1) Poison Puts: Corporate Governance Structure or Mechanism for Shifting Risk? Frederick L. Bereskin Assistant Professor Department of Finance Lerner College of Business & Economics University of Delaware, Newark, DE 19716 bereskin@udel.edu Helen Bowers Associate Professor Department of Finance Lerner College of Business & Economics University of Delaware, Newark, DE 19716 bowers@udel.edu September 8, 2015 Abstract A poison put, also referred to as an event-risk or change-in-control covenant, gives the bondholder the option of redeeming the bond before maturity if a specified event occurs, such as a change in control. We investigate the association of poison puts with corporate governance and their effects on firms’ merger activity. We do not find evidence that poison puts are associated with weaker governance. Rather, our evidence is most consistent with poison put issuance being driven by efficient contracting for bondholders, as opposed to managerial entrenchment. However, we also find that firms that issue poison puts are less likely to be either acquirers or targets, even though these firms are more likely to be in industries with acquisition activity. Other findings are consistent with an explanation based on efficient contracting – issues with poison puts exhibit other characteristics that we would also expect for bondholders who are concerned about potential risk-shifting – such as shorter maturity, higher treasury spreads, and significantly more covenants. 

2) 1 Introduction Poison puts (also referred to as proxy puts, change in control, or event risk covenants) give bondholders the right, but not the obligation, to redeem their bonds if a change in control occurs. These covenants give bondholders a put option requiring the issuer to offer to purchase all of the bonds (typically at 101% of par), when changes in control occur.1 Poison puts have traditionally been seen as the result of efficient contracting by bondholders to minimize their agency costs by protecting against risk-shifting from highly levered transactions or being acquired by firms with lower debt ratings. The effectiveness of covenants in this context has been studied by Asquith and Wizman (1990). They found that even though, on average, the bondholders of successful leveraged buyouts and restructurings over the period 1980-1988 experienced negative returns, bonds with relatively strong covenant protection gained value. Asquith and Wizman’s (1990) results also show that the covenants that were most effective in increasing bond value during this period functioned similarly to poison puts because the increase in bond value was related to those bonds being more likely to be called than unprotected bonds. The 2014 acquisition of Tim Hortons Inc. by Burger King Worldwide Inc. is a more recent example of the use of poison puts for efficient contracting by bondholders. Burger King’s debt was rated five levels below Tim Hortons’ investment grade debt. However, instead of experiencing a loss due to a ratings downgrade resulting from the acquisition, Tim Hortons’ 1 Specifically, changes in control are generally defined as one or more of the following events: 1) any person or group acquires 50% or more the issuer’s voting stock; 2) any time the majority of the board of directors ceases to be those who were directors at the time of issuance (continuing directors) or directors whose election was approved by a majority of the continuing directors (related to this are the notion of “dead hand” proxy puts; without the “dead hand” feature, board can approve the dissident slate of directors to avoid triggering the put – see Reindel (2015) for more detail); 3) the merger or consolidation of the company into another entity; 4) the sale, in one or a series of related transactions, of all or substantially all of the company’s assets; or 5) the adoption of a plan to dissolve or liquidate the company. Some poison put covenants include a “double trigger” where the change in control event must be associated with a ratings downgrade. 1 

3) bondholders were protected by poison puts that allowed them to redeem their issues at a slight premium to par.2 However, poison puts, like any anti-takeover defensive measure can also serve to entrench management. A recent New York Times article suggests that management is active in placing poison put covenants in indenture agreements for the purpose of entrenchment (Solomon, 2014). Poison puts are described as “tools companies use to forestall shareholder activists and hostile takeovers.” This argument places poison puts in the corporate governance arena as a mechanism of circumventing shareholder rights. Poison puts cause an additional concern because they are debt covenants and therefore, unlike poison pills, sidestep scrutiny as a takeover defense that could impinge on the shareholder franchise. Inclusion of a poison put in a debt covenant entails less conflict among shareholders than putting poison pill provisions in place – and are generally implemented without shareholder votes. Consequently, poison puts are an example of an “unregulable” takeover defense (Arlen and Talley, 2003) in that they can be imposed through contractual arrangements, and difficult to pinpoint as purely entrenching. Moreover, in contrast to poison pills, poison puts have the potential to significantly harm the firm, and not merely the acquirer. As a result of their potential for significant damage to the firm, their ability to be imposed without shareholder scrutiny, and their widespread and increasing use, it is important to understand their implications in greater detail. Decisions in two seminal cases in the Delaware Chancery Court, San Antonio Fire & Police Pension Fund v. Pharmaceuticals, Inc.3 and Kallick v. SandRidge Energy,4 dealt with C. Gutscher, “Tim Hortons poison put leaves creditors with indigestion”, BloombergBusiness, Aug 27, 2014, Retrieved http://www.bloomberg.com/news/articles/2014-08-27/tim-hortons-put-leaves-holders-with-indigestioncanada-credit on Mar 2, 2015. 3 San Antonio Fire and Police Pension Fund v. Amylin Pharms., Inc. 983 A.2d 304 (Del. Ch. 2009). 4 Kallick v. Sandridge Energy, Inc.68 A.3d 242 (Del. Ch. 2013). 2 2 

4) boards of directors’ responsibilities to protect stockholders from possible entrenchment effects around triggering events in poison put covenants. Both cases were decided on narrow grounds, but Byeff (2015) argues that because those poison puts that were triggered by the replacement of the majority of the company’s continuing directors, that poison puts can “have both the motive and effect of entrenching incumbent directors” and “that board entrenchment purposes had trumped potential agency cost reduction.”5 Although not endorsing the argument that poison puts exist solely to entrench management, the Delaware Chancery Court’s 2013 decision in Kallick recognizes that management can adopt them solely for that purpose. The Kallick decision states that, in light of the potential entrenching effects of poison puts, companies should strenuously bargain to exclude puts and accept them only in exchange for significant economic benefits and charges independent directors to monitor poison puts to ensure that they are not being adopted solely as entrenchment devices. Another avenue through which poison puts may entrench management is by discouraging activist investors. Reporting on bank loans, Hoffman (2015) states that “nearly 200 companies have struck new loan agreements since the beginning of 2014” that include poison put provisions. He argues that the increase in activist investors, particularly hedge funds, necessitates poison puts to protect lenders from the debt-financed buybacks, dividends, and restructurings, favored by such investors, which deteriorate credit. Poison puts are a unique means of entrenchment, because debtholders can argue that it is in their economic interest to know their borrowers and have faith that the business plan at the time of the issue will continue for the life of the loan. The increased focus given to poison puts in the courts and the popular press is understandable – despite their widespread attention in the early 1990s, their use has 5 See Byeff (2015) pp. 375 and 415. 3 

5) become significantly more common in recent years. For example, poison puts were included in over half of our sample debt issues in 2012, compared to a range of 8.1%-27.4% of debt issues in our sample in the 1990s. The purpose of our study is to investigate the association of poison puts with corporate governance and their effects on acquisition activity. We examine these two, non-mutually exclusive motivations for the inclusion of the poison put covenants. Poison puts could: (1) increase the agency costs associated with acquisitions, by potentially enhancing managerial entrenchment; or (2) protect bondholders from risk-shifting and associated wealth transfers that could result from acquisitions (i.e., efficient contracting). We do not find evidence that poison puts are associated with weaker governance.6 In particular, firms with bonds that have poison puts are more likely to have the characteristics of firms with strong corporate governance. Firms with bonds that have poison put covenants have slightly higher levels of institutional ownership, are less likely to have poison pills, and are no more or less likely to have classified boards. However, we also find that firms with bonds that have poison puts are more likely to be incorporated outside of Delaware7 and are less likely to be either acquirers or targets, even though these firms are more likely to be in industries with acquisition activity. Consistent with an explanation based on efficient contracting, issues with poison puts exhibit other characteristics that we would also expect for bondholders who are concerned about potential risk-shifting. In particular, issues with poison put covenants tend to have shorter maturities and higher treasury spreads (thus reflecting the risk), and significantly more covenants – both bondholder-protective covenants and issuer-restrictive. Our multivariate 6 7 We define governance as “weak” or “strong” from the shareholder’s perspective, in terms of shareholder rights. Daines (2001) notes that firms incorporated in Delaware are more vulnerable to takeovers. 4 

6) results confirm these findings, in that firms with more contracting-issues (both issue-specific and firm-specific) are more likely to include poison puts in their bonds. In contrast, we find no evidence of these firms having weaker corporate governance or being more prone to managerial entrenchment aside from entrenchment caused by the poison put itself. It is worth stressing that the additional covenants that are present among bonds with poison puts are particularly consistent with protection to bondholders from acquisition-related risk-shifting (as one would expect from the contracting-based explanation), as opposed to manager-specific and boardspecific restrictions on their investment activity and dividend policy. An alternative explanation for the generally strong governance of firms with poison puts is that these covenants could be one of the few means of entrenchment for these types of firms. Poison puts are not voted on by shareholders, and their effects are not widely understood, compared to other means of entrenchment. We then examine the effects of poison puts, in contrast to their motivation, and analyze our data in the context of acquisition activity. Poison puts appear to matter in that firms that had bonds with poison puts are less likely to be either acquirers or targets. Similarly, our regression results confirm that these firms are more likely to be in industries with acquisition activity. Together, these results suggest that bonds are more likely to contain poison puts if the issuing firms are more prone to acquisitions and in industries experiencing consolidation. Combined with our earlier results on governance and these firms’ contracting environments, our findings are consistent with these firms’ use of poison puts being driven by bondholders’ risk-shifting concerns. The effects of the firms experiencing fewer acquisition activity are also consistent with poison puts having an entrenchment effect. 5 

7) Nonetheless, we are unable to find consistent evidence that firms without poison puts experience more risk-shifting. This is not to say that poison puts are not effective at mitigating risk-shifting. One may expect that other firms mitigate risk-shifting and protect bondholders in other ways, whereas the type of firm that adopts poison puts does so out of limited alternative mechanisms. Our results have a number of interesting implications. First, we contribute to the literature by focusing on the governance implications and determinants of poison puts, which are not widely studied despite their widespread use and potential agency problems. Given our findings that firms issuing bonds with poison puts have reasonable shareholder governance and no significantly higher levels of managerial entrenchment (beyond the poison puts themselves), our results help explain why poison puts remain so prevalent. Moreover, their increased use may in part be driven by improved shareholder protection, to the degree that bondholders are more concerned about risk-shifting occurring. Our paper also presents evidence that firms are not opportunistically employing poison puts to position themselves as stronger acquirers. In contrast, poison puts are used by firms in consolidating industries that wish to avoid acquisition activity in general, which admittedly could also reflect an entrenchment effect. Poison puts are a manifestation of an environment in firms that have a unique contracting environment with bondholders, managers, and other parties. Finally, our lack of evidence of risk-shifting occurring is consistent with poison puts not being detrimental to bondholders. At the very least, they appear to offer a mechanism to protect bondholders who would otherwise not have credible forms of commitment. 6 

8) Our study also contributes to the broader literature of creditors’ role in corporate governance, and the mechanism by which their involvement occurs.8 Moreover, whereas other studies may focus on how particular covenants alleviate the bondholder-shareholder conflict, poison puts also highlight the shareholder-manager conflicts as well. Our paper continues as follows. In Section 2, we discuss the related literature, and build on it to develop our hypotheses. We review our data set and results in Section 3. Concluding remarks follow in Section 4. 2 2.1 Literature Review Pricing and value effects of poison puts on stockholders and bondholders The seminal work on poison put bonds dates to Crabbe (1991) and Cook and Easterwood (1994). Crabbe (1991) finds that over the period 1983-1988, the inclusion of event-risk protection reduced borrowing cost by approximately 20 to 30 basis points. Although his evidence is consistent with bondholders valuing these covenants, he also finds evidence that this effect declined with the reduction in corporate restructurings towards the end of his sample period.9 Cook and Easterwood (1994) compare the wealth effects on the existing debt and equity of firms issuing debt with poison put covenants with firms issuing debt without poison puts during 1988 and 1989. For issuers of poison put debt, stockholders experience an abnormal For example, recent studies have examined creditors’ role on the board of directors (Santos and Rumble, 2006), information intermediation (Ivashina et al., 2009), or focused on creditors’ role among firms experiencing financial distress (Baird and Rasmussen, 2006; Chava and Roberts, 2008; Ayotte and Morrison, 2009; Roberts and Sufi, 2009; Nini, Smith, and Sufi, 2012). 9 Crabbe (1991) restricts his sample to bond issues with “super poison put” covenants. In the more recent literature this type of put is usually described as having a “double trigger” where the change in control event must be associated with a ratings downgrade. 8 7 

9) return of negative 70 basis points, while existing bondholders experience a positive abnormal return of 52 basis points. There is no wealth reaction for stockholders and existing bondholders associated with non-poison put debt. Cook and Easterwood’s (1994) evidence is consistent with the “mutual interest” hypothesis that poison puts protect both managers and bondholders, at the expense of shareholders. Torabzadeh, Roufagalas and Woodruff (2000) find, in a sample spanning 1986 through 1990, that the inclusion of a poison put reduced yields by 58 to 79 basis points. Using a case study approach, Okamoto, Pedersen, and Pedersen (2011) exploit court decisions regarding the buyout of Bell Canada Enterprises to show that poison puts are, indeed, priced by the market. 2.2 Returns for change of control events and risk-shifting Billett, Jiang, and Lie (2010) study the effect of poison puts on leveraged buyouts over the period 1980-2006. In particular, they show that upon the announcement of a leveraged buyout, bondholders with poison puts experience average wealth effects of 2.3%; this compares to -6.8% for bonds without poison puts. Additionally, they find that poison puts reduce the probability of being targeted in both LBO and non-LBO takeovers. Their results are also consistent with Asquith and Wizman’s (1990) results concerning LBOs occurring between 1980 and 1988. They found that the buyout announcement return for bonds with strong covenant protection is 2.6%; this compares to announcement returns of -0.7% for bonds with weak protection and -5.2% for those with no protective covenants. Billett, King, and Mauer (2004) study acquisitions between 1979 and 1997; although the magnitude of their results is not as large as those found for leverage buyouts, they conclude that there is a coinsurance effect or risk-shifting associated with acquisitions. They find that returns 8 

10) for target bondholders are significantly higher when the target’s debt rating is below the acquirer’s, when the combination is expected to decrease leverage for the target, and when the target’s maturity is shorter. Overall, they find an abnormal return on announcement for target bondholders of 1.09%. However, for targets’ debt rated below investment grade, the announcement return the announcement abnormal return is 4.30%. The abnormal announcement return for acquirer bondholders is -0.17%, and there is no significant difference between investment grade and non-investment grade debt. Lehn and Poulsen (1991) conclude that the use of event-risk covenants, such as poison puts, was the response to the risk-shifting and consequent large losses experiences by target firms in LBOs. They found that alternative defenses against risk-shifting, such as hedging and the use of convertible debt, did not increase after the buy-outs and restructurings of the 1980s. 2.3 Firm characteristics of issuers with poison puts Nash, Netter, and Poulsen (2003) find that firms with high-investment opportunities were more likely to issue bonds with poison puts in 1989 than in 1996. Their results show that firms that have bond issues with poison put covenants tend to have lower interest coverage ratios, and that non-investment grade debt is not driven by firms’ investment opportunities – rather, it is largely driven by the likelihood of financial distress. Their evidence is consistent with bond contracts being carefully negotiated, depending on the firm’s and investors’ goals, as opposed to being “boilerplate” documents. The use of covenants (such as poison put covenants) is positively associated with growth opportunities, debt maturity, and leverage, consistent with firms using covenants to address potential conflicts between stockholders and bondholders over growth option exercise (Billett, 9 

11) King, and Mauer, 2007) or bondholder risk driven by strong shareholder control (Cremers, Nair, and Wei, 2007). In contrast, Chava, Kumar, and Warga (2010) note the role of managerial agency costs, and emphasize the degree that these agency issues can mitigate the bondholder-shareholder conflict, to the degree that managerial entrenchment can occasionally benefit bondholders at the expense of shareholders. Their evidence is consistent with the important role of agency issues; although entrenchment increases the use of covenants to control for inefficient investment, it is negatively associated with covenants on dividend payouts and accepting takeover offers. Qi, Roth, and Wald (2011) use international data to establish results similar to Chava et al. (2010); firms in countries with stronger shareholder control may have more severe shareholderbondholder conflicts, and thus prefer to use more covenants. There are other reasons that firms may wish to employ poison put covenants. For instance, Hege and Hennessy (2010) contribute to the literature examining why firms may wish to issue debt with poison puts. They present evidence that the use of poison put covenants, among other strategies, help an incumbent capture more of the surplus in the event of an acquisition. In summary, the aforementioned papers present evidence that the effect of entrenchment on the use of poison puts is unclear. 3 3.1 Data and Results Data and sample statistics Our data is collected from a number of sources. In particular, we use Mergent FISD as our primary source of data for poison puts. We complement this data with Compustat for 10 

12) financial and accounting data, Thomson for institutional ownership data, RiskMetrics for certain corporate governance measures, and SDC for acquisition data. Given the low frequency of poison put issues prior to the 1990s, we begin our sample in 1990 and continue to 2012. In Panel A of Table 1, we provide the distribution of poison puts by the percentage of bond issues and firm-years (of firms that issue bonds and with matching Compustat data). This table reflects the large frequency of poison puts among our sample firms, whether by bond-issuance activity or by firm-year. Our results highlight the relative resurgence in the use of poison puts. For example, although their popularity diminished in the late 1990s (i.e., to 8.1% of issues in 1998 and 11.3% of issues in 1999), their use significantly rebounded in the later part of the last decade – by 2012, 56.8% of bond issues had poison put covenants. In unscaled terms, 2012 was a record year, with 239 poison put issues and 172 firm-years where at least one poison put was present in a bond. This is consistent with the use and prevalence of poison puts not being diminished, despite more vigilant corporate governance over that period that weakened many other aspects of entrenchment. The prominence of poison puts despite the more rigorous recent governance environment would be consistent with poison puts reflecting bondholders’ contracting demands, as opposed to a means of managerial entrenchment. Alternatively, it is also plausible that poison puts are one of the few mechanisms of entrenchment for otherwise well-governed firms – potentially due to shareholders’ limited ability to vote on these covenants, or due to less shareholder scrutiny. In Panel B of Table 1, we continue examining the characteristics of the sample poison put firms by evaluating their relative prominence by Fama-French 48 industry group. This table highlights the significant heterogeneity among firms’ industries, and reflects the importance of industry characteristics in determining whether a firm is inclined to issue a bond with a poison 11 

13) put. Not surprisingly, among firms that are in industries with already prominent anti-takeover protections – such as banking firms and utilities, due partly to government regulation – issuance of bonds with poison puts is rare; for example, only 9.5% of issues in the financial industry and 15.1% in the utility industry contain poison puts. In contrast, among firms in industries with less apparent anti-takeover protections (such as healthcare firms, construction firms, and business services firms), issues with poison puts account for the majority of bond issues and of sample firm-years; for example, poison puts are present in 76.3% of sample healthcare firms’ issues. The results in this table are consistent with firms issuing poison puts if they are relatively more vulnerable to a hostile acquisition, and not issuing these types of bonds when consolidation in the industry is less frequent. In Panel C, we examine the prevalence of poison puts by credit ratings. This table highlights the weaker credit quality of firms that issue poison puts. Whereas highly-rated firms seldom issue bonds with poison puts, their use becomes very common among firms with poorer credit quality. For example, of firms A-rated and higher, our sample contains 208 issues with poison puts, out of 2,723 total issues; in contrast, firms rated BB+, BB, or BB- issued 476 bonds with poison puts, out of 652 total issues. This finding potentially reflects some of the contracting regularities among lower-rated firms, and the notion that these firms’ creditors value the poison puts (and, similarly, that issuing poison puts is less damaging to shareholder value for these firms than for more highly-rated firms). In conclusion, our results indicate that poison puts are common in general, and remain prevalent in current bond issues. Firms appear more likely to issue poison puts if they are in industries commonly associated with acquisition activity, and are less likely if they have other safeguards in place to protect themselves and their bondholders from potentially risk-shifting 12 

14) acquisitions. Additionally, poison puts’ concentration among low-rated and non-investment grade firms is consistent with contracting-based explanations for their use, as opposed to governance-based explanations, since governance is not thought to be reliably associated with credit ratings. In Figure 1 and Table 2, we evaluate the percentage of firms’ issues represented by poison puts, compared to non-poison put bonds. This is especially relevant for us to understand whether firms’ poison put issuance reflects a more significant change in firms’ behavior, as opposed to the decision to issue a bond with a poison put being a one-time decision unrelated to the firm’s other bond issues. In Figure 1, we show that firms appear to establish a policy decision of always issuing bonds with poison puts (851 firms) or never issuing bonds with poison puts (847 firms). In contrast, there are a smaller number of firms (344 firms) with some, but not all, issues containing a poison put – for 112 of those firms, poison put issues account for between 40% and 60% of their bond issues. To examine whether those firms issuing only certain bonds with poison puts experienced a policy shift when they began issuing these types of bonds, we study their issuance activity in Panel B of Table 2. In this Panel, we show that there are a large number of issues that occur prior to the firm’s first poison put issue, and a smaller number of issues that occur following the firm’s last poison put issue. In contrast, during the time between the first and last issues with poison puts, 76.7% of firms’ issues included bonds with poison puts. This is consistent with the decision to issue bonds with poison puts as one that is deliberate and reflective of a general decision by the firm to increase its anti-takeover protections. Additionally, the relatively small number of issues that occur following the “last” poison put issue in our sample is consistent both with firms’ reluctance to reduce their takeover protections once those protections are already in place, 13 

15) as well as new bondholders’ reluctance to accept covenants less favorable than those of prior bondholders. Finally, the issuing activity that we note in this table supports the subsequent use in our paper of an indicator variable equal to one if the firm had issued bonds with poison puts in the prior five years – in general, use of poison puts in bonds tend to reflect a policy shift in firms’ covenants, as opposed to being temporary events that are not repeated. In Table 3, we evaluate the sample statistics of our firms, to compare firm-years and bond issues with poison puts to those without poison puts. In Panel A, we evaluate the differences in means and medians of firm-years with at least one poison-put issue to those without a poison-put issue (all firm-years are in our Mergent sample, and thus include firms that issued at least one bond, to make the comparisons similar). Our results suggest that firms issuing bonds with poison puts tend to be smaller (median sales of $1.7 billion, compared to $4.9 billion for the control sample), have a higher market-to-book (median of 1.4, compared to the control sample’s 1.2), and have more leverage (37%, compared to the control sample’s 31%). The differences in means and medians for these two groups’ values of sales, market-to-book, and leverage are statistically significant. The differences in these characteristics are consistent with the notion that firms with poison puts are more prone to information asymmetries, thus helping explain why they are more likely to use poison puts: their greater levels of information asymmetries would naturally increase the need for bondholders to receive contractual assurances of protection from a riskshifting or otherwise damaging takeover event. We also study the presence of four important governance variables: institutional ownership, an indicator variable for firms incorporated in Delaware, a dummy for firms with poison pills, and an indicator variable for firms with classified boards. Poison-put firms have significantly higher levels of institutional ownership (64%, compared to 59% for other firms). 14 

16) This is consistent with these firms having, if anything, stronger external monitors – again, an explanation that is consistent with a governance structure complemented by contractual protections for bondholders (given these firms’ information asymmetries), and inconsistent with weak governance and/or managerial entrenchment leading to the issuance of poison puts. Firmyears with poison puts are significantly less likely to be from firms incorporated in Delaware. Given the increased exposure of Delaware firms to acquisitions (Daines, 2001) and evidence that Delaware firms are not more or less likely to replace CEOs due to poor performance (Jagannathan and Pritchard, 2013), this finding is not consistent with firms issuing poison puts having weaker governance. Interestingly, poison put firms are significantly less likely to have a poison pill – it is present in 35% of those sample firms, compared to 45% among the control firms. This, too, ought to be expected, as poison put firms’ investors have a need to protect themselves from takeovers. Finally, we do not find any evidence that poison put firms are more likely to have classified boards; the similar prevalence of classified boards across the two groups of firms is again consistent with any weaker governance among poison put firms. In Panel B of Table 3, we evaluate characteristics of issues with poison puts, compared to other issues. Our results are consistent with these firms experiencing greater information asymmetries – the treasury spread is significantly higher for issues with poison puts (a median spread that is 83 basis points higher than issues that do not have poison puts) and maturity is significantly shorter than non-poison put issues (mean maturity is 11.0 years for poison put issues and 13.1 for other issues). The offering amount is not reliably different between the samples – poison put issues exhibit a significantly smaller mean and a significantly larger median, compared to other issues. 15 

17) The number of covenants in poison put issues is consistent with bondholders demanding stronger contractual protections. The mean and median number of bondholder protective covenants in the issues (excluding poison put covenants) is significantly larger for issues with poison puts than those without. Similarly, the difference in mean and median is significantly present in issuer restrictive covenants as well; the mean number of these covenants is 3.95 among issues with poison puts, compared to 2.40 among other issues. This evidence is consistent with poison puts being reflective of bondholders’ desires for optimal contracting, in contrast to them being reflective of managerial opportunism. Managers seeking to entrench themselves would be particularly reluctant to issue bonds that have significantly higher amounts of covenants (and especially issuer restrictive covenants). 3.2 Results In Table 4, we confirm these results with regressions on the determinants of poison puts. The dependent variable is equal to one if the issue has a poison put, and is zero otherwise. Our regression results are consistent with our univariate analysis. In particular, we find that the number of covenants is positively and significantly associated with the presence of poison puts. We also find that treasury spreads are positively associated with the presence of poison puts. Larger issues are positively associated with poison puts, consistent with the median of poison put issues being larger than other issues and in contrast to the mean of poison put issues being smaller. Shorter maturity is also positively and significantly associated with the presence of poison puts. In short, these issue-specific variables are consistent with our univariate tests. With the exception of market-to-book, the firm-specific variables are also consistent with our univariate tests. Lower levels of sales (i.e., firm size) are associated with a greater likelihood 16 

18) of a poison put issue. Additionally, leverage is positively and significantly associated with the presence of a poison put when controlling for year fixed effects, reflecting the distinct contracting environment for firms at risk of financial distress. In contrast, market-to-book is negatively associated with the presence of a poison put, suggesting that the firms with fewer growth options are more likely to acquiesce to limiting their vulnerability to being taken over. In Table 5, we continue from our earlier result by examining the characteristics associated with a firm-year having a poison put issue, conditional on the firm issuing bonds in that year. Similar to our results in Table 4, the likelihood of a firm-year experiencing the issuance of at least one bond issue with a poison put is greater in firms that are smaller, firms with more leverage, and firms with lower market-to-book. Our other variables are also as expected, and are inconsistent with poison puts being driven by poorly governed firms. In particular, we find that (1) institutional ownership is higher for these firms, reflecting stronger external monitoring coinciding with the use of poison puts; (2) an inconsistent effect of being incorporated in Delaware; (3) depending on the fixed-effects, poison pills are either insignificantly associated or negatively associated with the firm’s propensity to issue bonds with poison puts; and (4) the use of classified boards is not significantly associated with the use of poison puts. In conclusion, we do not find any evidence that the issuance of bonds with poison puts is driven by these firms having weaker governance. One way to study how the use of poison puts corresponds with other aspects of the firms’ contracting environment is to test what bond covenants are associated with poison puts. In Panel A of Table 6, we show the percentage of bonds with given covenants, conditional on them having poison puts. In this table, we find that the use of poison puts corresponds with other, related covenants – likely reflecting bondholders with legitimate concerns related to acquisitions 17 

19) and risk shifting. For example, the covenants most frequently seen in bonds with poison puts relate to consolidations/acquisitions, asset sales, and cross-acceleration. In Panel B of Table 6, we examine corresponding statistics of the proportion of poison put covenants present in bonds that have certain types of covenants. Again, our results are broadly consistent with poison puts being associated with protecting bondholders from the types of events that would occur in somewhat risky firms. For example, whereas poison puts would be common among bonds that have ratings decline triggers and asset sale clauses, they are less sensitive where bondholders’ concerns are less clear. As an example, only 22.2% of bonds with defeasance covenants contain poison puts. We continue by examining how poison puts are associated with subsequent acquisition activity in particular industries. In Panel A of Table 7, we compare acquisition activity of firms with poison puts with those that do not have poison puts. Consistently, we find that firms with poison puts are less inclined to be involved in acquisitions, either as acquirers or targets. For example, an average of 15% of sample poison put firm-years experience a year as an acquirer, compared to 21% for firms without poison puts. Similarly, 7% of firm-years where the firm issued poison puts are acquisition targets, compared to 10% of other firm-years.10 This is consistent with poison puts reflecting a desire by the firm to avoid being involved in acquisitions, as opposed to it using poison puts opportunistically to aggressively acquire firms while being a less likely target itself. Interestingly, we also find that industries less prone to acquisition activity (either as acquirers or as targets) are also more likely to have firms with poison puts. In particular, for poison put firms, the industry-average percentage of acquisitions that occur are 6% 10 Our results are very similar when omitting withdrawn deals. 18 

20) of acquisitions and 3% of targets, values that are significantly lower than 8% of firm-years and 4% of targets for non-poison put firms. Given the earlier results, we also examine how acquisitions of poison put targets compare to other firm-years. Our results in Panel B of Table 7 present mixed evidence regarding potential risk-shifting occurring at acquisitions in which the target has a poison put.11 First, these transactions are not significantly more likely to be hostile. Although this is not particularly surprising given the paucity of hostile transactions, fewer hostile transactions would suggest that poison puts work to deflect hostile takeovers. The transaction value is significantly larger for firms with poison puts, with a median value of $72 billion compared to $53 billion for other sample firms; admittedly, this partly reflects our inclusion of typically smaller firms without public debt issues in our sample of non-poison put targets. The median market-to-book of poison put targets is not significantly different than targets without poison puts, although the mean value is significantly less, at 1.58, compared to 1.93 for non-poison put firms. The market-to-book of the acquiring firm is not significantly different across subsamples. This is not consistent with risk-shifting occurring. We find broadly similar results for other variables related to risk-shifting. In particular, although leverage is significantly higher for firms with poison puts (for example, the median book and market value of leverage for these firms is 40% and 35%, respectively), the leverage of acquirers is lower than that of the targets – for example, the median market value of leverage for these firms is 20%, compared to 19% for non-poison put targets. This difference is significant at the 10% level, and the difference in mean values of leverage are insignificantly different between Our sample size here is larger since we are also including non-Mergent firms in the sample as “non-poison put targets.” Our results are robust to limiting both samples to firms that are in Mergent. 11 19 

21) the two subsamples. Although the mean and median of the book value of leverage of acquirers of poison put firms (32% and 30%, respectively) are significantly higher than the corresponding values for non-poison put firms (23% and 21%, respectively), the differences are not as surprising when considering that they are comparable to the differences in book leverage between the target firms. Moreover, the acquiring firm’s leverage is consistently below that of the target’s for poison put firms – something that is inconsistent with risk-shifting occurring. Finally, we consider earnings volatility. We find that poison put firm targets have significantly lower earnings volatility compared to non-poison put firm targets (i.e., a mean value of 3% compared to 9%), as do their acquirers (i.e., a mean value of 2%, compared to 10%). Additionally, among targets with poison puts, the earnings volatility of the acquirers is less than that of the targets (i.e., 2% compared to 3%), whereas this is not the case with non-poison put targets (i.e., 10% compared to 9%). If anything, risk-shifting in favor of poison put target bondholders seems to be occurring. This is consistent with the types of firms that issue poison puts being particularly aware of the potential damage associated with risk-shifting, and inclined to protect their bondholders more than otherwise comparable firms not issuing poison puts. In Table 8, we regress the recent issuance of poison put bonds against other firm and industry-specific characteristics. Firms in industries that are more acquisition-intensive (based on the acquisition activity in the prior year) are more likely to issue bonds with poison puts. In economic significance, a one-standard deviation increase in the industry’s acquisition intensity as targets (acquirers) is associated with a 14.5% (17.5%) increase in the likelihood of the firm issuing poison puts. This is consistent with these firms endeavoring simply to be less involved in acquisition activities, and their issuance of poison puts is likely driven by being in a consolidating industry; it is inconsistent with the sample statistics, and reflects the effect of the 20 

22) year and industry fixed-effects. Additionally, as before, leverage and institutional ownership are both positive and significant, and firm size is negative and significant. For our other variables, market-to-book is less significant than before, likely having much of its effect subsumed by our variable for the intensity of consolidation in the industry-year. Additionally, earnings volatility is insignificant, indicating that, when controlling for all other factors, firms particularly concerned about risk-shifting occurring are not more likely to issue poison puts. In Table 9, we examine the risk-shifting hypothesis in more detail, and find no consistent evidence of risk-shifting occurring. The aspects that we consider for risk-shifting are the acquirer’s market-to-book, leverage, and earnings volatility. Model (5) for leverage and Model (7) for earnings volatility are the only regressions where the coefficient associated with poison puts is significant. In these regressions, we find that the leverage or earnings volatility of the acquirer are slightly lower when the target has a poison put – in contrast to what would be expected from risk-shifting. However, these results are not robust to including year fixed-effects. 4. Conclusion In this study, we examine the prevalence and determinants of firms’ use of poison put covenants. Given the potential for these covenants to entrench management, and their continued popularity (particularly in certain industries), it is important to understand how they interact with other governance mechanisms that firms face. Our evidence is consistent with poison put covenants being reflective of efficient contracting. We do not find evidence of firms that issue these covenants experiencing weaker governance, and present evidence that poison puts are driven by risk-shifting concerns. We also 21 

23) find that firms with poison puts are less likely to be involved in acquisitions as either acquirers or targets, even though these firms are in industries experiencing consolidation. Our evidence is consistent with poison puts being employed in a manner consistent with firms’ contracting environments, although their effect on mergers would increase entrenchment as well. The increased prominence of poison put covenants and potential scrutiny from regulators, combined with the potential for poison puts to facilitate managerial entrenchment, shows the importance of understanding their role. Our results provide an example of a corporate finance decision that seems to be primarily driven by contracting considerations, as opposed to governance failures. 22 

24) References Arlen, J. and E. Talley, 2003. Unregulable defenses and the perils of shareholder choice. University of Pennsylvania Law Review 152(2), 957-1023. Armstrong, C.S., K. Balakrishnan, and D. Cohen, 2012. Corporate governance and the information environment: Evidence from state antitakeover laws. Journal of Accounting and Economics 53(1&2), 185-204. Asquith, P. and T. Wizman, 1990. Event risk, covenants, and bondholder returns in leveraged buyouts. Journal of Financial Economics 27(1), 195-213. Ayotte, K.M. and E.R. Morrison, 2009. Creditor control and conflict in chapter 11. Journal of Legal Analysis 1(2), 511-551. Baird, D.G. and R.K. Rasmussen, 2006. Private debt and the missing lever of corporate governance. University of Pennsylvania Law Review 154(5), 1209-1251. Billett, M.T., Z. Jiang, and E. Lie, 2010. The effect of change-in-control covenants on takeovers: Evidence from leveraged buyouts. Journal of Accounting and Economics 16(1), 1-15. Billett, M.T., T.-H. King, D.C. Mauer, 2004. Bondholder wealth effects in mergers and acquisitions: New evidence from the 1980s and 1990s. Journal of Finance 59(1), 107135. Byeff, S., 2015. Note: The spirit of Blasius: Sandridge as an antidote to the poison put. Columbia Law Review, 115, 375-415. http://www.lexisnexis.com/hottopics/lnacademic/ Chava, S., P. Kumar, A. Warga, 2010. Managerial agency and bond covenants. Review of Financial Studies 23(3), 1120-1148. Chava, S. and M.R. Roberts, 2008. How does financing impact covenants? The role of debt covenants. Journal of Finance 63(5), 2085-2121. Chemmanur, T.J. and K. Simoyan, 2010. What drives the issuance of putable convertibles: Riskshifting, asymmetric information, or taxes? Financial Management 39(3), 1027-1067. Cheng, A.F. and T.-H. King, 2006. An empirical examination of poison puts in U.S. corporate debt. Working paper, UNC Charlotte. Cook, D.O. and J.C. Easterwood, 1994. Poison put bonds: An analysis of their economic role. Journal of Finance 49(5), 1905-1920. Corporate Bonds: Control Freaks, 16 Mar 2006. The Economist. 21 Feb 2015. http://www.economist.com/node/5637646. Crabbe, L., 1991. Event risk: An analysis of losses to bondholders and “super poison put” bond covenants. Journal of Finance 46(2), 689-706. Cremers, K.J.M., V.B. Nair, and C. Wei, 2007. Governance mechanisms and bond prices. Review of Financial Studies 20(5), 1359-1388. Daines, R., 2001. Does Delaware law improve firm value? Journal of Financial Economics 62(3), 525-558. 23 

25) De Franco, G. and Vasvari, F. P. and Vyas, D. and Wittenberg-Moerman, R., 2013. Sticky covenants. Chicago Booth Research Paper No. 13-61; Rotman School of Management Working Paper No. 2288723. Available at SSRN: http://ssrn.com/abstract=2288723. Fabozzi, F. J. and S. Mann, 2012. The Handbook of Fixed Income Securities 8th ed., (New York: McGraw Hill). Feldhütter, P., E. Hotchkiss, and O. KarakaÅŸ, 2015. The value of creditor control in corporate bonds. Journal of Financial Economics, forthcoming. Francis, B.B., I. Hasan, K. John, and M. Waisman, 2010. The effect of state antitakeover laws on the firm’s bondholders. Journal of Financial Economics 96(1), 127-154. Gutscher, C. 27 Aug 2014. Tim Hortons poison put leaves creditors with indigestion. BloombergBusiness. 2 Mar 2015. http://www.bloomberg.com/news/articles/2014-0827/tim-hortons-put-leaves-holders-with-indigestion-canada-credit. Hege, U. and C. Hennessy, 2010. Acquisition values and optimal financial (in)flexibility. Review of Financial Studies 23(7), 2866-2899. Hoffman, L., 28 Apr 2015, Lenders feel heat as suits ramp up. Wall Street Journal. 29 Apr 2015. http://www.wsj.com/articles/banks-feel-the-heat-from-lawsuits-1430259260. Ivashina, V., V.B. Nair, A. Saunders, N. Massoud, and R. Stover, 2009. Bank debt and corporate governance. Review of Financial Studies 22(1), 41-77. Jagannathan, M., and A.C. Pritchard, 2013. Do Delaware CEOs get fired? Working paper, Binghamton University and University of Michigan. Kallick v. Sandridge Energy, Inc.68 A. 3d 242. Del. Ch. 7 Mar 8 2013. LexisNexis Academic. Web. 24 Jun 2015. Kahan, M. and M. Klausner, 1993. Antitakeover provisions in bonds: Bondholder protection or management entrenchment? UCLA Law Review 40, 931-981. http://www.lexisnexis.com/hottopics/lnacademic/ Katz, D. A., 15 Mar 2013. Delaware court raises bar for use of “poison put” provisions. Harvard Law School Forum on Corporate Governance and Financial Regulation. 2 Mar 2015. http://corpgov.law.harvard.edu/2013/03/15/delaware-court-raises-bar-for-use-of-poisonput-provisions/. Lehn, K. and A. Poulsen, 1991. Contractual resolution of bondholder-stockholder conflicts in leveraged buyouts. Journal of Law and Economics, 34(2-2), 645-673. Levine, M., 17 Dec 2014. Oil company finds gusher in bond indenture. Bloomberg View. 21 Feb 2015. http://www.bloombergview.com/articles/2014-12-16/oil-company-finds-gusher-inbond-indenture. Mansi, S.A., W. F. Maxwell, and J. K. Wald, 2009. Creditor protection laws and the cost of debt. Journal of Law and Economics 52(4), 701-717. McGee, P. and M. Wirz, 3 Feb 2013.New worry for bondholders: LBOs. The Wall Street Journal. 21 Feb 15. http://www.wsj.com/articles/SB10001424127887324761004578281660084115812. 24 

26) Nash, R.C., J.M. Netter, and A.B. Poulsen, 2003. Determinants of contractual relations between shareholders and bondholders: Investment opportunities and restrictive covenants. Journal of Corporate Finance 9(2), 201-232. Nini, G., D.C. Smith, and A. Sufi, 2012. Creditor control rights, corporate governance, and firm value. Review of Financial Studies 25(6), 1713-1761. Okamoto, K., D. Pedersen, and N. Pedersen, 2011. The price effects of event-risk protection: The results from a natural experiment. Journal of Empirical Legal Studies 8(3), 878-903. Petitt, B., J. E. Pinto, and W. L. Pirie, 2015. Fixed Income Analysis 3rd ed., (Hoboken, NJ: Wiley). Qi, Y., L. Roth, and J. K. Wald, 2011. How legal environments affect the use of bond covenants. Journal of International Business Studies 42(2), 235-262. Qi, Y. and J. K. Wald, 2008. State laws and debt covenants. Journal of Law and Economics 51(1), 179-207. Reindel, F.W., 2015. “Dead hand proxy puts” – what you need to know. Harvard Law School Corporate Governance Forum. June 10. Roberts, M.R. and A. Sufi, 2009. Control rights and capital structure: An empirical investigation. Journal of Finance 64(4), 1657-1695. San Antonio Fire and Police Pension Fund v. Amylin Pharms., Inc. 983 A. 2d 304. Del. Ch. 12 May 2009. LexisNexis Academic. Web. 12 May 2009. Santos, J.A.C. and A.S. Rumble, 2006. The American keiretsu and universal banks: Investing, voting and sitting on nonfinancials’ corporate boards. Journal of Financial Economics 80, 419-454. Solomon, S. D., 26 Nov 2014. A defense against hostile takeovers develops a downside. New York Times. 29 Nov 2014. http://dealbook.nytimes.com/2014/11/25/a-defense-againsthostile-takeovers-develops-a-downside/?_r=0. Torabzadeh, K., J. Roufagalas, and C. Woodruff, 2000. Self-selection and the effects of poison put/call covenants on the reoffering yields of corporate bonds. International Review of Economics and Finance 9(2): 139-156. Wargo, A. and I. Welch, 1993. Bondholder losses in leveraged buyouts. Review of Financial Studies 6(4), 959-982. 25 

27) Figure 1: Histogram of percentage of firms’ issues containing poison puts For each sample firm, this figure provides the proportion of the firm’s total issues that have poison put covenants. Number of Firms 900 851 847 800 700 Axis Title 600 500 400 300 164 200 100 71 79 30 0 Percent of Issues with Poison Puts 26 

28) Table 1: Sample statistics This table provides the number and proportion of poison put issues that occur, by various firm characteristics. Panel A provides the distribution by firm-year, Panel B provides the distribution by Fama-French 48 industry group, and Panel C provides the distribution by the bond’s S&P credit rating (where available). Panel A: Distribution by year % Poison Put Firm-Years Number of Issues in Sample Number of Firms in Sample 25.4% 27.9% 189 122 47 15.3% 22.1% 359 213 89 19.9% 27.6% 558 322 176 151 25.4% 38.8% 692 389 1994 77 73 27.4% 38.2% 281 191 1995 40 65 10.7% 24.9% 375 261 1996 49 109 16.7% 38.8% 293 281 1997 31 70 9.9% 27.8% 312 252 1998 26 63 8.1% 23.8% 320 265 1999 27 46 11.3% 22.0% 239 209 2000 16 55 9.9% 33.1% 162 166 2001 23 66 8.5% 28.1% 271 235 2002 18 42 7.0% 19.9% 258 211 2003 31 109 11.5% 37.8% 270 288 2004 20 73 11.5% 39.9% 174 183 2005 16 29 9.1% 19.1% 175 152 2006 35 58 16.5% 32.4% 212 179 2007 155 135 56.2% 61.6% 276 219 2008 108 97 47.0% 58.4% 230 166 2009 195 175 50.1% 60.3% 389 290 2010 205 164 53.8% 61.2% 381 268 2011 176 128 50.4% 59.3% 349 216 2012 239 172 56.8% 57.3% 421 300 Year Number of Poison Put Issues Number of Poison Put firm-years % Poison Put Issues 1990 48 34 1991 55 1992 111 1993 27 

29) Panel B: Distribution by industry-group Industry Group Coal Fabricated Products Real Estate Agriculture Healthcare Non-Metallic and Industrial Metal Mining Measuring and Control Equipment Construction Precious Metals Electrical Equipment Rubber and Plastic Products Steel Works Etc Recreation Apparel Restaraunts, Hotels, Motels Entertainment Construction Materials Wholesale Business Services Textiles Communication Business Supplies Retail Petroleum and Natural Gas Tobacco Products Electronic Equipment Chemicals Almost Nothing Personal Services Pharmaceutical Products Food Products Medical Equipment Shipbuilding, Railroad Equipment Consumer Goods Machinery Shipping Containers Transportation Computers Aircraft Printing and Publishing Insurance Candy & Soda Automobiles and Trucks Defense Utilities Beer & Liquor Trading Banking Number of Poison Put Issues 9 3 11 9 58 22 30 74 8 23 18 35 12 9 38 29 37 51 69 6 123 69 176 163 16 45 81 8 6 59 65 21 3 29 40 11 64 33 20 11 86 10 14 3 91 11 39 29 Number of Poison Put firm-years 10 6 12 7 67 15 24 81 8 17 14 39 12 9 43 40 41 66 114 6 159 48 150 179 7 95 58 13 12 85 51 33 4 27 47 11 71 41 14 9 66 6 20 4 82 9 70 48 % Poison Put Issues 100.0% 100.0% 91.7% 90.0% 76.3% 73.3% 68.2% 66.7% 66.7% 53.5% 51.4% 50.0% 48.0% 45.0% 43.7% 43.3% 42.5% 40.5% 40.1% 40.0% 37.7% 36.3% 35.7% 34.8% 34.8% 34.6% 34.5% 33.3% 33.3% 33.0% 32.8% 31.3% 30.0% 29.6% 28.8% 28.2% 26.8% 25.6% 20.0% 19.0% 18.2% 16.9% 15.7% 13.0% 10.9% 10.5% 8.8% 3.2% 28 % Poison Put Firm-Years 100.0% 85.7% 85.7% 87.5% 77.9% 68.2% 64.9% 73.6% 50.0% 53.1% 51.9% 56.5% 57.1% 45.0% 55.8% 61.5% 56.9% 50.0% 62.0% 46.2% 56.4% 37.2% 48.9% 48.8% 38.9% 63.3% 37.4% 50.0% 63.2% 51.8% 39.8% 49.3% 57.1% 36.0% 40.5% 40.7% 40.8% 45.6% 26.9% 21.4% 17.1% 17.6% 28.6% 40.0% 15.1% 18.4% 19.4% 9.5% Number of Issues in Sample 9 3 12 10 76 30 44 111 12 43 35 70 25 20 87 67 87 126 172 15 326 190 493 468 46 130 235 24 18 179 198 67 10 98 139 39 239 129 100 58 473 59 89 23 838 105 443 916 Number of Firms in Sample 10 7 14 8 86 22 37 110 16 32 27 69 21 20 77 65 72 132 184 13 282 129 307 367 18 150 155 26 19 164 128 67 7 75 116 27 174 90 52 42 387 34 70 10 543 49 361 507 

30) Panel C: Distribution by credit rating Credit rating Number of Poison Put Issues AAA - AA+ - AA AA- Number of Poison Put firm-years % Poison Put Issues Number of Issues in Sample Number of Firms in Sample 0.0% 0.0% 100 65 1 0.0% 4.0% 33 25 10 7 3.3% 4.1% 299 169 12 8 3.0% 3.7% 399 216 A+ 76 45 10.3% 11.4% 737 396 A 110 83 9.5% 12.7% 1,155 654 A- 166 94 22.9% 20.5% 726 459 BBB+ 202 156 24.6% 28.3% 821 551 BBB 262 207 26.4% 29.6% 993 699 BBB- 225 199 37.6% 40.2% 599 495 BB+ 123 128 58.0% 63.7% 212 201 BB 179 194 71.0% 78.5% 252 247 BB- 174 235 92.6% 89.0% 188 264 B+ 173 220 94.0% 92.1% 184 239 B 59 94 93.7% 91.3% 63 103 B- 18 60 100.0% 89.6% 18 67 CCC+ 6 17 85.7% 89.5% 7 19 CCC 2 6 100.0% 66.7% 2 9 CCC- 1 1 50.0% 50.0% 2 2 CC 1 2 100.0% 66.7% 1 3 1 0.0% 100.0% - 1 C - - % Poison Put FirmYears 29 

31) Table 2: Proportion of poison put issuance across sample This table examines the proportion of issues that contain poison put covenants. Panel A, which corresponds to Figure 1, provides the distribution by firm, for our sample period. Panel B provides the distribution of issues for the subsample of firms that issued at least one bond with a poison put, and for which less than 100% of its bonds contain poison puts. Panel A: Proportion of sample firms’ issues containing poison put covenants % of issues with a poison put x=0% 0< x ≤5% 5%< x ≤10% 10< x ≤15% 15%< x ≤20% 20< x ≤25% 25%< x ≤30% 30< x ≤35% 35%< x ≤40% 40< x ≤45% 45%< x ≤50% 50< x ≤55% 55%< x ≤60% 60< x ≤65% 65%< x ≤70% 70< x ≤75% 75%< x ≤80% 80< x ≤85% 85%< x ≤90% 90< x ≤95% 95%< x <100% x =100% Number of firms 847 5 9 14 24 19 13 33 21 15 82 3 12 7 29 28 8 8 11 3 0 851 30 

32) Panel B: Proportion of sample firms’ issues containing poison put covenants Number of poison put issues (1): Prior to first poison put issue (2): During the time that the firm issues its first poison put issue (3): After final poison put issue % of poison put issues (dollar weighted) 1,224 Number of issues % of poison put issues 76.7% 83.4% 917 1,595 389 31 

33) Table 3: Differences in means and medians; issue and issuer characteristics This table examines the differences in means and medians of firm-years where the firm issues at least one issue with a poison put covenant in the firm-year (Panel A), or issue characteristics of issues with poison put covenants (Panel B). Variables are defined as follows: Revenue (Firm’s sales in that year, in millions), Market to Book (the market value of assets, scaled by the book value of assets), Leverage (book value of total debt, scaled by the book value of assets), Institutional Ownership (the proportion of the firm owned by institutional investors), Delaware (an indicator variable equal to one if the firm is incorporated in Delaware, and zero otherwise), Poison Pill (an indicator variable equal to one if the firm has a poison pill, and zero otherwise), Classified Board (an indicator variable equal to one if the firm has a classified board, and zero otherwise), Treasury Spread (the treasury spread of the bond issue, in basis points), Offering Amount (the par value of the debt originally issued, in thousands), Maturity (the maturity of the issue, in years), Number of BP Covenants (the number of bondholder protective covenants in the same issue, excluding the poison put), and Number of IR Covenants (the number of issuer restrictive covenants in the same issue). *, **, and *** denote significant differences in means and medians at the 10%, 5%, and 1% levels respectively. Panel A: Firm-year characteristics Poison put firm-year N Mean Median Revenue 2,050 6,413 1,674 Market to Book 2,050 1.67 1.37 Leverage 2,050 0.41 0.37 Institutional Ownership 2,050 0.55 0.64 Delaware 831 0.42 0.0 Poison Pill 831 0.35 0.0 Classified Board 831 0.51 1.0 Non-poison put firm-year N Mean Median 3,328 13,548*** 4,882*** 3,328 1.48*** 1.19*** 3,328 0.31*** 0.31*** 3,328 0.54 0.59*** 1,259 0.46** 0.0** 1,259 0.45*** 0.0*** 1,259 0.52 1.0 Panel B: Issue characteristics Poison put bond N Mean Median Treasury Spread 1,877 216.62 183.00 Offering Amount 1,877 409,892 300,000 Maturty 1,877 10.99 10.02 Number of BP Covenants 1,877 3.46 4.00 Number of IR Covenants 1,877 3.95 3.00 32 Non-poison put bond N Mean Median 5,309 123.20*** 100.00*** 5,309 415,592 250,000*** 5,309 13.13*** 10.02 5,309 2.54*** 3.00*** 5,309 2.40*** 3.00*** 

34) Table 4: Determinants of poison put covenants in an issue This table presents the results of logistic regressions, in which the dependent variable is equal to one if the issue contains a poison put covenant, and zero otherwise. Explanatory variables include: LogBP (the natural log of one plus the number of bondholder protective covenants, excluding poison puts), LogIR (the natural log of one plus the number of issuer restrictive covenants), TreasurySpread (the treasury spread of the bond issue), LogOffAmount (the natural log of one plus the offering amount), LogMaturity (the natural log of one plus the number of days to maturity), LogRevenue (the natural log of one plus the firm’s sales), MarketToBook (the market to book value of assets), Leverage (total debt scaled by total assets). Year and FamaFrench 12 industry fixed effects are included, where applicable. Firm-clustered standard errors are provided in parentheses; *, **, and *** denote significant differences in means and medians at the 10%, 5%, and 1% levels respectively. LogBP LogIR TreasurySpread LogOffAmount LogMaturity LogRevenue MarketToBook Leverage Constant Observations Pseudo R2 Year fixed-effects Industry fixed-effects (1) 0.477*** (4.570) 1.209*** (10.983) 0.005*** (18.730) 0.012*** (15.368) -0.148*** (-2.793) 0.001 (0.030) -0.277*** (-4.854) -0.154 (-0.652) -2.703*** (-2.982) (2) 0.240** (2.004) 1.304*** (11.086) 0.004*** (9.441) 0.011*** (12.617) -0.098 (-1.539) -0.353*** (-9.119) -0.511*** (-6.919) 1.266*** (4.581) 7.274*** (6.574) 7,186 30.58% 7,186 45.63% No Yes Yes Yes 33 

35) Table 5: Determinants of firm-year with at least one poison put covenant issued This table presents the results of logistic regressions, in which the dependent variable is equal to one if the firm-year is associated with the issuance of a bond with a poison put covenant, and zero otherwise. Explanatory variables are as defined in Table 4, with the exception of: InstitutionalOwn (the percentage of ownership by institutional investors), Delaware (an indicator variable equal to one if the firm was incorporated in Delaware, and zero otherwise), PoisonPill (an indicator variable equal to one if the firm had a poison pill, and zero otherwise), ClassifiedBoard (an indicator variable equal to one if the firm had a classified board, and zero otherwise). Year and Fama-French 12 industry fixed effects are included, where applicable. Firm-clustered standard errors are provided in parentheses; *, **, and *** denote significant differences in means and medians at the 10%, 5%, and 1% levels respectively. LogRevenue MarketToBook Leverage InstitutionalOwn (1) -0.367*** (-17.798) -0.133*** (-2.839) 1.822*** (9.113) 0.321*** (2.857) (2) -0.622*** (-23.269) -0.149*** (-2.744) 2.396*** (10.551) 0.671*** (4.483) 14.587*** (21.981) (3) -0.452*** (-10.847) -0.239*** (-2.980) 1.047*** (2.912) 2.057*** (8.886) -0.186* (-1.749) -0.835*** (-7.270) -0.135 (-1.222) 9.270*** (9.190) (4) -0.772*** (-14.406) -0.220** (-2.274) 1.857*** (4.356) 1.071*** (4.099) 0.523*** (3.644) -0.149 (-1.065) 0.059 (0.456) 18.034*** (13.507) 7.402*** (15.365) 5,378 19.27% 5,378 32.54% 2,090 19.66% 2,090 36.69% No Yes Yes Yes No Yes Yes Yes Delaware PoisonPill ClassifiedBoard Constant Observations Pseudo R2 Year fixed-effects Industry fixed-effects 34 

36) Table 6: Covenants associated with poison puts This table presents a description of the covenants associated with a poison put issue. In Panel A, we provide the proportion of issues with poison puts that also have certain covenants. In Panel B, we provide the prevalence of poison puts among issues that have certain covenants. The other covenants are defined as follows: After_acquired_property_clause (property that is acquired after the sale of current debt issues will be used in the current mortgage), Asset_sale_clause (requirement that the issuer uses sale proceeds of certain assets to redeem bonds), Consolidation_merger (restriction of a consolidation or merger with another entity), Covenant_defeas_wo_tax_conseq (provides issuer the right to defease the indenture covenants without tax consequences to the bondholder), Cross_acceleration (enables bondholders to accelerate their debt, if any other debt of the firm has been accelerated due to default), Cross_default (activates an act of default if any other debt of the firm experiences default), Declining_net_worth (triggers certain bond provisions when the issuer’s net worth falls below a certain level), Defeasance_wo_tax_conseq (provides issuer the right to defease the issue’s monetary portion without tax consequences to the bondholder), Dividends_related_payments (limits to some ratio the payments made to shareholders or other entities), Economic_cov_def (provides issuer the right to defease indentures), Fixed_charge_coverage (requirement that issuer has a minimum amount of earnings available for fixed charges), Funded_debt (restriction from issuing additional “funded debt,” defined as any debt with a maturity of at least one year), Indebtedness (restriction on incurring additional debt beyond a given dollar amount or as a percentage of total capital), Investments (restriction on the issuer’s investment policy), Legal_defeasance (provides the issuer the right to defease the security’s monetary portion); Leverage_test (restricts the issuer’s total indebtedness), Liens (provides bondholders with unpaid obligations the right to sell mortgaged property in the event of bankruptcy), Maintenance_net_worth (requirement of issuer to maintain a minimum net worth), Negative_pledge_covenant (requirement to secure current issue on a pari passu basis if the firm issues secured debt), Net_earnings_test_issuance (requirement of certain profitability levels for firm to issue additional debt), Rating_decline_trigger_put (triggers a bondholder put provision in the event of a rating decline), Restricted_payments (restriction on issuer for making payments not related to dividends to shareholders and other parties), Sale_assets (restriction on the issuer’s ability to sell assets or on using the proceeds from asset sales), Sales_leaseback (restriction on the type or amount used in sale-leaseback transactions), Senior_debt_issuance (restriction on the amount of senior debt that may be issued), Stock_issuance_issuer (restriction on issuing common stock), Stock_transfer_sale_disp (restriction on transferring or selling its own common stock), Subordinated_debt_issuance (restriction on the issuance of subordinated or junior debt), Transaction_affiliates (restriction on certain types of transactions with subsidiaries). 35 

37) Panel A: Issues with poison puts – proportion of other covenants that are also present Covenant Proportion Consolidation_merger 92.8% Sale_assets 92.6% Cross_acceleration 80.7% Defeasance_wo_tax_conseq 76.2% Covenant_defeas_wo_tax_conseq 73.7% Negative_pledge_covenant 70.3% Sales_leaseback 54.7% Indebtedness 34.7% Restricted_payments 30.2% Transaction_affiliates 30.2% Fixed_charge_coverage 21.7% Asset_sale_clause 14.9% Dividends_related_payments 9.4% Legal_defeasance 9.2% Subordinated_debt_issuance 7.7% Cross_default 7.0% Economic_cov_def 6.1% Rating_decline_trigger_put 6.1% Liens 5.1% Investments 4.3% Stock_transfer_sale_disp 3.5% Stock_issuance_issuer 3.0% Maintenance_net_worth 2.6% Declining_net_worth 1.8% Funded_debt 1.5% Senior_debt_issuance 0.5% After_acquired_property_clause 0.2% Net_earnings_test_issuance 0.1% Leverage_test 0.0% 36 

38) Panel B: Issues with certain covenants – proportion that also has poison puts Covenant Proportion Asset_sale_clause 99.3% Rating_decline_trigger_put 99.1% Restricted_payments 91.9% Transaction_affiliates 90.0% Investments 83.3% Declining_net_worth 82.9% Subordinated_debt_issuance 75.9% Indebtedness 64.6% Fixed_charge_coverage 58.2% Stock_issuance_issuer 55.9% Senior_debt_issuance 47.4% Dividends_related_payments 43.6% Cross_acceleration 37.9% Maintenance_net_worth 33.3% Funded_debt 33.3% Liens 32.8% Covenant_defeas_wo_tax_conseq 31.7% Sales_leaseback 31.0% Defeasance_wo_tax_conseq 30.7% Sale_assets 29.1% Consolidation_merger 29.1% Negative_pledge_covenant 28.8% Economic_cov_def 28.5% Cross_default 24.0% Legal_defeasance 22.2% Stock_transfer_sale_disp 19.6% After_acquired_property_clause 1.9% Net_earnings_test_issuance 0.9% Leverage_test 0.0% 37 

39) Table 7: Differences in means and medians; poison puts and mergers This table examines the differences in means and medians for firms that had recently issued poison puts. In Panel A, we examine the propensity of firms to experience M&A-related activities, depending on their past issuance of poison puts. In Panel B, we examine deal characteristics where the target has a poison put. We distinguish a firm as having a poison put if it issued at least one bond with a poison put within the prior five years; all other firms are regarded as non-poison put firms. Our other variables in Panel A are: Acquirer (an indicator variable equal to one if the firm was an acquirer in that firm-year, and zero otherwise), Acquirer_year_mean (the average of firm-years in the industry where firms have Acquirer equal to one), Target (an indicator variable equal to one if the firm was a target in that firm-year, and zero otherwise), Target_year_mean (the average of firm-years in the industry where firms have Target equal to one). In Panel B, our variables are: Hostile (an indicator variable equal to one if the deal is hostile, and zero otherwise), Transaction Value (the deal’s transaction value), Market to Book (the market to book value of assets of the acquirer or target, as specified in the table), Book Leverage (the book value of leverage of the acquirer or target, as specified in the table), Market Leverage (the market value of leverage of the acquirer or target, as specified in the table), Earnings Volatility (the earnings volatility based on the prior five years of income before extraordinary items scaled by the book value of assets, for the acquirer or target). *, **, and *** denote significant differences in means and medians at the 10%, 5%, and 1% levels respectively. Panel A: M&A activity for firms with and without poison put covenants Poison put firm Non-poison put firm N Mean Median N Mean Median Acquirer 2,416 0.15 0.00 2,962 0.21*** 0.00*** Acquirer_year_mean 2,416 0.06 0.07 2,962 0.08*** 0.09*** Target 2,416 0.07 0.00 2,962 0.10*** 0.00*** Target_year_mean 2,416 0.03 0.03 2,962 0.04*** 0.04*** Panel B: Deal characteristics for firms with and without poison put covenants Poison put targets Non-poison put targets N Mean Median N Mean Median Hostile 222 0.02 0.00 7,833 0.02 0.00 Transaction Value 193 183.91 72.00 6,961 141.60** 53.00*** Market to Book (target) 222 1.58 1.36 7,833 1.93*** 1.28 Market to Book (acquirer) 162 1.99 1.52 5,874 2.23 1.48 Book Leverage (target) 222 0.42 0.40 7,833 0.26*** 0.22*** Book Leverage (acquirer) 162 0.32 0.30 5,874 0.23*** 0.21*** Market Leverage (target) 222 0.39 0.35 7,833 0.29*** 0.24*** Market Leverage (acquirer) 162 0.26 0.20 5,874 0.25 0.19* Earnings Volatility (target) 222 0.03 0.01 7,833 0.09*** 0.01* Earnings Volatility (acquirer) 162 0.02 0.01 5,874 0.10*** 0.01 38 

40) Table 8: Determinants of poison put issuance, merger characteristics This table provides the results of logistic regressions on the firm having issued a bond with a poison put covenant within the prior five years. All variables are as defined in Tables 4 and 7. Firm-clustered standard errors are provided below in parentheses. *, **, and *** denote significant differences in means and medians at the 10%, 5%, and 1% levels respectively. Target_year_mean Acquirer_year_mean LogRevenue MarketToBook BookLeverage EarningsVolatility InstitutionalOwn Constant (1) 14.738*** (4.155) (2) 9.645*** (3.649) -0.513*** -0.513*** (-11.220) (-11.218) -0.118 -0.127* (-1.635) (-1.765) 2.157*** 2.140*** (6.196) (6.182) -0.041 -0.057 (-0.072) (-0.101) 0.951*** 0.938*** (3.821) (3.794) 12.904*** 12.923*** (11.629) (11.637) Observations Pseudo R2 5,378 29.33% 5,378 29.32% Year fixed-effects Industry fixedeffects Yes Yes Yes Yes 39 

41) Table 9: Effects of poison puts on risk-shifting This table provides the results of regressions for the risk-shifting hypothesis. The dependent variables are of the acquirer’s (1) market to book value of assets; (2) market leverage; (3) earnings volatility, as defined in Table 7. The independent variables are: PoisonPut (an indicator variable equal to one if the firm issued a bond with a poison put covenant within the last five years), LogRevenue (the natural log of one plus the target’s sales), MarketToBook (the target’s market to book value of assets), BookLeverage (the target’s book value of leverage), and EarningsVolatility (the target’s earnings volatility based on the prior five years of income before extraordinary items, scaled by assets). Year and Fama-French 12 fixed-effects are provided where applicable. Firm-clustered standard errors are provided below in parentheses. *, **, and *** denote significant differences in means and medians at the 10%, 5%, and 1% levels respectively. Market to Book Market Leverage Market Leverage Earnings Volatility (1) (5) (6) (7) (8) -0.034 -0.039* -0.039 -0.009* -0.010 (-0.199) (-1.299) (-1.485) (-1.652) (-1.624) (-1.652) (-1.646) 0.040 0.038 0.006 0.003 0.000 -0.003 -0.001 -0.002 (0.352) (1.042) (0.400) (0.012) (-0.366) (-0.754) (-0.862) 0.035 -0.123 -0.016 -0.012 0.017 0.015 -0.003 -0.004 (-1.040) (-1.247) (-1.224) (1.109) (1.004) (-1.091) (-1.680) -0.145 -0.299 0.228*** 0.221*** -0.007 -0.007 (-0.180) BookLeverage (4) -0.031 (0.208) MarketToBook (3) -0.043 (0.522) LogRevenue (2) -0.131 (-0.431) PoisonPut (-1.197) (3.632) (2.973) (-0.462) (-0.612) MarketLeverage 0.284*** 0.247*** (4.753) (3.258) EarningsVolatility Constant 0.088 0.058 (0.877) (0.694) 0.941 0.664 0.020 0.121 0.094 0.219 0.040 0.048 (0.470) (0.244) (0.130) (0.697) (0.637) (1.190) (1.036) (0.910) 494 494 494 494 494 494 494 494 5.27% 10.58% 21.64% 26.00% 23.11% 26.55% 3.43% 5.75% Year fixed-effects No Yes No Yes No Yes No Yes Industry fixed-effects Yes Yes Yes Yes Yes Yes Yes Yes Observations R 2 40