CEOs, abandoned acquisitions, and the media

Abstract

Do the media play a role in corporate governance and, if so, how? Those questions are broad and their answers have broad implications. This is especially so in countries such as the U.S. that are characterized by a free and vigorous business press. By corporate governance, we mean the traditional role of corporate governance in monitoring corporate management to ensure that top managers act in shareholders’ interests. So the questions are whether this active media coverage plays a role in guiding corporate managers to act in shareholders’ interests and, if so, how do they do it.

Academic studies have proposed that the media can play such a role by influencing the value of top managers’ reputational capital.1 In this framework, a manager's reputational capital is viewed as the present value of his future wages and employment opportunities.2 The media are said to affect such values by reporting on managers’ actions and by shaping perceptions of those actions. And to the extent that they influence managers’ reputational capital, the media can play a role in guiding managers’ actions. Whether they do so—and whether they do so in ways that are in shareholders’ interests—are open questions. We address those questions in one specific set of circumstances: namely, when would-be acquirers are considering whether to carry out or abandon acquisition attempts that the market perceives as “value-reducing.”

Several prior studies have reported that would-be acquirers are significantly more likely to abandon takeover attempts when the market responds to the announcement of the proposed acquisition with a downward revision of the potential acquirer's stock price. A common interpretation of this finding is that “managers listen to the market.” But this begs the question: why do managers listen to the market?

Our answer to that question is that acquirers’ top managers—their CEOs—have two sets of skin in the game. First, and perhaps obviously, the CEO owns stock in the acquiring company. Call this his financial capital. To the extent that cancellation of a proposed “value-reducing” takeover results in recovery of the announcement period stock price decline and the CEO owns shares in the company, the CEO stands to gain from that price recovery.

Second, we propose that the CEO stands to gain from the recovery of his personal reputational capital that may also have been diminished as a result of the market's perception that the announced takeover attempt is value-destroying. The media influence the CEO's reputational capital by interpreting and disseminating information about the CEO's acquisition decisions. The worse the tone of the media coverage and the broader its dissemination, the greater the negative impact on the CEO's reputational capital. To the extent that the CEO's reputational capital has been diminished by media coverage of the takeover attempt, abandonment of that attempt may be associated with a recovery of that loss.

There are at least three testable predictions that follow from this reasoning. First, the greater the dollar amount of the CEO's ownership of the shares of the acquirer, the more likely the CEO is to abandon a value-reducing acquisition attempt. Second, the broader the media coverage of the takeover attempt, the more likely the takeover is to be abandoned. Third, the broader the media coverage of the takeover attempt in combination with a more negative tone, the more likely the takeover attempt is to be abandoned. It is this last prediction that lies at the heart of our proposals. It is this interaction of the breadth and tone of media coverage that influences managerial reputational capital.

All of these predictions depend, of course, on the expectation that abandonment of the value-reducing acquisition attempt will be associated with a recovery of whatever financial and reputational capital are lost at the announcement of the value-reducing attempt and the associated media coverage.

In a study whose findings are reported in an article recently published in the Journal of Financial Economics, we tested these three predictions by analyzing 636 acquisition attempts by U.S. public companies during the period of 1990–2010. Each of the deals had a proposed transaction value of at least $100 million, and each experienced a negative stock market reaction at its initial announcement.

Consistent with our propositions, we find that the greater the dollar value of the loss experienced by the CEO through his ownership of the acquirer's shares, and therefore the greater his hoped-for recovery in the event of deal abandonment, the greater is the likelihood of abandonment. We further find that such acquisition attempts are more likely to be abandoned the more negative the tone of the media coverage of the proposed transaction and the broader that coverage.

A further feature that distinguishes our statistical analysis from prior studies of abandoned takeover attempts is that when we take account of the CEO's loss in share value and our proxy for his loss of reputational capital in our statistical models, the decline in share price alone is no longer a statistically significant predictor of the likelihood of acquisition abandonment. The implication is that potential acquirers abandon value-reducing acquisition attempts when it is in the managers’ interests to do so. That is, managers listen to the market when it is in their self-interest to do so. The further implication is that the media can and do play a role in corporate governance by influencing the value of managers’ human capital.

Keep in mind that each of these predictions is predicated on managers’ expectation that the reversal of their decisions to undertake the acquisition will be associated with at least some recovery of whatever monetary and reputational capital were lost at the announcement. Thus, to round out our analysis, we examined the stock prices of acquirers around the public announcements of cancellations of value-reducing takeover attempts. We find that announcements of cancellations of value-reducing acquisition attempts are accompanied by positive average stock price reactions. What's more, the price response tends to be more positive, the more negative the stock price response at the initial announcement. This evidence supports the proposition that managers can and do recover value, both monetarily and reputationally, when value-reducing acquisition attempts are abandoned.

A number of studies have reported finding that the likelihood of an acquisition attempt being abandoned is greater the more negative the stock price reaction of the potential acquirer at the initial announcement of the attempt.3 This evidence has been interpreted to mean that “managers listen to the market” when considering abandonment (or completion) of takeover opportunities.

But a more careful interpretation of this evidence is that managers sometimes listen to the market. After all, not all acquisition attempts that receive a negative market response are abandoned. Many studies have suggested that managers receive certain observable and unobservable benefits from acquisitions, many of which may be the result of managing a larger enterprise.4 If such benefits are significant, the question then becomes: why do managers ever abandon proposed acquisitions? We propose that they do so because they have both financial and reputational capital at stake. Thus, our working hypothesis is that managers listen to the market when the benefits of abandoning the acquisition outweigh the benefits of completing it.

More specifically, we propose that there are two main benefits of abandoning value-reducing acquisitions, and both relate to recovery of value lost (and the additional value that could be lost) if the deal goes forward.

First, to the extent that the CEO owns shares in his company, those shares will have declined in value when the stock price declined at the initial announcement. To the extent that abandonment of the attempt is associated with recovery of that stock price decline, the CEO can expect, or hope, to recover some or all of that value if the attempt is abandoned.

Second, to extent that the stock price declined at the announcement of the attempt, the CEO may also have lost reputational capital in the form of the labor market's perception of his abilities as a manager. If so, the CEO may also be able to recover some or all of that loss should the attempt be abandoned. But one of the main challenges in designing our study was how to measure the loss and, thereby, the potential recovery of that reputational capital.

That is where the media enter the picture. Media coverage, both in the breadth and the tone of the coverage, can be considered as a source of reputational value lost—and therefore potentially recovered.

As mentioned earlier, studies starting with Luigi Zingales (2000) have argued that the media can and do play a role in corporate governance. Dyck et al. (2008) built upon that observation by making the connection between corporate governance and the media both more specific and more general. They set forth a model in which top corporate managers have human capital at risk in making corporate decisions. Their human capital as managers is the present value of their future wages and employment opportunities as managers. These wages and opportunities lie in the realm of the managerial labor market. To the extent that the managerial labor market is informed by the media about corporate (i.e., a given CEO's) actions, media coverage has the potential to influence the CEO's actions by influencing the CEO's future opportunities in the managerial labor market. And in this sense, the media, by shaping perceptions of corporate events through the tone and extent of that coverage, can serve as a monitor of corporate activities and play a role in corporate governance.

With this thinking in mind, we used value-reducing acquisition attempts as the laboratory to study whether the media play a role in corporate governance.

To identify the acquisition attempts used in our analysis, we obtained a set of 636 proposed acquisitions by 537 publicly traded U.S. companies during the period of 1990–2010. Each proposed transaction had a value of at least $100 million and, at the announcement of the transaction, the acquiring firm's cumulative abnormal stock return (CAR) was negative. This announcement period CAR was measured as the sum of the differences between the acquiring firm's daily stock returns and the CRSP value-weighted market returns over the 3-day interval surrounding the announcement of the proposed acquisition. This sample and the announcement dates of the proposed transactions are from the Thomson Financial Securities Data Company's (SDC) US Mergers and Acquisitions database. To our knowledge, the set of acquisitions used in our analysis includes all takeover attempts that meet our criteria.

Panel A of Table 1 (column 2) gives the time series of the value-reducing acquisition attempts. By far the largest number of such attempts took place in 2000, when there were 67. The year in our sample with the fewest such attempts was 1991, when there were just seven. Abandoned attempts (as reported in column 3) also varied considerably over time, with none in 1991 and a high of 14 in 1996. In total, 19% of the takeover attempts with negative announcement period CARs were abandoned. Interestingly, the percentage of abandoned attempts is more evenly distributed across our sample period than is the number of such attempts. Nonetheless, none of the seven value-reducing acquisitions in 1991 and the 20 such attempts in 2009 was abandoned.

Panel B of Table 1 gives the industry classifications of the industries with the greatest representation of value-reducing acquisition attempts where industry is based on the potential acquirer's primary two-digit SIC code. In total, acquirers come from 51 different two-digit SIC industries. Not surprisingly, Business Services, which encompasses the largest number of firms in general, also had the greatest representation among value-reducing attempts, with 131. As shown in the last column of Panel B, however, the percentage of attempts that are abandoned is much more evenly distributed across industries than is the number of attempts.

The variables of primary interest to us are the following two: (1) the change in the dollar value of the CEO's shares upon announcement of the proposed transaction; and (2) the extent and tone of the media coverage given to the takeover attempt. For each abandoned acquisition attempt, we gathered CEO stock ownership from ExecuComp. We measured the change in the value of the CEO's ownership as the number of his shares times the acquirer's announcement period CAR.

We measured the breadth of the media coverage given to each acquisition attempt by counting the number of transaction-specific news stories that appeared in the Wall Street Journal (WSJ), the New York Times (NYT), and the Dow Jones News Service (DJNS) over the ten calendar days beginning with the announcement day of the proposed transaction. We labeled this variable media attention.

We measured the tone of media coverage given to each acquisition attempt by calculating the number of “negative” words divided by the number of total words in transaction-specific news stories.5 We called this variable media tone.

Table 2 sets forth some preliminary statistics regarding our key variables. Each of these simple comparisons is consistent with our propositions, but it will be the interactions of these variables, as reported later in Table 3, that provide a robust formal test. As shown in the first row of Table 2, and consistent with prior studies, both the mean and median announcement period CARs are more negative for the proposed transactions that are eventually abandoned (mean = −8.73%; median = −6.77%) than for those that are completed (mean = −7.55%; median = −5.72%). These results, as already noted, are consistent with the idea that managers listen to the market in making decisions to abandon (or complete) proposed acquisition attempts. The unanswered question, however, is why do managers continue with some value-reducing deals while abandoning others? That is the main question our study aims to answer.

The beginning of an answer can be seen in the later rows of Table 2. As reported in the second row, the dollar value of CEO ownership of shares is significantly greater in transactions that are abandoned (mean = $487.35 million; median = $10.74 million) than in those that are completed (mean = $98.89 million; median = $8.31 million). Further, and more consequentially, as shown in the third row, the change in the value of the CEO's shares upon announcement of the proposed transaction is also more negative in transactions that are abandoned (mean = −51.76 million; median = −$0.50 million) than in those that are completed (mean = −$5.86 million; median = −$0.42 million). These simple comparisons indicate that, ignoring other factors, the CEOs of acquiring firms are more likely to abandon proposed acquisitions when the proposed transaction means a significant loss in the value of their own financial capital.

The fourth and fifth rows of the table report our findings for the variables that are meant to capture the role of the media, if any, in this setting. As shown in the fourth row, media attention was far more pronounced in abandoned attempts (mean = 7.51; median = 5.00) than those that were completed (mean = 4.01; median = 3.00), and the differences between the two are highly statistically significant. Greater media coverage of value-reducing acquisition attempts is clearly associated with a greater likelihood of transaction abandonment. The simple comparisons of media tone point in the direction of the media playing a role in shaping perceptions, but, standing alone, they do not appear to be especially powerful. For abandoned attempts, the mean and median media tone are 5.28 and 4.99; for completed attempts the mean and median are 5.17 and 4.96.

The simple comparisons of Table 2 set the stage for more formal statistical consideration of our propositions. Recall that our propositions are that managers are more likely to abandon value-reducing acquisitions when the gains to the manager for abandonment are larger. The gains come in two forms: recovery of the CEO's financial capital, which we measure as the initial loss in the value of the CEO's shares when the deal is announced; and recovery of the CEO's reputational capital, which we try to capture with the interaction of the announcement period CAR, media attention, and media tone. That is, to the extent that abandonment of the deal recovers value that appears to have been lost because of extensive media coverage—especially coverage with a negative slant—the CEO has more to recover by canceling the transaction.

We consider these ideas in a step-wise series of tests whose findings are reported in Table 3. Table 3 shows the results of probit analyses where the dependent variable is a 1 or 0 depending upon whether the proposed transaction was abandoned or completed. Each of the estimations includes the variables of primary interest along with a set of control variables that are defined in the Appendix.

In the first model of the table (whose findings are reported in column 1 of Table 3), we include only the acquirer's announcement period CAR (along with the control variables). As reported in prior studies, the relation between the likelihood of deal abandonment and CAR is negative and highly statistically significant. Thus, the more negative the CAR, the greater the likelihood that the deal will be abandoned. These results are consistent with managers listening to the market. But why do they do so? The subsequent models attempt to address that question.

The model in column 2 includes the change in the dollar value of the CEO's shares along with the acquirer's announcement period CAR (and the control variables). The coefficient of the dollar value change in the CEO's shares is negative and statistically significant. Thus, as might be expected, CEOs are more likely to abandon proposed acquisitions the greater their financial loss when announcing the deal and, therefore, the greater their hoped-for recovery when abandoning the transaction. At the same time, however, the coefficient of CAR remains statistically significant—and this, in turn, suggests that the CEO's personal financial loss (and hoped-for recovery) is not the only factor that influences his decision to abandon (or complete) the acquisition.

The models in the remaining columns of the table address the question of whether one of those factors is the CEO's loss of (and hoped-for recovery of) his reputational capital. Model (3) includes the acquirer's CAR, the change in the dollar value of the CEO's shares, media attention, and media attention interacted with CAR (i.e., CAR × media attention). Model (4) includes the acquirer's CAR, the change in the dollar value of the CEO's shares, media tone, and media tone interacted with CAR (along with the control variables).

In both estimations, the coefficients of the change in the value of the CEO's shares are negative and statistically significant. Further, in model (3) the coefficient of the interaction of CAR with media attention and in model (4) the coefficient of the interaction of CAR with media tone are negative and statistically significant. In neither model is the coefficient of CAR alone statistically significant. In model (3) the coefficient of media attention alone is not statistically significant and in model (4) the coefficient of media tone alone is not statistically significant.

The importance of this set of results is fourfold. First, the insignificance of CAR alone indicates that it is not just the market reaction to the announcement of the proposed transaction that plays a role in the decision to abandon (or complete) the acquisition. Second, the insignificance of media attention and media tone alone imply that it is not just the level and slant of the media coverage that matter. Third, the significance of the interaction of media attention and media tone with the loss in share value (i.e., the level of the negative CAR) implies that it is the transformation of media attention and media tone into a monetary value that is of consequence in the decision to abandon the proposed acquisition. Fourth, the continued significance of the change in the dollar value of the CEO's shares indicates that the loss in the personal financial value to the CEO plays a role in the abandonment decision. These latter two results are consistent with our proposition that it is the possible recovery of financial and reputational capital that are lost at the initial announcement of the value-reducing acquisition that influences the decision to abandon (or complete) the proposed takeover.

The final model of Table 3 includes all of the variables in models (1)–(4) along with the interaction of media attention and media tone (i.e., media attention × media tone) and the triple interaction of CAR, media attention, and media tone (i.e., CAR × media attention × media tone). Arguably, it is this model that addresses our propositions in full. As shown in the last column of the table, the coefficient of CAR is not statistically significant, the coefficient of the change in the value of the CEO's shares is negative and statistically significant, the coefficient of the interaction of media attention and media tone is negative and statistically significant, and the coefficient of the 3-way interaction of CAR, media attention and media tone is negative and statistically significant. Simply stated, managers do not listen to the market just because the stock price declined. Rather, the data point to the manager paying attention when he has financial and reputational capital at stake.

Key to our propositions and our interpretation of the results is that managers expect to recover their losses in financial and reputational capital when value-reducing acquisition attempts are abandoned. To address the reasonability of that expectation, we examine acquirers’ CARs over the 3-day interval surrounding the “Withdrawn Date” (we label this the “Abandonment CAR”).

We classify the abandonments into three groups based on their initial announcement period CARs. Those with the most negative CARs are in the Low CAR group and those with the least negative CARs are in the High CAR group. We then calculate the average Abandonment CAR for each group.

For the entire set, the average Abandonment CAR was 2.71%. Thus, at least on average, abandonments of value-reducing acquisition attempts are good news for shareholders and are consistent with CEOs recovering financial and reputational value lost at the time of the initial announcement. More important for our purposes, Abandonment CARs increase in magnitude as we move across groups from the Low to High. For the Low CAR group, the Abandonment CAR is 5.71%; for the Middle CAR group, the Abandonment CAR is 1.65%; and for the High CAR group, it is 0.73%.

This evidence supports the presumption that the CEO of the acquiring firm can reasonably expect to recover at least part of his losses in stock ownership and reputational capital when abandoning the proposed value-reducing acquisition attempts.

In 2008, Microsoft launched a bid for Yahoo. On February 1, 2008, the announcement day of the proposed takeover, Microsoft shares fell by 8.4%. Within 7 days, the deal received 62 media reports with 9.02% of the words in the articles carrying a negative tone (in other value-reducing deals in our sample, the averages for these numbers are 4.7 and 5.12%). Steve Balmer, the CEO of Microsoft, owned $13.7 billion worth of Microsoft shares and options. Mr. Ballmer abandoned the deal on May 3, 2008.

In 2008, HP initiated an attempt to buy Electronic Data System. On the announcement day, May 13, 2008, the company's stock price declined by 8.5%. In contrast with the Microsoft-Yahoo transaction, the HP/EDS deal received only seven news articles within 7 days with 3.74% of the words carrying a negative tone. The CEO of HP, Mark Hurd, owned $26 million worth of HP shares and options. The deal was completed on May 12, 2008.

Both deals occurred in 2008 within the same sector. Also, the companies are of similar size and both experienced stock price declines of about −8.5% during the 3-day period around the announcement of the proposed takeover. However, there are striking differences with regard to CEO ownership, media attention, and the tone of media coverage. The attempt with more CEO ownership involved, greater media coverage and a more negative tone to that coverage was canceled.

Prior studies have shown that managers’ decisions to abandon proposed corporate acquisitions are negatively correlated with the stock market reaction to the announcement of the proposed transaction. One interpretation of this result is that managers “listen to the market” when considering whether to complete or abandon proposed acquisitions. The unanswered question is: why do managers listen to the market?

Part of the answer may be managers’ sense of obligation to their shareholders, along with possible concern about corporate control consequences—including the possibility of being subjected to a takeover offer themselves—of failing to maximize shareholder value. Another part of the answer is their own shareholdings in the firm, and the possibility of recovering the loss in the value of their own financial capital incurred at the announcement. But, in a recently published study whose findings are reviewed in this article, we propose that these are only part of the answer. We propose that managers also have reputational capital at risk, and that the reversal of the acquisition decision may also reverse any loss in that value associated with the announcement of the proposed transaction.

Using measures of the level and tone of media attention given to the proposed transaction together with the announcement period stock price reaction to such acquisitions (as a proxy for the change in the CEO's reputational capital), we find that this three-way interaction variable is significantly correlated with the likelihood that the proposed acquisitions will be abandoned. We interpret this to mean that managers are sensitive to the stock market reaction to proposed “value-reducing” acquisitions at least in part because of the impact on their reputational capital. In short, it appears to be CEOs’ concerns about the effects of value-reducing acquisitions on both their financial and reputational capital that induces them to listen to the market.