Beyond the target: M&A decisions and rival ownership
Diversified acquirer shareholders can profit from value-destroying acquisitions not only through their target stakes, but also through stakes in non-merging rival firms. Announcement losses are largely mitigated for the average acquirer shareholder when accounting for wealth effects on their rival stakes. Ownership by acquirer shareholders in non-merging rivals is negatively associated with deal quality and positively associated with deal completion. Funds with more rival ownership are more likely to vote in favor of the acquisition. Overall, these results show that many so-called “bad deals” are often in the interest of acquirer-firm shareholders.
Acquisitions are important corporate events that are often viewed as value destroying. The average returns to bidder firms that acquire other public corporations are negative around merger announcements, while average returns to target firms are positive. This finding has been interpreted as evidence of empire building, CEOs pursuing a personal agenda or CEOs’ overconfidence.1 Why do acquirer shareholders not stand up and fight against value-destroying acquisitions?2 Matvos and Ostrovsky (2008) (MO) provided a potential explanation to this puzzle by examining target ownership: since the acquirer’s institutional investors may hold shares in the target, the increase in value of the target may offset the losses on the acquirer side. This explanation was contested by Harford et al. (2011) (HJL), who argued that cross-ownership at the shareholder level was not large enough to compensate the acquirer shareholders in value-reducing acquisitions.
In this paper we aim to shed new light on this puzzle by considering the role of ownership in non-merging industry rival firms by acquirer shareholders. While the debate so far has focused on the wealth effects of the acquiring and target firms3, mergers generally have effects beyond them, impacting other rivals in the industry that are not involved directly in the acquisition. Extensive empirical evidence documents that, on average, there is a positive effect of takeover announcements on rival firm stock returns (Eckbo, 1983, Eckbo, 1985, Mitchell, Mulherin, 1996, Song, Walkling, 2000, Shahrur, 2005, Servaes, Tamayo, 2013). In addition, there is substantial ownership by acquiring shareholders in rival firms. As we will see, taking this ownership into account helps understand M&A decisions and the voting behaviour of acquirer shareholders.
When an acquirer firm conducts a horizontal merger, its non-merging industry rivals may gain for different reasons. There is a rich literature documenting the existence of both informational and real effects of merger announcements on rivals. Informational effects are changes in the valuation of rivals due to new information being revealed in the market without any change in the underlying fundamentals nor corporate policies of the firms. For example, rivals may gain because the merger increases the perceived probability that they could also become targets in the future (Song and Walkling, 2000), regardless of the form and outcome of the merger, or a reassessment of valuation of the firms in the sector.
Mergers also have real effects on rivals: rivals may gain due to improved efficiency at the expense of the merging firms (Eckbo, 1983, Shahrur, 2005, Clougherty, Duso, 2009, Bernile, Lyandres, 2019), increased market power due to a reduced number of industry participants (Eckbo, 1983, Eckbo, 1985, Sapienza, 2002, Fathollahi, Harford, Klasa, 2021), and more efficient corporate policies due to increased takeover threats in the industry (Servaes and Tamayo, 2013).4 Overall, these papers highlight the importance of real effects in explaining rivals’ stock price reactions to M&A horizontal deals.
We argue that diversified acquirer shareholders, holding broad portfolios of industry firms, can internalize not only target gains, but also gains in non-merging rival firms. This may lead to lower incentives to oppose deals with negative acquirer returns. In fact, many acquirer shareholders end up gaining from the merger, even when they lose from their position in the acquirer, supporting this internalization hypothesis.
Consider the following example: when Microsoft announced the $26.2 billion acquisition of LinkedIn in June 2016, the deal was perceived as value-destroying by the market and led to a loss of 1.46% for Microsoft shareholders in the 3-day window around the announcement. With a market capitalization of over $400 billion at the time, the losses for Microsoft’s largest shareholders were substantial, ranging from $72 million to $373 million, as shown in Fig. 1. Nine of these top ten shareholders also owned shares in the target. While LinkedIn did enjoy a large announcement gain of 45.97%, only two acquirer shareholders were able to offset its loss on Microsoft with a gain from LinkedIn.
However –and this is the main point of this paper– nine of Microsoft’s top ten institutional shareholders obtained a net gain thanks to the wealth effects from their ownership in rival firms. This was because, among Microsoft’s top twenty industry rivals, fifteen gained during the 3-day window around this announcement, and these gains were more than enough to compensate these shareholders for their losses from their holdings in Microsoft. Therefore, even though the deal may have seemed to be value-destroying, it actually created value for most of the top 10 shareholders of Microsoft.
In a sample of 1800 horizontal mergers among public firms from 1988 to 2016, we find that the pattern of acquirer holdings in non-merging rivals compensating for losses in the acquiring firm is fairly common. In particular, we show that the returns around deal announcement on rival stakes are on average positive –ranging from 0.14% to 0.28%, depending on the industry definition. Rival gains also continue beyond deal announcement with an average return of 0.84% from deal announcement to completion. The largest shareholders of the acquiring firm hold an average stake of 2.2% in each rival, which is double the average stake in the target (1.1%). In addition, for 76% of acquirer shareholders, a larger fraction of their industry portfolio is held in rivals than in the acquirer and target combined.
Furthermore, in “bad deals” nearly one third of the acquirer shareholders are winning shareholders, i.e., achieve a net gain when accounting for stakes in both the target and non-merging rivals. That is, these deals may not be “bad deals” after all, as they are not value-destroying for an important subset of acquirer shareholders.5
To understand whether winning shareholders have enough clout within firm to affect the outcome of the merger, we measure the aggregated voting power of winning shareholders. Of course, if we only take into account the acquirer CAR, none of the acquirer shareholders are winners in “bad deals”. If we take into account only holdings in the acquirer and the target, we find that in only 4% of the “bad deals” the majority of the votes are held by winning shareholders. However, if we take into account the acquirer, target, and rival holdings, we find that in 29% of “bad deals” the majority of the votes are held by winning shareholders. This suggests that taking into account rival holdings could be important to understand why in many cases the majority of shareholders do not oppose deals with negative acquirer CARs.
Rival ownership by acquirer shareholders can also have implications for expected acquirer returns around announcement. Suppose management is considering a “value-destroying” acquisition, in the sense that the acquirer return is negative. If most shareholders expect to gain from the deal through their target and rival holdings, management might conclude that shareholders are not likely to strongly oppose the deal, and this reasoning would make managers more likely to pursue the deal. Therefore, we would expect higher rival ownership by acquirer shareholders to be associated with lower acquirer returns.
We find that ownership in non-merging industry rivals is indeed negatively associated with acquirer CAR. This result is consistent with the hypothesis: shareholders care less about losses on their acquirer holdings because they are compensated by their holdings in non-merging rivals, and therefore, we should observe lower acquirer return when the latter are large.6
An alternative to the internalization hypothesis is that diversified shareholders may allow negative acquirer CAR deals purely because the negative returns are more diluted when they represent a smaller fraction of their portfolio. That is, diversified acquirer shareholders may not oppose “bad deals” simply because the large number of firms in their portfolio reduce their incentive to monitor. If this were the reason for the association between rival ownership and negative CARs, then we should also expect to see an association between deal quality and ownership in non-rival firms, beyond the industry. However, we find that acquirer shareholders’ stakes in non-rival firms in their portfolios have no predictive power for deal quality. This result suggests that our results are unlikely to be simply driven by a dilution effect.
Also supporting the internalization hypothesis, we find that ownership in rivals is positively associated with the probability of a “bad deal” being completed. If rivals jumped due to purely informational reasons, we should not expect to see any association between rival ownership and the probability of a deal being completed, since shareholders would not need the deal to go through to benefit from for the jump in the rivals. In a more direct test, we show that when there are more top ten shareholders that benefit from the deal at the industry portfolio level, a bad deal is more likely to be completed. Conversely, when there are more top ten shareholders that lose from the deal at the industry portfolio level, a good deal is less likely to be completed.
This link between investors’ portfolio returns and the probability of deal completion suggests that there is some form of engagement between investors and managers. Investors can vote against management in acquisition deals that require shareholder approval (Becht, Polo, Rossi, 2016, Li, Liu, Wu, 2018). Even absent formal voting requirements, acquirer shareholders with significant stakes can oppose the deals through the threat of exit or behind-the-scenes interventions (Chen, Harford, Li, 2007, McCahery, Sautner, Starks, 2016). While we obviously cannot observe the threat of exit or behind-the-scenes engagement for each deal, we do know that such engagement is common (McCahery et al., 2016). We can observe, however, the voting behavior of shareholders at the individual deal level, at least for a subset of the deals. We thus examine shareholder voting on acquisition decisions, and in particular whether voting behavior is different for shareholders with rival ownership. Only deals that are financed by issuing equity of more than 20% of the acquirer firm’s outstanding shares require shareholder approval, and therefore it is important to note that deals that require voting are a selected sample.7
We find that funds with more rival ownership are less likely to vote against an acquisition when the ISS recommends voting in favor, compared to funds with lower rival ownership. This finding is robust to the inclusion of deal and fund fixed effects, so that we are comparing voting behavior across funds within a deal, and controlling for the fact that some funds are more likely to vote against management in general. We also find that when a fund experiences a significant negative return at the industry portfolio level, it is more likely to vote against the deal even when ISS recommends voting for or when it is a seemingly good deal.
In summary, we find that acquirer shareholders often gain around the announcement of so-called “bad deals” once we account for the gains from their target and rival stakes. Thus, our findings provide a rational explanation for why acquirer shareholders fail to oppose a large number of deals that would seem to go against their interest if one narrowly focuses solely on the acquirer return. Our study provides evidence to the idea that diversified shareholders often take a portfolio value maximization view when evaluating an individual firm’s corporate actions, and thus sometimes may lack the incentive to pursue an active monitoring role against poor decisions at the firm level.
The remainder of this paper is organized as follows: Section 2 provides a descriptive analysis of shareholder returns around mergers; Section 3 analyses the association between acquirer shareholders’ ownership in rivals and deal quality, as well as probability of completion. It also tests the alternative explanations for our results, and the direct tests using realized returns; Section 4 presents the empirical analyses on voting actions by acquirer shareholders after accounting for their ownership in the target and industry rivals; Section 5 concludes.
Our sample includes all horizontal deals from 1988 to 2016 from SDC Thomson-Reuters. We keep a deal if the acquirer owns less than 50% of the target prior to the announcement and is seeking to own more than 50% of the target. For completed deals, we require the acquirer to own more than 90% of the target after deal completion. Mergers involving firms with multiple securities are dropped, as in MO. We define a horizontal M&A deal based on historical Compustat 3-digit SIC codes following
Relation between rival holdings and acquirer CARs
In this section, we examine the relation between holdings in target and rival firms and deal quality. Our hypothesis conjectures that when acquirer shareholders have large holdings in rival firms, they would have less incentives to deter managers from pursuing deals in which the acquiring firm’s return is expected to be negative, i.e. bad deals. This is because if diversified acquirer shareholders expect general gains from the acquisition by other industry firms they hold and evaluate the
Shareholder voting in M&A deals
Perhaps an even more direct way to test the internalization hypothesis is to examine the voting behavior of shareholders with different levels of target and rival ownership. If the internalization hypothesis were correct, we would expect to see that, within a deal, the shareholders with more target and rival ownership are less likely to vote against the deal than shareholders with less target and rival ownership. Focusing on vote against management or dissent is very common in the governance
The literature on M&A that focuses on negative acquirer returns has historically highlighted the conflicts of interest between shareholders and managers as the most likely explanation for why these deals get announced and completed. However, more recently, Matvos and Ostrovsky (2008) and Harford et al. (2011) have pointed out that there is also conflict of interest within the group of shareholders due to portfolio heterogeneity. In particular, some shareholders are highly diversified within the
Research data for this article
for download under the CC BY licence
Supplementary Data S1
Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/