isindicated by ∗∗∗, ∗∗, and ∗.

All Observations Non-Zero Observations(1) (2)

Negotiation period dummy −7.372 −162.729(0.208) (0.229)

Negotiation period × Fixed ratio 2.814 182.634(0.739) (0.284)

Firm-Deal fixed effects Yes YesF−test 1.070 0.740p−value (0.342) (0.477)Observations 87,675 5,040

16

Internet Appendix Table XRobustness Tests for Differences-in-Differences of Media CoverageThis table presents coefficient estimates from firm-deal fixed effects regressions of media coverage.These regressions correspond to identical tests presented in Table V of the main paper. See thecaption of Table V for additional details. In Panel A, the sample includes only mergers where theacquirer had no articles in the top three domestic newspapers that included a merger-related wordin the title during the negotiation period. Merger-related words are all variations of merge, deal,bid, acquire, acquisition, and takeover. In Panel B, the sample includes only mergers where therewas no collar used, according to SDC. Heteroskedasticity- and autocorrelation-robust p−values arein parentheses. Significance at the 0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Media Source All Media Domestic Newspapers Newswires

Panel A. Excluding Deals With Merger-Related Words in Media Coverage

Negotiation period dummy −0.137 −0.015∗∗ −0.130∗∗

(0.149) (0.017) (0.029)

Negotiaton period × Fixed ratio 0.247∗∗ 0.013∗ 0.175∗∗

(0.021) (0.063) (0.011)

Turnover and |returns| controls Yes Yes YesFirm-Deal fixed effects Yes Yes YesF−test 8.380 7.500 5.870p−value (< 0.001) (< 0.001) (< 0.001)Observations 72,539 72,539 72,539Deals 444 444 444

Panel B. Excluding Deals With Collars

Negotiation period dummy −0.222 −0.011 −0.249∗∗

(0.221) (0.345) (0.033)

Negotiation period × Fixed ratio 0.350∗ −0.003 0.276∗∗

(0.079) (0.793) (0.031)

Turnover and |returns| controls Yes Yes YesFirm-Deal fixed effects Yes Yes YesF−test 10.040 8.560 7.640p−value (< 0.001) (< 0.001) (< 0.001)Observations 72,943 72,943 72,943Deals 431 431 431

17

Internet Appendix Table XIDifferences-In-Differences Tests of Analysts’ Estimates and RecommendationsThis table presents coefficient estimates from firm-deal fixed effects regressions of analysts’ earnings-per-share (EPS) estimates for the next quarter, the next year, and the next two years, and analysts’stock recommendation (1=Sell, 2=Underperform, 3=Hold, 4=Buy, 5=Strong Buy) for the acquirersin our sample. Observations are firm-days in the pre-negotiation and negotiation periods, definedin Table I of the main paper. Analysts’ estimates and recommendations data are from I/B/E/Sand are recorded at the daily-level based on the announcement date of the forecast or recommenda-tion. See the caption of Table V of the main paper for additional details. Heteroskedasticity- andautocorrelation-robust p−values are in parentheses. Significance at the 0.01, 0.05, and 0.10 levels isindicated by ∗∗∗, ∗∗, and ∗.

EPS Forecast Recommendation

1 quarter 1 year 2 years(1) (2) (3) (4)

Negotiation period dummy 0.70 1.97 2.75 −0.04(0.359) (0.414) (0.320) (0.351)

Negotiation period × Fixed ratio −0.62 −1.57 −2.25 0.03(0.411) (0.515) (0.416) (0.520)

Firm-Deal fixed effects Yes Yes Yes YesF−test 7.44 13.19 15.16 0.48p−value (0.001) (< 0.001) (< 0.001) (0.620)Observations 24,609 30,606 28,741 9,267

18

Internet Appendix Table XIICalendar-Time Regressions of a Long-Short Portfolio with a Long Position in FloatingRatio Bidders and a Short Position in Fixed Ratio BiddersThis table presents coefficient estimates from time-series regressions. Each month, from January 2000to December 2009, equal-weighted portfolios are formed from all sample firms that have completed amerger during a previous holding period, from one year to five years, with long positions in floatingexchange ratio bidders and short positions in fixed exchange ratio bidders. The portfolio is rebalancedmonthly to drop all companies that reach the end of the holding period and add all companies thathave just completed a merger in the prior month. The table reports coefficient estimates from time-series regressions of these long-short portfolio returns on four factors and an intercept. The factorsare zero-investment portfolios representing the excess return of the market, ‘RM-RF’; the differencebetween a portfolio of small stocks and big stocks, ‘SMB’; the difference between a portfolio of highbook-to-market stocks and low book-to-market stocks, ‘HML’; and the difference between a portfolioof stocks with high past one-year returns minus a portfolio of stocks with low past one-year returns,‘UMD’. Factors are from Kenneth French’s website. The intercept measures the average monthlyabnormal return, given the model. Newey-West p−values are in parentheses. Significance at the0.01, 0.05, and 0.10 levels is indicated by ∗∗∗, ∗∗, and ∗.

Portfolio Holding Period

1-year 2-year 3-year 4-year 5-year

Intercept 0.001 0.002 0.002 0.001 < 0.001(0.714) (0.352) (0.374) (0.774) (0.971)

RM-RF −0.241∗∗ −0.284∗∗∗ −0.259∗∗∗ −0.320∗∗∗ −0.313∗∗∗

(0.038) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

SMB −0.107 −0.075 −0.069 −0.118 −0.151∗

(0.256) (0.313) (0.387) (0.143) (0.063)

HML 0.571∗∗∗ 0.536∗∗∗ 0.592∗∗∗ 0.538∗∗∗ 0.526∗∗∗

(0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)

UMD 0.258∗∗ 0.145∗∗ 0.148∗ 0.134∗ 0.156∗∗

(0.013) (0.034) (0.087) (0.098) (0.016)

F−test 8.020 13.240 12.770 14.950 14.540p−value (< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001)Observations 120 120 120 120 120

19

III. CVS-Caremark Merger Background

As filed with the Securities and Exchange Commission on January 9, 2007SECURITIES AND EXCHANGE COMMISSION

Washington, D.C. 20549AMENDMENT NO. 1

TOForm S-4

REGISTRATION STATEMENTUNDER

THE SECURITIES ACT OF 1933CVS CORPORATION

Background of the MergerOn October 7, 2005, Thomas M. Ryan, chairman, president and chief executive officer of CVS, and E.

Mac Crawford, chairman, president and chief executive officer of Caremark, spoke by telephone to arrangea subsequent meeting.

On October 20, 2005, Mr. Ryan and Mr. Crawford met in Providence, Rhode Island. During this initialmeeting, Mr. Ryan and Mr. Crawford discussed the potential strategic fit of the two organizations andthe complementary nature of the services provided by each company. Mr. Ryan and Mr. Crawford agreedto evaluate the possible synergies that might be derived from a potential strategic transaction between theparties.

In late October 2005, Evercore was requested to act as a financial advisor to CVS in connection with apotential strategic transaction with Caremark.

On November 2, 2005, the CVS board of directors met and, at the meeting, Mr. Ryan apprised the CVSboard of directors of his meeting with Mr. Crawford and the possibility of exploring a strategic transactionwith Caremark.

On November 9, 2005, the Caremark board of directors met and discussed possible strategic opportunitiesfor Caremark.

In mid-November 2005, UBS was requested to act as a financial advisor to Caremark in connection witha potential strategic transaction with CVS.

On November 22, 2005, Mr. Ryan and Mr. Crawford met at the Westin Hotel in Providence, RhodeIsland to discuss potential growth opportunities for a combined company.

On December 5, 2005, representatives of Evercore, as a financial advisor to CVS, and UBS, as a financialadvisor to Caremark, also met to discuss potential financial terms of a potential business combination betweenCVS and Caremark.

On January 4, 2006, the CVS board of directors met and received an update from Mr. Ryan on thestatus of discussions with Mr. Crawford regarding a possible strategic transaction with Caremark. At thismeeting, Mr. Ryan discussed the strategic rationale for such a combination as well as certain governanceand organizational issues.

On January 30, 2006, a meeting between representatives of CVS and representatives of Caremark tookplace at the Ritz-Carlton Hotel in Atlanta, Georgia. During this meeting, the parties discussed a potentialbusiness combination between CVS and Caremark and agreed to use a third party consultant to calculatepossible synergies that might result from such a transaction. The parties also entered into a non-disclosureagreement.

On February 2, 2006, CVS and Caremark jointly retained through legal counsel an outside consultantto assist with the calculation of potential synergies, and on February 6, 2006, representatives of CVS andrepresentatives of Caremark participated in a conference call with representatives from this firm to commencethis process. During February and March 2006, CVS and Caremark provided data to the synergy consultant.The third party consultants activities were limited to the performance of mathematical calculations based

20

on the data provided by CVS and Caremark to the third party consultant. The substantive aspects of theanalyses were performed by management of CVS and Caremark.

On March 1, 2006, the CVS board of directors met and received an update from Mr. Ryan on the statusof discussions with Caremark.

On March 7, 2006, the synergy consultant participated in a conference call with representatives of CVSand representatives of Caremark to discuss their preliminary findings with respect to potential synergies thatmay arise from a combination.

On March 14, 2006, Mr. Ryan and Mr. Crawford met at the St. Regis Hotel in New York, New York todiscuss the possible financial and strategic implications of a business combination between the two parties.Mr. Ryan and Mr. Crawford also discussed potential organizational and governance matters, includingboard composition of the combined company as well as the effect of a potential business combination onemployees, clients, payors and customers of both CVS and Caremark.

In late March 2006, Mr. Ryan called Mr. Crawford, and the parties agreed to terminate any furtherpreliminary discussions of a potential business combination between CVS and Caremark until CVS hadcompleted the then pending acquisition of the Osco/Sav-on stand-alone drugstores.

On April 5, 2006, the Caremark board of directors met telephonically and received an update from Mr.Crawford on the status of discussions with CVS and other strategic opportunities.

On May 11, 2006, the CVS board of directors met and received a presentation by certain members ofCVS management on the PBM industry landscape and the strategic rationale and benefits of a businesscombination with a pharmacy benefit management company.

On May 11, 2006, the Caremark board of directors also met and received a presentation from Caremarkssenior management on potential strategic opportunities for Caremark.

On June 2, 2006, CVS completed the Osco/Sav-on stand-alone drugstore acquisition from Albertsons.On August 16, 2006, the Caremark board of directors met and received a presentation from Caremarks

senior management on potential strategic opportunities for Caremark, including a discussion of a potentialtransaction with a retail pharmacy chain.

On August 22 and 23, 2006, Mr. Ryan and Mr. Crawford met in North Carolina, and the parties agreed tomove forward with their discussions and evaluation of a business combination between CVS and Caremark.At this meeting, Mr. Ryan and Mr. Crawford further discussed the strategic rationale for the transactionas well as corporate governance and organizational matters.

On August 30, 2006, Mr. Ryan met with David B. Snow, Jr., the chief executive officer of Medco HealthSolutions, Inc., referred to as Medco, to discuss the existing business relationship between the companies, aswell as industry trends. At the meeting, Mr. Snow raised the subject of a strategic transaction but specificswere not discussed nor were these discussions further pursued.

On September 13, 2006, Mr. Ryan met with George E. Paz, the chief executive officer of Express Scripts,Inc., to discuss existing business arrangements between the companies, as well as industry trends especiallyin light of the trend towards consumerism in the healthcare industry. No specific transactions between theparties were discussed.

On September 20, 2006, the CVS board of directors met and received a presentation from members ofmanagement on strategic considerations relating to the PBM business generally and the strategic rationaleand benefits of a business combination with Caremark, as well as potential strategic alternatives. In addition,Evercore provided a preliminary financial overview of a potential combination of CVS and Caremark for theCVS board of directors at this meeting. At the conclusion of the meeting, the CVS board authorizedmanagement to begin due diligence on Caremark and to engage in negotiations on a potential combination.

On September 25, 2006, the Caremark board of directors met telephonically and received an updatefrom Mr. Crawford on the status of discussions with CVS. The Caremark board of directors discussed thepotential business combination with CVS, other strategic growth opportunities and the industry generally.

On September 26, 2006, Mr. Ryan and Mr. Crawford met at the Hyatt Hotel in Dulles, Virginia todiscuss further the strategic rationale for, and benefits expected to be derived from, a business combinationbetween the parties. Mr. Ryan and Mr. Crawford also discussed possible governance structures for thecombined company and other management issues.

21

In September 2006, Lehman Brothers was requested to act as a financial advisor to CVS in connectionwith a potential strategic transaction with Caremark.

In late September 2006, JPMorgan was requested to act as a financial advisor to Caremark in connectionwith a potential strategic transaction with CVS.

On October 3, 2006, the Caremark board of directors met telephonically and reviewed with UBS andJPMorgan financial aspects of a possible business combination of CVS and Caremark. In addition, theCaremark board of directors discussed potential structures of a transaction with CVS, the potential businesscombination with CVS in the context of the industry and the potential benefits of such a transaction forCaremark stockholders. The Caremark board authorized Mr. Crawford and his management team to proceedwith a due diligence review of CVS and to continue to explore a possible business combination with CVS.

On October 5, 2006, the parties financial advisors met to discuss the exchange ratio for the potentialtransaction and the due diligence and transaction process to be followed. A subsequent meeting amongrepresentatives of CVS and representatives of Caremark and their respective legal and financial advisorsalso took place at the offices of CVS legal counsel, Davis Polk & Wardwell, in New York, New York. Theparties discussed the proposed due diligence process to be followed, regulatory matters raised by a businesscombination and the parties respective approaches with respect to these issues, the synergies analysis beingcompleted by the outside consulting firm, timing for negotiation and completion of a definitive agreement andprocess for moving forward with the parties consideration and negotiation of a possible business combinationbetween CVS and Caremark.

On October 6, 2006, CVS and Caremark executed a revised non-disclosure agreement. In addition, onOctober 6, 2006, each party commenced a due diligence review of the others operations. Members of CVSsenior management and members of Caremark senior management and their internal and external legal,accounting and financial advisors conducted certain due diligence reviews from an operational, financial,accounting, tax and legal perspective, including discussions with the other partys management. The bulk ofthe due diligence continued through October 20, with additional follow-up due diligence taking place betweenOctober 20 and October 31, 2006.

On October 9, 2006, representatives of CVS and representatives of Caremark and each of their regulatorylegal counsel, Mintz Levin Cohn Ferris Glovsky and Popeo, which is referred to as Mintz Levin, and JonesDay, respectively, met to discuss a joint communications strategy in connection with the proposed businesscombination.

On October 9, 2006, CVS legal counsel, Davis Polk & Wardwell, distributed an initial draft mergeragreement to Caremark and its advisors. During the month of October, negotiations on the merger agreementcontinued.

On October 9, 2006, Mr. Crawford met with David B. Snow, Jr., chief executive officer of Medco, todiscuss, among other things, the possibility of a business combination between the parties. Caremark andMedco subsequently entered into a confidentiality agreement with each other dated October 23, 2006 andon October 26, 2006, members of Caremarks management participated in a telephone call with membersof Medcos management. No confidential business information was exchanged between the parties, and noproposals for a potential business combination were made or discussed by either party, including in respectof price or other terms. No further discussions occurred between the parties.

On October 10, 2006, regulatory legal counsel to CVS and regulatory counsel to Caremark met with theoutside consulting firm to discuss the synergy analysis. In addition, on October 10, 2006, representatives ofCVS and representatives of Caremark and each of their respective legal counsel arranged a conference callto further discuss the overall due diligence process.

On October 12, 2006, representatives of CVS and representatives of Caremark and each of their respectivelegal counsel and financial advisors met at the New York City offices of King & Spalding, Caremarks counsel,to discuss current and pending litigation, regulatory and investigation matters facing the companies.

On October 18, 2006, representatives of CVS and representatives of Caremark and each of their respec-tive legal counsel and financial advisors met at the Hyatt Hotel in Dulles, Virginia for presentations bymanagement of each party about their respective businesses.

22

On October 20, 2006, the Caremark board of directors met telephonically during which Caremarks seniormanagement, legal counsel and financial advisors reviewed with the Caremark board of directors the status ofdiscussions with CVS regarding the possible transaction. At the meeting, Caremarks legal counsel reviewedwith the Caremark board of directors the draft merger agreement and the Caremark board of directorsdiscussed with Caremarks senior management, legal counsel and financial advisors the issues raised by thedraft agreement, matters identified to date by the due diligence review, corporate governance matters andrisks associated with the potential transaction and transaction timing. The Caremark board of directorsalso discussed synergies expected to be derived from a business combination between CVS and Caremark,the potential transaction in the context of the industry and possible reactions of the market.

On October 23, 2006, the CVS board of directors met to discuss various matters relating to a possiblebusiness combination of CVS and Caremark. Prior to this meeting, the board members had received avariety of background materials for their review. Mr. Ryan opened this meeting with a brief overview of thepossible transaction and the status of the negotiations. Thereafter, CVS financial advisors, Evercore andLehman Brothers, gave their respective financial overview of the transaction and other strategic alternatives.They also presented their respective views as to the strategic benefits and value-creating potential of thecombination. Mr. David B. Rickard, CVS chief financial officer, thereafter, presented on the financial duediligence conducted on Caremark up to that time. Thereafter, a Davis Polk & Wardwell partner summarizedand analyzed for the CVS board of directors the terms of the merger agreement as they stood at that timeand the material open issues that remained to be resolved in negotiations. He also reviewed the fiduciaryduties of the CVS board of directors in its consideration of the transaction, and reported on legal duediligence conducted on Caremark. Finally, a Mintz Levin partner and CVS chief legal officer, DouglasA. Sgarro, presented to the CVS board of directors on certain litigation and regulatory matters reviewedduring the course of legal due diligence on Caremark, as well as on the regulatory process applicable to thetransaction. At the conclusion of these various management and advisor presentations, the CVS board ofdirectors discussed the transaction in detail, including the risks (including any regulatory risks) and strategicbenefits and synergies expected to be derived from the transaction and the directors financial analysis of theproposed transaction. At the end of this discussion, the CVS board of directors authorized CVS to continueto negotiate and explore the possible transaction.

On October 24, 2006, representatives of CVS and representatives of Caremark and each of their respectivelegal counsel and financial advisors met at the Hyatt Hotel in Dulles, Virginia to discuss financial due diligencematters, preliminary results of the synergies analysis conducted by the outside consulting firm and a jointcommunications strategy in connection with the proposed transaction. The parties also further negotiatedopen issues in the merger agreement.

On October 24 and October 30, 2006, the parties financial advisors discussed Caremarks and CVS viewson the exchange ratio for the potential transaction.

On October 25, 2006, the Caremark board of directors met telephonically. At this meeting Caremarksmanagement and King & Spalding LLP, legal counsel to Caremark, updated the Caremark board of directorsregarding the status of the negotiation of the merger agreement. King & Spalding LLP and Caremarksmanagement discussed with the Caremark board of directors a number of provisions in the current draftof the merger agreement, including the definition of material adverse effect, the obligations of the partiesto obtain clearance for the merger from the Department of Justice and U.S. Federal Trade Commissionfrom an antitrust perspective, the circumstances under which Caremark would be permitted to enter intodiscussions and share non-public information with a third party, the Caremark board of directors abilityto recommend another transaction to the Caremark stockholders, each partys termination rights and thecircumstances under which a break up fee would be payable. The Caremark board of directors also discussedthe fact that these provisions were reciprocal and applied equally to CVS as well as to Caremark. King& Spalding LLP reviewed with the Caremark board of directors deal protection provisions contained inprecedent merger of equals transactions, and Caremarks management, legal counsel and financial advisorsalso discussed with the Caremark board of directors the amount of the proposed break up fee in the mergerand the amount of break up fees in selected precedent merger of equals transactions. After a thoroughdiscussion, the Caremark board of directors authorized Caremarks management to continue to negotiate the

23

merger agreement, including the deal protection provisions described above. The Caremark board of directorsalso discussed with Caremarks senior management, legal counsel (including antitrust counsel) and financialadvisors other issues raised by the revised draft agreement, the due diligence process, risks associated withthe potential transaction and a regulatory analysis of the transaction. The Caremark board of directors alsodiscussed the potential management team of the combined company and synergies expected to be derivedfrom a business combination between the parties. In addition, the Caremark board of directors receivedan update on the discussions of the parties financial advisors regarding the exchange ratio for the potentialtransaction.

On October 30, 2006, the Caremark board of directors met telephonically to evaluate the possible businesscombination with CVS. Prior to the meeting, the Caremark board of directors received various materials,including a draft of the merger agreement. At the meeting, Caremarks legal counsel reviewed with theCaremark board of directors its legal duties and responsibilities in connection with the possible transactionand reviewed the material terms and conditions of the merger agreement and open points in the mergeragreement that were still subject to negotiation. Caremarks senior management reviewed with the Care-mark board of directors the strategic benefits of the possible transaction, the results of the due diligencereview of CVS and the risks associated with the possible transaction. Caremarks financial advisors updatedthe board of directors on their discussions with CVS financial advisors regarding the exchange ratio. Athorough discussion took place among the members of the Caremark board of directors concerning the pos-sible transaction, including a discussion of the risks of the transaction, the expected strategic benefits ofthe transaction, possible regulatory considerations in connection with the transaction, synergies expected tobe derived from the business combination, the interests of directors and officers in the merger and financialaspects of the proposed transaction. At the conclusion of the meeting, the Caremark board of directorsauthorized management to continue negotiations with CVS to seek to resolve the remaining outstandingissues and to continue due diligence with respect to the potential transaction.

On October 31, 2006, representatives of CVS and representatives of Caremark and each of their respectivelegal counsel and financial advisors participated in a conference call to resolve the remaining open pointsin the merger agreement. Between approximately noon on October 31, 2006 and 7:30 a.m., Eastern Time,on November 1, 2006, senior management of CVS and senior management of Caremark and their respectivelegal and financial advisors finalized the merger agreement and other proposed definitive documentation.

At approximately 8:30 a.m., Eastern Time, on November 1, 2006, the Caremark board of directors metat JPMorgans offices in New York, New York to consider and act upon the proposed business combinationbetween CVS and Caremark. Prior to this meeting, the Caremark board of directors received variousmaterials, including a substantially final draft of the merger agreement. During this meeting, Caremarkslegal counsel reviewed with the Caremark board of directors the legal duties and responsibilities of theCaremark board of directors in connection with the proposed transaction and provided an update on thematerial terms and provisions of the merger agreement and changes to the merger agreement that had beennegotiated since the last meeting of the Caremark board of directors.

King & Spalding LLP reviewed with the Caremark board of directors the provisions of the draft mergeragreement and discussed, among other things, the definition of material adverse effect, the obligations ofthe parties to obtain clearance for the merger from the Department of Justice and U.S. Federal TradeCommission from an antitrust perspective, the Caremark board of directors fiduciary duties under Delawarelaw, the provisions of the merger agreement relating to the circumstances under which Caremark may engagein discussions and share non-public information with a third party, the Caremark board of directors abilityto recommend another transaction to the Caremark stockholders, each partys termination rights and thecircumstances under which a break up fee would be payable. King & Spalding LLP also reviewed withthe Caremark board of directors revisions made to the merger agreement to address points raised by theCaremark board of directors. In approving the negotiated deal protection provisions contained in the mergeragreement, the Caremark board of directors considered, in particular, the following:

• The fact that the deal protection provisions contained in the merger agreement regarding the cir-cumstances under which the Caremark board of directors could enter into discussions and sharenon-public information with a third party and change its recommendation permitted the Caremark

24

board of directors to properly exercise its fiduciary duties to Caremark stockholders under Delawarelaw.

• The resolution that led to the reduction in the amount of the break up fee and the circumstancesunder which the break up fee is payable;

• That revisions were made to the merger agreement in response to the Caremark board of directorsconcerns that Caremark could be responsible for paying the break up fee, but Caremark would notsimultaneously be able to enter into a binding agreement with a third party with respect to anacquisition proposal to cover the break up fee, including the revision that, even though Caremarkmust still submit the merger to the Caremark stockholders for a vote even if the Caremark boardof directors no longer supports the merger and changes its recommendation, Caremark is permittedto enter into a binding letter agreement with a third party making a superior proposal, and thisletter agreement may (1) provide that the third party is obligated, on behalf of Caremark, to payany termination fee payable by Caremark in connection with the merger agreement and (2) attachas an exhibit a fully negotiated and executed merger agreement relating to the superior proposal solong as the effectiveness of such agreement and plan of merger is conditioned upon the terminationof the merger agreement and the payment of the termination fee by Caremark to CVS in connectionwith the merger agreement;

• A survey of the deal protection provisions contained in precedent transactions of similar size andtype as the merger;

• That the deal protection provisions are reciprocal and are customary for strategic business transac-tions of this type; and

• That the deal protection provisions are the product of extensive negotiations between the parties.

Each of the foregoing factors, among others, was considered by the Caremark board of directors in approvingthe merger agreement.

Members of Caremarks senior management then updated the Caremark board of directors on the resultsof due diligence, the risks associated with the transaction and the strategic benefits of the transaction andrelated transaction matters. In addition, UBS and JPMorgan reviewed with the Caremark board of directorstheir joint financial analysis of the exchange ratio and each delivered to the Caremark board of directorsan oral opinion, each of which was confirmed by delivery of a written opinion dated November 1, 2006, tothe effect that, as of that date and based on and subject to various assumptions, matters considered andlimitations described in such opinion, the exchange ratio was fair, from a financial point of view, to holders ofCaremark common stock. Following a thorough discussion of the proposed transaction, the Caremark boardof directors unanimously voted to approve the merger and the merger agreement and authorized managementof Caremark to take certain actions designed to accomplish the transactions contemplated by the mergeragreement.

Following the approval of the merger and the merger agreement, Mr. Ryan was introduced to the Caremarkboard of directors.

At approximately 9:00 a.m., Eastern Time, on November 1, 2006, the CVS board of directors met tele-phonically or in person in New York, New York at the offices of Davis Polk & Wardwell. Before the CVSboard of directors convened, directors received a package of materials relating to their review of the proposedtransaction. Mr. Ryan opened this meeting by summarizing the then current status of deal negotiations anddevelopments since the board had last met. At the conclusion of Mr. Ryans summary, Evercore and LehmanBrothers each gave its own presentation as to the financial analyses performed by it relating to the merger.At the conclusion of each of their respective presentations, each of Evercore and Lehman Brothers renderedtheir respective opinions orally (each of which opinions was later followed up in writing) to the effect that, asof that date and based on and subject to the assumptions made, procedures followed, matters considered andqualifications and limitations described in each such opinion, the exchange ratio was fair, from a financialpoint of view, to CVS. After these opinions were rendered, a Davis Polk & Wardwell partner summarizedthe principal deal terms for the members of the CVS board of directors focusing, in particular, on changesto those terms since the CVS board of directors October 23 meeting including as to the resolution of thedeal protection provisions. Thereafter, senior CVS executives reviewed again for the members of the CVS

25

board of directors the strategic rationale for, and benefits expected to be derived from, the transaction. Adiscussion of the transaction followed this review. At the conclusion of this discussion, the CVS board ofdirectors unanimously approved the merger and the merger agreement, resolved to recommend these andrelated matters to CVS stockholders for their approval and authorized CVS management to take certainactions to bring the transaction negotiations to a successful conclusion.

Following the meetings of the CVS board of directors and the Caremark board of directors, CVS andCaremark executed the merger agreement and thereafter issued a joint press release on November 1, 2006announcing the transaction.

On December 18, 2006, Caremark received an unsolicited offer from Express Scripts, Inc., referred toas Express Scripts, pursuant to which Express Scripts offered to acquire all of the outstanding shares ofCaremark common stock for $29.25 in cash and 0.426 shares of Express Scripts common stock for eachshare of Caremark common stock. The receipt of Express Scripts common stock by Caremark stockholderswould be tax free, and the cash portion received by Caremark stockholders would be taxable to Caremarkstockholders. Express Scripts offer was made subject to, among other things, completion of a confirmatorydue diligence review of Caremark, termination of the merger agreement with CVS and negotiation andexecution of a merger agreement with Express Scripts. The cash portion of the consideration would befinanced by Express Scripts and Express Scripts stated that it had commitment letters from CitigroupCorporate and Investment Banking and Credit Suisse to provide the requisite financing to complete thetransaction. Caremark did not receive copies of these commitment letters from Express Scripts nor did itreceive a conditional merger agreement from Express Scripts.

On that same day, Caremark issued a press release announcing that the Caremark board of directorswould review the terms of the proposal submitted by Express Scripts in a manner consistent with its obliga-tions under the merger agreement with CVS and applicable Delaware law. CVS also issued a press releasereaffirming its support of the merger with Caremark and its confidence in the long-term strategic value ofthe combination as well as the benefit to the stockholders of both Caremark and CVS.

Also on that day, the Caremark board of directors met telephonically to discuss Express Scripts proposal.King & Spalding LLP, legal counsel to Caremark, gave a presentation to the Caremark board of directorsregarding the directors fiduciary duties under Delaware law and reviewed with the Caremark board ofdirectors the requirements of the merger agreement with CVS in connection with receipt and analysis ofExpress Scripts proposal. Caremarks financial advisors also participated in the meeting.

On December 20, 2006, the initial waiting period for the merger with CVS required by the HSR Actexpired without a Request for Additional Information (commonly referred to as a second request) from theU.S. Federal Trade Commission.

On December 22, 2006, the Caremark board of directors met in Nashville, Tennessee to discuss ExpressScripts proposal. During this meeting, King & Spalding LLP and Wachtell, Lipton, Rosen & Katz, specialcounsel, reviewed with the Caremark board of directors the legal standards under Delaware law and relatedlegal considerations applicable to the Caremark board of directors review of Express Scripts proposal, in-cluding a thorough analysis of the applicable requirements of the merger agreement with CVS. Jones Day,Caremarks antitrust counsel, reviewed possible antitrust implications of the Express Scripts proposal withthe Caremark board of directors. Caremarks financial advisors reviewed with the Caremark board of direc-tors financial aspects of Express Scripts proposal as compared to those of the merger with CVS. Managementof Caremark gave a presentation to the Caremark board of directors regarding strategic and business con-siderations of a possible business combination with Express Scripts as compared to the merger with CVS.The Caremark board of directors then met in an executive session.

On December 22, 2006, the CVS board of directors met telephonically. At the commencement of themeeting, Mr. Ryan provided a general update to the CVS board of directors regarding developments withrespect to Express Scripts. Davis Polk & Wardwell thereafter summarized Express Scripts proposal. Fol-lowing that summary, representatives of CVS financial advisors, Evercore and Lehman Brothers, discussedExpress Scripts proposal from a financial and market perspective. Davis Polk & Wardwell then summarizedpotential next steps. Thereafter, the CVS board of directors met in executive session.

26

On December 23, 2006, Mr. Crawford telephoned Mr. Ryan to discuss CVS possible response to ExpressScripts proposal. Mr. Ryan stated that CVS will have an upcoming board meeting and indicated to Mr.Crawford that CVS remains fully committed to the pending merger with Caremark and will move forwardpromptly to present the pending merger to the respective stockholders of Caremark and CVS for theirrespective approvals as contemplated under the merger agreement with CVS.

On December 27, 2006, the Caremark board of directors met telephonically to discuss Express Scriptsproposal further. King & Spalding LLP also participated in the meeting. The Caremark board of directorsdetermined to reconvene on January 5, 2007 for further review and discussion of the proposal.

On January 2, 2007, Caremark retained Banc of America Securities LLC to act as an additional financialadvisor to Caremark.

On January 5, 2007, Morgan Stanley & Co. Incorporated was retained to act as a financial advisor toCVS in connection with the merger.

On January 5, 2007, the Caremark board of directors met in Nashville, Tennessee to further reviewthe Express Scripts proposal. After thoroughly reviewing the Express Scripts proposal as set forth in itsDecember 18th letter and after consulting with its outside legal and financial advisors and upon considerationof a variety of factors, the Caremark board of directors unanimously determined that the Express Scriptsproposal does not constitute, and is not reasonably likely to lead to, a Superior Proposal as defined inthe merger agreement with CVS. As a result, the Caremark board of directors unanimously concludedthat, consistent with its fiduciary duties under Delaware law, the Express Scripts proposal does not meet therequisite standards established under the merger agreement for permitting Caremark to engage in discussionswith Express Scripts and that engaging in discussions with Express Scripts is not in the best interests ofCaremark and its stockholders. The Caremark board of directors has therefore determined not to enterinto discussions with Express Scripts or to provide Express Scripts with access to confidential informationconcerning Caremark consistent with the board of directors obligations under the merger agreement withCVS.

The Caremark board of directors considered a number of factors in connection with its evaluation ofExpress Scripts proposal, as described in more detail below. In view of the number of factors and complexityof these matters, the Caremark board of directors did not find it practicable to, nor did it attempt to,quantify, rank or otherwise assign relative weight to the specific factors it considered.

Lacks Strategic RationaleThere is no logical or compelling strategic rationale for a combination of Express Scripts and Caremark.

Simply creating a larger pharmacy benefits manager does not address evolving market dynamics, includingan increasingly consumer-centric healthcare environment, and greater demand for access to information,personalized pharmaceutical and disease management services, and the ability to better manage costs.

In contrast to the merger with CVS, Express Scripts proposal would not provide Caremark with anyunique services, programs or tools that could serve to address emerging healthcare trends, improve clinicaloutcomes, or help consumers manage costs and improve their health.

Creates the Risk of Significant Customer Attrition and Destruction of Stockholder ValueExpress Scripts proposal creates substantial business risks and potential near- and long-term destruction

of stockholder value resulting from the likely loss of existing and prospective customers. Given that theCaremark stockholders would own 57

Caremark believes that uncertainties created by pursuing Express Scripts proposal would disrupt andadversely impact at a minimum the 2007 selling season, which extends through the first three quarters ofthe year. Caremark also believes that the uncertainties presented by the Express Scripts proposal wouldvirtually eliminate new business opportunities and make it difficult to renew existing contracts.

Express Scripts recently reported financial results have raised questions about its ability to effectivelyintegrate recent acquisitions. Additionally, all of Express Scripts past acquisitions have been significantlysmaller than those completed by Caremark and CVS, and Express Scripts has no experience integrating abusiness the size of Caremark. In contrast, Caremark and CVS have an established track record of successfullyintegrating large-scale acquisitions, and these strong management teams are committed to remaining in placeto integrate and manage the combined company.

27

Intended to Derail the Strategic and Compelling CVS/Caremark Merger and Is a Defensive Move

by Express ScriptsUnlike Caremarks and CVS decision to enter into a merger from positions of strength and provide plan

sponsors and consumers with the products and services they desire, Express Scripts proposal is reactionaryand defensive. Caremark believes that Express Scripts interference with the CVS/Caremark merger re-flects Express Scripts concerns about the enhanced competition due to the innovative services created by aCVS/Caremark combination and the change in industry landscape that would result. Several of Caremarksclients have notified Caremark that Express Scripts has contacted them in an effort to disrupt Caremarksrelationship with the client. Also, Express Scripts has lost approximately $1 billion in net business to Care-mark over the last three years. Clearly, Express Scripts fears a greater loss in business when facing an evenstronger combined CVS/Caremark.

Questionable Assumptions Regarding the Calculation of SynergiesCaremark has substantial doubts regarding the reliability of the synergy estimates espoused by Express

Scripts, based on Caremarks own knowledge and experience in developing synergy estimates. Express Scriptssynergy analysis also fails to address the substantial potential negative synergies which could arise from acombination of Express Scripts and Caremark. In addition to the potential loss of business, factors such asthe two companies vastly different formularies, contracting practices, and adjudication engines and systemscreate both integration risk and potential for additional negative synergies.

Faces Significant, if Not Insurmountable, Antitrust Risks and Timing DelaysExpress Scripts proposal carries significant antitrust risk that could substantially delay closing, could

prevent closing altogether, or could result in the imposition of conditions that could adversely impact thebusiness, projected synergies, Express Scripts ability to obtain financing for such a transaction, and theterms of such financing. Potential remedies that could be sought by antitrust regulators would be difficultto execute in the pharmacy benefits management business given its service orientation and could negativelyaffect any potential benefits projected by Express Scripts.

Caremarks merger with CVS has received antitrust clearance without a second request and is expectedto close in the first quarter of 2007, while Express Scripts has just begun the antitrust approval processand Express Scripts does not expect to close its proposed transaction until the third quarter of 2007 at theearliest.

Highly Leveraged and Weakened Business with Diminished Financial Strength and FlexibilityExpress Scripts proposal would create one of the most leveraged public companies in the healthcare

services industry. With Express Scripts free cash flow dedicated to debt reduction for several years, thecombined company would be severely restricted in its ability to invest in its business, pursue other strategicopportunities, react to changing market dynamics, or engage in value creating financial transactions beneficialto stockholders such as share repurchases and dividends. The concern over leverage associated with ExpressScripts proposal is further evidenced by S&Ps recent decision to place Express Scripts on negative watch,a measure that is in contrast to the positive watch that CVS received following the merger announcementwith Caremark. Negative synergies associated with lost customer accounts or antitrust remedies would onlyfurther increase these risks and concerns.

A combined CVS/Caremark will have significant free cash flow, an investment grade credit rating, andconsiderable borrowing power, giving it substantial financial flexibility to invest in its business and pursueopportunities to enhance stockholder value, including through a continuation of dividend payments andpotential future share buybacks.

Other FactorsThe Caremark board of directors also considered certain other factors relating to the merger agreement

with CVS and the Express Scripts proposal including:Express Scripts did not disclose its financing and any conditions to the financing that may exist, such as

a material adverse effect condition.Express Scripts has expressly conditioned its willingness to enter into any binding agreement with Care-

mark on the prior termination of the merger agreement with CVS, as opposed to offering to execute aconditional merger agreement, as permitted by Section 8.07(g) of the merger agreement with CVS. Because

28

there is no conditional merger agreement, which is typical in competing merger proposals, Express Scriptshas failed to offer to advance the break-up fee that could become payable to CVS. The Caremark boardcannot envision any scenario where it would be willing to trigger the imposition of a $675 million break upfee without having a competing party obligated to fund that payment.

On January 5, 2007, following the Caremark board of directors meeting, Mr. Crawford called Mr. Ryanto advise him of the Caremark board of directors determination.

29

IV. CVS-Caremark News Articles

Example of News Articles and Press Releases During Merger NegotiationsBelow are a random sample of 50 media articles from the 1,931 articles for the merger of CVS andCaremark, announced on November 1st, 2006. The date merger negotiations began was November 20th,2005 and the deal closed on March 22nd, 2007. The pre-negotiation period began May 2nd, 2005. Amonth after the deal was announced, a secondary bidder, Express Scripts, entered and was rebuked byCaremark.

Media Source Date Article Title

Business Wire 20050602 CVS Corporation Holds 2005 Analyst/Investor Meeting; Company Raises EarningsGuidance

The Boston Globe 20050607 CVS AGREES TO PAY $110M TO SETTLE SUIT ; COMPLAIN ALLEGESSECURITIES VIOLATION

Associated Press Newswires 20050702 CVS files proposed $3 million settlement over 401K lawsuitThe Wall Street Journal 20050922 Volatility Rises on Walgreen, CVSM2 Presswire 20051104 Breaking Market News for Traders! November 4, 2005The Wall Street Journal Europe 20051122 CVS fans see room for gains — Drug retailer might be set to narrow valuation gap

with bigger rival WalgreenBusiness Wire 20060120 Zacks Brokerage Buy List: Apache Corporation, CVS Corporation, PepsiCo and

Wells Fargo & CompanyChicago Sun-Times 20060121 Biz BriefsFinancial Times (FT.Com) 20060123 FT.com site : Albertson’s agrees $9.8bn sale to Supervalu group.Associated Press Newswires 20060302 CVS same-store sales up in FebruaryReuters News 20060316 UPDATE 3-Miller to run some of Albertsons, leave Wild OatsReuters News 20060413 TEXT-Moody’s on CVS CorporationReuters News 20060511 UPDATE 2-Sovereign to put 484 ATMs in NYC-area CVS storesAssociated Press Newswires 20060525 S&P announces index changesBusiness Wire 20060601 CVS Corporation Reports Record May Sales of $3.2 Billion, up 10.7%Associated Press Newswires 20060601 CVS May same-store sales up 8.4 percentMarket Wire 20060620 ERUG Continues Growth: Announces the Visit of Its 25,000th Patient! June 20, 2006Associated Press Newswires 20060710 Sector Snap: CVS sales lift drug storesReuters News 20060710 UPDATE 2-CVS June same-store sales rise 8.4 percentBusiness Wire 20060717 Paula A. Price Joins CVS as Senior Vice President, Controller and Chief Accounting

OfficerPittsburgh Post-Gazette 20060731 SNOWBIRDS TAKE THEIR JOBS WITH THEMDow Jones News Service 20060803 UPDATE: CVS Logs 2Q Profit Gain, Issues Upbeat OutlookDow Jones International News 20060824 Rite Aid To Acquire Eckerd, Brooks DrugstoresReuters News 20060905 Walgreen, CVS August sales top expectationsDow Jones News Service 20061006 NYSE Issues, 4 pm Net Change Percentage Gainers & LosersDow Jones News Service 20061101 UPDATE:CVS 3Q Earnings Rise On Strong Sales, MarginsDow Jones News Service 20061101 UPDATE: CVS To Buy Caremark In $21B ‘Merger Of Equals’Dow Jones Capital Markets Report 20061101 CVS Bonds Rally As No Debt Will Be Involved In MergerPR Newswire (U.S.) 20061101 CVS and Caremark Confirm DiscussionsBusiness Wire 20061101 Fitch Places CVS & Caremark Rx on Rating Watch Positive After Merger

AnnouncementM2 Presswire 20061121 Shareholder Alert for Walgreen Co.Reuters News 20061130 Rite Aid sets vote on Jean Coutu deal for Jan. 18Dow Jones News Service 20061218 UPDATE: IN THE MONEY: Express Scripts Bid Seems FamiliarPR Newswire (U.S.) 20061218 CVS Corporation Statement Released TodayDow Jones News Service 20061220 Update Regarding Merger Of Caremark Rx, Inc. And CVS CorpAssociated Press Newswires 20061221 Express Scripts reiterates offer for Caremark after CVS’ offer gets antitrust clearanceAssociated Press Newswires 20061222 Judge refuses to force Caremark to review Express Scripts bidPR Newswire (U.S.) 20070104 Express Scripts Sends Letter to Caremark StockholdersReuters News 20070108 UPDATE 8-Caremark favors CVS pact, rebuffs Express ScriptsReuters News 20070108 Express Scripts ‘committed’ to Caremark combinationDow Jones News Service 20070116 Express Scripts Announces Exchange Offer For CaremarkDow Jones News Service 20070213 Big Caremark Holders Ask Court To Delay Vote On CVS DealDow Jones News Service 20070213 UPDATE: Caremark Holders Ask To Delay Vote On CVS DealDow Jones News Service 20070223 Delaware Court Won’t Block Vote On CVS Offer For CaremarkThe Wall Street Journal 20070309 CVS Makes ‘Final’ Bid To Secure CaremarkBusiness Wire 20070313 CVS Comments on ISS Recommendation in Support of CVS/Caremark Merger and

Express Scripts’ AnnouncementDow Jones News Service 20070315 4th UPDATE: CVS Shareholders Vote To Approve Caremark MergerDow Jones Chinese Financial Wire 20070320 DJ CVS-Caremark Deal Seen Closing Late This Week,Early Next WeekBusiness Wire 20070322 CVS/Caremark Merger Closes, Creating the Nation’s Leading Pharmacy Services

CompanyAssociated Press Newswires 20070322 CVS completes $26.5 billion acquisition of Caremark