Dividend Policy and the Method of Payment in Mergers and Acquisitions

Jin Q Jeon*
Dongguk Business School Dongguk University 3-26 Pil-dong, Chung-gu Seoul 100-715, KOREA +822-2260-8884 jjeon@cba.ua.edu

James A. Ligon
Department of Economics, Finance & Legal Studies The University of Alabama P.O. Box 870224 Tuscaloosa, AL 35487-0224 205-348-6313 jligon@cba.ua.edu

Charn Soranakom
College of Management Mahidol University 69 Vipawadee Rangsit Rd. Phayathai, Bangkok, 10400 THAILAND +662-206-2000 Ext. 2105 cmcharn@mahidol.ac.th

Last modified: July 2010

* Corresponding Author.

Dividend Policy and the Method of Payment in Mergers and Acquisitions

ABSTRACT This study examines how relative differences in dividend policies affect the choice of payment method in mergers and acquisitions. Using the dividend clientele hypothesis, we hypothesize that, at the margin, the method of payment is more likely to be stock if the dividend policies of the two firms involved in a merger or acquisition are quite similar, but more likely to be cash if the dividend policies are much different. The empirical data support the relevance of the dividend clientele hypothesis for the method of payment in mergers and acquisitions. We also find some evidence that in stock-based deals a difference in dividend policies is negatively correlated with announcement returns.

JEL classification: G32, G34, G35 Keywords: Mergers and Acquisitions; Method of Payment; Dividend Clientele; Dividend Policies; Announcement Returns

Dividend Policy and the Method of Payment in Mergers and Acquisitions 1. Introduction The two most common currencies for payments in corporate mergers and acquisitions are cash and stock. Some mergers involve all cash and some all stock, but a mixture of both currencies is also widely observed. Stock became increasingly important as a payment method during the 1990s, but its use has declined somewhat in the first decade of this century. 1 Studies on the method of payment have identified many possible factors that influence the choice of payment method used in mergers and acquisitions including managerial ownership, asymmetric information, and relative cost, among others. In this study, we look at how the dividend policies of the two firms involved in a merger or acquisition affect the method of payment used in the deal. According to the dividend clientele literature, different payout policies appeal to different classes of investors (Elton and Gruber, 1970; Kalay, 1982; Scholz, 1992; Allen, Bernardo and Welch, 2000). Some investors prefer to receive the payout from firms as dividends, while others may prefer to receive the payout as capital gains. In stock based mergers and acquisitions, shareholders of either company or both companies may have an incentive to liquidate their holdings if the payout policy of the surviving firm no longer fits their preference for payout form. Elton and Gruber (1970, p. 68) comment that " a change in dividend policy might cause a change in clientele and this could be costly." They explain that, "One type of cost would be the transaction costs incurred by both buyers and sellers as the firm's clientele changes. Furthermore, there could be at least a short-run unfavorable
1

Heron and Lie (2002), for instance, find that about 31% of their takeover sample (427 out of 1,376 deals) are entirely stock financed between 1985 and 1997. Faccio and Masulis (2005) use European takeover data and report that 26.8% of their sample deals are financed only with stock.

price movement as the change in dividend policy is more apparent to those investors who find it less favorable (present stockholders) than to those who find it more favorable." Baker, Coval, and Stein (2007) show that if the demand curve for the acquirers stock is downward sloping, then the cost of a stock for stock merger is decreasing in the fraction of target shareholders who passively accept and retain acquirer shares. Each share retained by a passive target shareholder is one less share that must be absorbed by market investors with a lower evaluation of the acquirer. Demand curves for stocks may slope down for a number of reasons. Miller (1977) suggested the combined effects of differences of opinion and shortsale constraints. The empirical literature provides clear support for downward-sloping

demand curves (e.g. Harris and Gurel, 1986; Shleifer, 1986; Bagwell, 1992; Hodrick, 1999; Kaul, Mehrotra, and Morck, 2000; Wurgler and Zhuravskaya, 2002; and Greenwood, 2005). In the context of mergers, Mitchell, Pulvino, and Stafford (2004) focus on price pressure around mergers and Baker, et al. (2007) find evidence in support of inertial behavior by target shareholders and that acquirer returns are lower when inertia is lower. We argue that if the target firm's dividend policy is very different from the acquiring firm's, this may serve as a disincentive to a stock based acquisition. If the dividend clientele hypothesis holds and the dividend policies of the acquirer and the target are materially different and stock is used in the acquisition, this may cause a change in the shareholder composition of the targets shareholder base. That is, target shareholders may decide to rebalance their portfolios. They may decide to sell their position in the target, before the merger is consummated, or in the survivor firm, after the merger is consummated, if they do not like the dividend policy of the acquiring firm. In the language of Baker, et al. (2007), a similarity in dividend policies between the target and the acquirer increases inertia, while a

dissimilarity in dividend policies between the target and the acquirer decreases inertia, other things equal. Selling activities, resulting from shareholders exiting the target firm's clientele base, can have immediate adverse effects on target stock prices if some non-passive shareholders exit immediately upon the announcement, although their exit could occur at any point prior to or possibly after the merger. Since acquirer shareholders would not necessarily expect a change in dividend policy post-merger, they would have no immediate reason to sell at the announcement date for dividend policy related reasons. An acquirer related announcement effect might be possible, however, if merger arbitragers anticipate and respond immediately to effects on the target share price. Whether merger arbitragers respond immediately to dividend related sales of target shares, or respond only with a lag, is an empirical question. However, potential long-term shareholders who might prefer the dividend policy of the acquirer would delay acquisition at least until the price drops sufficiently to cover the transaction costs of rebalancing and until the merger has a high enough probability of being consummated. However, if the dividend policies of the two firms are very similar to one another, a stock based acquisition may be less likely to cause selling related to the acquisition because there is very little or no mismatch between the dividend policy of the combined firm and the shareholders' preferences for payout. Hence, managers of an acquiring firm who anticipate this effect may find it rational to take into consideration the dividend clientele effect when they consider the method of payment for the acquisition. Our empirical results generally support the hypothesis that the likelihood of acquirers using stock as the payment method in takeovers increases with the degree of similarity in dividend policies. To arrive at this conclusion, we first consider the case where acquirer

management faces the qualitative decision to pay in the form of either stock, cash, or some mix of the two. The results of our multinomial regressions show that the degree of difference in dividend policies is significantly higher for pure cash deals than pure stock deals. We alternatively examine the determinants of the proportion of cash payment used in takeovers using a two-limit Tobit approach. The results are similar to our discrete analysis showing that a difference in dividend policies significantly increases the percentage of cash payment. In addition, consistent with previous studies, we find that several deal, target, and acquirer characteristics serve as important determinants of the payment choice. We also examine the relationship between the degree of similarity in dividend policies and takeover abnormal announcement returns. According to the dividend clientele hypothesis, target shareholders who do not prefer the dividend policy of the acquiring firm may sell at any point from the announcement date until after the merger consummation. However, clienteles preferring the dividend policy of the acquirer would not buy until the likelihood of merger consummation was high enough and the price drop large enough to compensate them for the risk of merger failure and the transaction costs of rebalancing. Target announcement returns will be lower if some dividend related selling occurs at the announcement date and the effect is not fully transferred by merger arbitragers immediately to the acquirer. Acquirer announcement returns will be lower if merger arbitragers immediately transfer, at least partially, the effects of any selling by target shareholders to the acquirer. Therefore, our second hypothesis states that in stock based acquisitions abnormal announcement returns of targets, acquirers, or both are lower if the level of difference in dividend policies is larger.

To test this hypothesis, we conduct both OLS and a 2-step procedure that controls for the potential endogeneity of the choice of the payment methods. Our results show that target cumulative abnormal returns (CARs) at the takeover announcement date for stock deals are significantly lower when one firm pays a dividend but the other does not and decreases in the difference in dividend policies between target and acquiring firms when dividend yield is used to measure the differential in dividend policies. We also find some evidence that acquirer CARs decrease in the difference in dividend policies. In addition, our results are largely consistent with adverse selection explanations that announcement returns are lower for stock based acquisitions because acquirers may use their stock to pay for acquisitions only when their stock is overvalued (Myers and Majluf, 1984; Travlos, 1987; Amihud, Lev and Travlos, 1990). In sum, this study provides a new perspective regarding the takeover method of payment choice by showing that, in addition to other factors related to traditional explanations of method of payment, the dividend policies of target and acquiring firms are a key determinant of the payment choice. The remainder of this paper is organized as follows. In Section 2, we present the research hypotheses and describe the dependent and test variables and the testing techniques. Section 3 details the sample selection process and provides the descriptions of the control variables and the rationales for including them, including references to relevant literature, and presents the descriptive statistics. Section 4 analyzes the determinants of the payment method. Section 5 analyzes announcement returns, and Section 6 concludes.

2. Hypotheses Development, Dependent and Test Variables, and Testing Techniques 2.1. Hypotheses We construct and test two hypotheses in this study. The first hypothesis states that, if the dividend policies of the two firms involved in an acquisition are quite similar, it is more likely that a method of payment for the acquisition is going to be stock, other things equal. A method of payment is more likely to be cash if the dividend policies are much different, other things equal. Ceteris paribus, the management should consider using a stock based acquisition when the dividend policies are similar and use cash when the dividend policies are fundamentally different. Our second hypothesis states that for stock based mergers, abnormal stock returns on the announcement day for the target, the acquirer, or both are lower if the level of difference in dividend policies is greater. In this case, selling activities by target shareholders who do not like the change in dividend policy that a successful merger would entail may put downward pressure on the target stock price. Potential buyers who prefer the new dividend policy may delay their purchase until the merger has a reasonably high probability of consummation and they can cover the transaction costs of rebalancing through a temporarily depressed stock price. These factors may result in a lower announcement return, ceteris paribus, for targets when the dividend policies of the target and the acquirer differ substantially. A negative effect of dividend differences on announcement returns for stock based mergers for acquiring firms is also possible if dividend related effects are anticipated and arbitraged by the market. Both may experience negative return if any effect is only partially arbitraged.

2.2. The Dependent and Test Variables 2.2.1. Dependent Variables : Method of Payment and Announcement Returns We take both dummy variable and continuous variable approaches to measure the payment method. First, we categorize deals into three groups in terms of their payment methods; Cash Only, Mixed Payment and Stock Only. Cash Only includes deals where at least 90% of consideration is paid with cash. Stock Only includes deals where at least 90% of consideration is paid with acquirer stock.2 Mixed Payment includes deals financed with both stock and cash. Alternatively, we use a continuous variable, %Cash PMT or %Stock PMT, which takes on any value between zero and one. Announcement returns, CARs, are the cumulative abnormal returns over the threeday window [-1, +1] or five-day window [-2, +2] around the bid announcement using the firm return minus the CRSP value-weighted market return. We calculate the announcement returns for both target and acquiring firms. 2.2.2. Measures of a Difference in Dividend Policies We obtain cash dividends for each acquiring firm and target firm from the COMPUSTAT database. We collect quarterly dividend data as of the end of the quarter immediately preceding the quarter when a merger is announced. We then average actual dividend payments over the last 4 quarters in order to account for seasonality in dividend payments. For example, if a merger is announced in the second quarter of 2006, we average quarterly dividends from the second quarter of 2005 to the first quarter of 2006.3 We use three measures of a difference in dividend policies between an acquirer and
We use alternative definitions for cash- or stock-based acquisitions and the results are similar to those presented here. 3 In unreported results, we also use the quarterly dividend prior to the announcement date. The results are similar to those presented here.
2

target. First, One is Payer is a dummy variable equal to 1 if one of the two firms involved in the merger pays any dividend during the last 4 quarters prior to the deal announcement, while the other does not, and equals 0 otherwise. Dividend policies are significantly different in the case that only one of the two firms pays a dividend. There is less difference in dividend policies if both or neither firm pay dividends. We also calculate the absolute value of differences in the dividend yield between targets and acquirers, Diff.DivYield, and, in order to ensure that our results are not affected by stock price variation (Grinstein and Michaely, 2005; Li and Zhao, 2008), we also use the absolute value of differences in the dividend to book value ratio between targets and acquirers, Diff.Div/Book. DivYield is defined as the ratio of dividends per share to the market price per share and Div/Book is the ratio of the amount of dividends to the book value of assets. For example, for a merger between a firm with a dividend yield of 2.0% and a firm with a dividend yield of 5.0%, the dividend yield difference between the two companies would be |2.0% - 5.0%| = 3.0%. The closer this number is to zero, the more similar the dividend policies between the two firms are. For the coefficient on Diff.DivYield to be meaningful, we add a control variable Acquirer DivYield to control for the levels of dividend yields of the acquiring firms. Likewise, we add a control variable Acquirer Div/Book when Diff.Div/Book is included as a variable of interest. We do so because a 3% value reported in Diff.DivYield may result, for example, from a 3%-dividend-yield acquirer takes over a nondividend-paying firm, or it may result from a non-dividend-paying acquirer buys a 3%dividend-yield firm. The impact on the method of payment from these two cases may be different. Hence, we use the variable Acquirer DivYield to control for the differences in the acquirers dividend yields.

2.3. Testing Techniques To test the first hypothesis, we adopt the multinomial logit model as well as the twolimit Tobit model. First, we consider the case where acquirer management faces the qualitative financing decision. Faccio and Masulis (2005) argue that in many mixed deals, target shareholders have a choice to receive cash or stock, implying that acquiring firms do not determine the fraction of cash financing.4 Accordingly, we categorized the sample into three groups in terms of their payment methods. An indicator variable, MOP, takes a value of 0 for pure cash deals, 1 for mixed deals (whether the mix is acquirer or target determined), and 2 for pure stock deals. Using MOP as a dependent variable, we estimate the following multinomial logit regression : (1) where P(MOPi=K) is the probability that a deal i will have the Kth payment method. The variables of interest are the measures of the degree of similarity in dividend policies, One is Payer, Diff.DivYield, or Diff.Div/Book. Following previous literature, we include a number of other control variables including; deal characteristic variables such as Deal Premium, ln(Deal Value), Related Deal, Hostile, and Tender Offer, acquirer characteristic variables such as Insider Ownership, Insider Ownership Squared, Leverage, Change in Leverage, PPE/Book, Market/Book, Cash/Deal Value, and Prereturns, and target characteristic variables including Relative Size, Insider Ownership, Insider Ownership Squared, Institutional Ownership,

The fact that target shareholders have a choice to receive cash or stock should not affect our investigation of the relationship between method of payment and dividend differences since, in equilibrium, the more similar the dividend policies of the two firms the more likely stock is the method of payment, ceteris paribus, irrespective of whether the choice is made by the acquirer or target shareholders. In our examination of announcement returns, our key test variable is an interaction term related to stock only acquisitions.

Leverage, and Market/Book. All the control variables and the rationales for including them are discussed fully in Section 3.2. We also examine the effect of a difference in dividend policies on the fraction of cash financing by adopting the two-limit Tobit model. In this model, the dependent variable, %Cash PMT, can be thought of as a latent variable that is truncated at both lower and upper limits, namely at zero and one. %Cash PMT values that we actually observe are bounded between zero and one. Specifically, we estimate the following model : (2) where x is the set of explanatory variables and, %CashPMTi = 1 %CashPMTi = %CashPMTi * %CashPMTi = 0 if %CashPMTi * 1, all cash

if 0< %CashPMTi * <1, mix of cash and stock if %CashPMTi * 0, all stock

We estimate this Tobit equation employing the quasi maximum likelihood estimation (Q-MLE) that uses Huber-White sandwich estimators. This method enables us to have robust estimators in the presence of possible model misspecification in the nonlinear framework.5 This model has some advantages over other models. First, we do not need to limit our choice dependent variable to a binary type or multinomial type of variable. Second, our dependent variable now is a continuous variable, but we have overcome the inherent limitations of a standard OLS regression. An OLS regression model may return a %Cash PMT value of less than zero or greater than one. The two-limit Tobit model solves that problem. To test the second hypothesis, we estimate the following OLS regression :
In a nonlinear framework, the Q-MLE uses a robust sandwich estimator that maximizes a log-likelihood function that is possibly misspecified (See Cameron and Trivedi, 2005. pp. 146). The role of robust standard errors in a nonlinear framework is quite different from that in a linear model.
5

10

CARs=+1(Stock Only Measures of Dividend Differences) + 2 Measures of Dividend Differences + 3 Stock Only+ 4 X+e (3)

Announcement returns, CARs, are measured as the cumulative abnormal returns of targets and acquirers, respectively, relative to the CRSP value-weighted market index, computed for the event windows of [-1, +1] and [-2, +2] days around a bid announcement date. Measures of Dividend Differences refers to the three measures of dividend differences previously discussed. As discussed in our second hypothesis, we expect the interaction term of a dividend difference and a stock based acquisition dummy, Stock Only Measures of Dividend Differences, is negatively correlated with target and/or acquirer announcement returns. We include several control variables (designated by the vector X) in the

announcement returns regressions; including the acquirers ln (Market Value), Market/Book, Leverage, and PPE/Book and the targets Relative Size, Institutional Ownership, Leverage, Stock Return Volatility, and Market/Book. The control variable descriptions and the rationales for including them are again provided in Section 3.2. A potential concern in equation (3), however, is that the payment method, Stock Only, is an endogenous variable because it is not randomly determined (Faccio and Masulis, 2005). If the payment method is indeed endogenous, then the interaction term of the payment method and a difference in dividend policies also becomes an endogenous variable. As a result, controlling for this endogeneity is integral to estimating the model. In order to control for the potential endogenous problem in our regressions, we additionally estimate the following 2-stage least squares (2SLS) equations: 1st reduced form: %Stock PMT = 1 z1+ 2 z2+ 3X+u

11

 2nd reduced form: %Stock PMT Measures of Dividend Differences= 1 z1+ 2 z2+ 3 X+w

Main equation: CARs=+1P[%Stock PMT Measures of Dividend Differences] + 2 Measures of Dividend Differences +3 P[%Stock PMT]+ 4X+e where P[] refers to the predicted value from the first stage regressions.6 In this model, we have two sets of instrumental variables; one for %Stock PMT and the other for %Stock PMTMeasures of Dividend Differences. The set of instrumental variables for %Stock PMT is z1. Since %Stock PMTMeasures of Dividend Differences is an interaction term, the instrumental variables, z2, are z1Measures of Dividend Differences. Note that in order to satisfy the over-identification condition, all instrumental variable sets (z1 and z2) and the exogenous variable set, X, must be included in both first and second reduced forms.7 Instrumental variable sets, z1 and z2, are described in Section 3.2. 3. Sample Selection, Control Variable Descriptions, and Summary Statistics 3.1. Sample Our sample consists of all mergers and acquisitions that were announced from January 1, 2001 to December 31, 2007, obtained from the Securities Data Corporation (SDC) Platinum Mergers and Acquisitions database. All mergers and acquisitions must satisfy the following screening criteria : 1) deal value is greater than one million dollars and is publicly disclosed, 2) the percentage of shares of the target firm held by an acquirer at announcement is less than 50%, 3) stock prices are available in the Center for Research in Securities Prices
Instead of the discrete variable, Stock Only, we use a continuous variable for stock payment, %Stock PMT , defined as the percent of stock financing.
7 6

(4)

See Section 6.2 of Wooldridge (2002) for detailed discussion of this model.

12

(CRSP) database, 4) Financial data are available in COMPUSTAT for both targets and acquirers, and 5) insider ownership data are available in the Thomson Financial Network (TFN) Insider Filing database for both targets and acquirers. The sample restrictions result in a final sample of 1,022 deal observations. 3.2. Control Variable Descriptions The literature has suggested that various factors may have some impact on the outcome of a method of payment in acquisitions. The variables mentioned below will be used as our control variables for these factors in our payment choice regression models. 3.2.1. Ownership Structure Managers of an acquiring firm who value control may prefer to use cash as a means of payment in an acquisition because stock dilutes their ownership in the combined firm (Amihud, Lev, and Travlos, 1990; Harris and Raviv, 1988; Martin, 1996; Mayer and Walker, 1996; Stulz, 1988; Yook, Gangopadhyay, and McCabe, 1999). Conversely, managers of a target firm may be interested in retaining control in the combined firm after the merger or acquisition (Ghosh and Ruland, 1998). Ghosh and Ruland find that the target firms

managerial ownership is an even more important factor than the acquiring firms managerial ownership in explaining the method of payment in acquisitions. Faccio and Masulis (2005) indicate the role of managerial ownership may be non-linear and we control for this as well. Thus, we include the acquiring firm's percentage of insider ownership, Acquirer Insider Ownership, and the target firm's percentage of insider ownership, Target Insider Ownership, and their squares in our empirical models. The data on ownership is obtained from the TFN Insider Filing database. The more stake the management of the acquiring firm has in the firm, the more likely the acquiring firm will use cash as a means of payment for the

13

acquisition. So, we expect the coefficient of Acquirer Insider Ownership to have a positive sign. On the contrary, the higher the percentage of target firm's insider ownership, the more likely the means of payment will be stock. Thus, we expect the coefficient of Target Insider Ownership to have a negative sign.8 Baker, et al. (2007) show that if the proportion of passive shareholders who accept acquirer stock in a stock for stock deal decreases, acquirer returns to the acquisition decrease. They find that institutional shareholders are less likely to be passive than individual shareholders. Accordingly we control for the percentage of institutional target shareholders using Target Institutional Ownership. The larger this percentage, the less attractive a stock merger is for the acquirer and hence acquirers would be less likely to use stock and if they do so the merger is likely to have lower announcement returns. 3.2.2. Relative Cost of Funds According to Myers (1984) pecking order theory, firms should fund their investment opportunities from internally generated cash flow whenever feasible. If cash flow is

inadequate, debt should be the next financing option the firms should consider. Equity financing, being the most expensive, should be considered last, only when the firms have no other financing option. This implies that if the acquiring firm has a lot of free cash flow, the firm is more likely to use cash as a means of payment for the acquisition. Martin (1996), and Mayer and Walker (1996) find acquirers, who have an ample amount of cash on the balance sheet or who can generate a large amount of free cash flow and have a low level of leverage, tend to use cash and/or debt to finance their acquisitions. Our variable, Cash/Deal Value,
Faccio and Masulis (2005) provide the tradeoff hypothesis between corporate control concerns and debt constraints. They argue that an acquirers payment decision is influenced by debt capacity and existing leverage and, simultaneously, by insiders desire to maintain their control of firms. We consider leverage below.
8

14

measures the amount of acquirer cash plus marketable securities normalized by the value of the merger or acquisition. In the case where the acquirer does not have enough cash, but it is not already highly leveraged, the acquirer can issue new debt to fund the acquisition. The acquirer can also use the unused lending capacity from the target firm if the target firm is under-leveraged. Chaney, Lovata, and Philipich (1991) find that acquiring firms that use cash acquisitions tend to be highly levered small firms with high return on assets, while acquiring firms that use stock acquisitions tend to have large asset bases, low leverage, low return on assets, and high price-earnings ratios. Acquirer Leverage and Target Leverage measure the debt to assets ratio of acquirers and target firms, respectively. Faccio and Masulis (2005) control for the borrowing power of the acquirer using a variable related to the collateral capacity of the firm, which they measure using property, plant, and equipment (PPE) over the book value of total assets. We also include PPE/Book to control for the collateral value of the acquiring firms assets. Yook (2003) suggests that the leverage effect plays an important role in payment decisions. An increase in leverage caused by cash payment reduces the free cash flow problem that self-interested managers invest money on less profitable (or negative NPV) projects. The benefits of this leverage effect should be stronger when a firm has lower growth opportunities and large free cash on hand. To address this, we include Change in Leverage, defined as the change in the acquirers leverage (total liabilities/total assets) from t-2 to t-1 (where t is the deal announcement year). When the stock price runs up considerably, it makes equity financing relatively less expensive. Acquirers could take advantage of such an occurrence by using stock as a means

15

of payment when their stock prices have gone up prior to the merger or acquisition announcement. Conventional wisdom suggests that targets would recognize such overvalued stock offers and refuse them. However, research by Rhodes-Kropf and Viswanathan (2004) suggests that stock price appreciation of bidders and targets may be positively correlated. Mutual overvaluation can lead bidders to make, and targets to accept, stock offers introducing a possible positive correlation between stock price appreciation, merger frequency, and use of stock as a method of payment. Shleifer and Vishny (2003) construct a model where mutual overvaluation can also lead to a positive correlation between stock price performance and the use of stock as a method of payment in acquisitions. We use a 90trading-day market-adjusted cumulative return at the 30th trading day prior to the announcement date, Prereturns, to measure how much the stock price of each acquiring firm has run-up. We predict a negative sign on this coefficient. The higher the stock returns prior to mergers or acquisitions announcements, the higher the likelihood that acquirers will use stock. The growth opportunities of acquiring firms also affect the payment choice. An acquirer with a high growth rate may be able to use their high multiple stock to pay for acquisitions. Acquirer Market/Book is defined as the sum of total assets and market value of equity minus book value of equity divided by total assets. We also include Target Market/Book. According to Carleton, Guilkey, Harris and Stewart (1983), a high target market to book ratio represents potentially high capital gains for target shareholders and nondeductible goodwill for acquirers, hence low market to book ratios of targets are associated with the use of cash.

16

3.2.3. Asymmetric Information Payment for acquisitions in the form of stock can help alleviate an asymmetric information problem. Asymmetric information exists in acquisitions because acquirers may know more about the value of their own firms than do target firms, but may not be able to derive a correct estimate of the true value of target firms, and vice versa. Hansen (1987) regards a stock offer as a contingent pricing mechanism. He finds that when target firms know their values better than acquirers, the acquirers will prefer to use stock, which has desirable contingent-pricing characteristics, rather than cash. When asymmetric information exists on both acquirers and target firms sides, a signaling equilibrium develops whereby targets regard both the method of payment used and the size of the stock offer as signals of the value of the acquiring firms. However, he finds only minimal supportive evidence. Fishman (1988) and Eckbo, Giammarino, and Heinkel (1990) also provide adverse selection based models of the acquirers choice of payment. When stock is used as a payment in acquisitions, the risks of a miscalculated firm valuation are shared between acquirers and targets. This type of risk is likely to be small when an acquirer is a much larger firm than is a target firm. As the target firms size increases, the risks are larger. Following Martin (1996), Relative Size is our proxy variable for information asymmetry which measures the relative size of target firm to acquiring firm. Relative Size is a ratio of targets market value of equity to the sum of targets market value and acquirers market value as of the year-end prior to the deal announcement.9 The larger the size of the target firm relative to the acquirer, the higher the risks of valuation

In unreported regressions we substitute an alternative relative size measure, which is the ratio of target firms total assets to the sum of targets total assets and acquirers total assets. The results are essentially unchanged.

17

miscalculation and, thus, the higher the probability of a use of stock as a payment consideration. 10 Target Return Volatility, an alternative proxy for information asymmetry that may be particularly relevant to announcement returns, is the standard deviation of daily stock returns during the one year prior to the bid announcement date. Officer, Poulsen, and Stegemoller (2009) find that acquirer announcement returns are greater in stock based acquisitions when target stock return volatility is greater, which implies that targets market value is difficult to estimate. 3.2.4. Deal Characteristics and Acquirer Size We examine several variables that capture deal characteristics. Deal Premium is defined as an acquirers offer value for the target over the pre-offer market value of the target minus one. We follow the approach of Officer (2003) to calculate deal premium. Ln(Deal Value) is the natural logarithm of deal value. Related Deal is a dummy variable equal to 1 if an acquirer and target share the same primary 2-digit SIC code and 0, otherwise. Hostile is a dummy variable equal to 1 if deal attitude is hostile and 0 if friendly or unsolicited as classified by the SDC. Tender Offer is a dummy variable equal to 1 if an acquirer involves a tender offer as reported in SDC and 0, otherwise. In the announcement return regressions we include a number of the variables previously discussed. We also include the natural log of the acquirers market value,

ln(Market Value) in the announcement return regressions. Our control variables in the

10

Mayer and Walker (1996) use both the interaction between earnings predictability and the market to book ratio of acquirers, and the ratio of market value of equity for target firms to that of acquirers to proxy for the information asymmetry. However, they find that these variables have only minimal impact on the method of payment.

18

announcement returns regressions follow previous literature such as Moeller, Schlingemann, and Stulz (2004) and Officer, Poulsen, and Stegemoller (2009). 3.2.5. Instrumental Variables for Equation (4) The set of instrumental variables for the percentage stock financing, %Stock PMT , in the first stage estimation associated with Equation (4) includes two variables. Average Acquirer Industry %Stock PMT is the acquirer industry-average (based on the 1-digit acquirer SIC code) proportion of stock financing during the quarter prior to the bid announcement. Average Target Industry %Stock PMT is the target industry-average %Stock PMT, computed in the same manner. We construct instrumental variables for an interaction variable of the proportion of stock financing and our three measures of differences in dividend policies, %Stock PMTMeasures of Dividend Differences, by interacting the instrumental variables for %Stock PMT and the measures of differences in dividend policies. 3.3. Descriptive Statistics [Table 1 about here] Table 1 presents the summary statistics when deals are classified by the payment method and dividend policy. There are 379 deals (37.1% of total sample) financed only with cash, while 271 deals (26.5%) are financed only with stock. The frequency of stock-based deals is the highest in the year 2001 (87 out of 208 deals), and is then generally decreasing over the sample years. The average of the cash portion of consideration is about 50%, while that of the stock portion is about 43%. During the sample period, 355 target and 535 acquiring firms pay dividends. The average quarterly dividend yield of target and acquiring

19

firms are 0.802% and 0.956%, respectively.11 The average dividend to book value ratio is 1.233% for targets and 1.555% for acquiring firms. In 303 mergers, only one firm (either a target or an acquirer) pays a positive dividend during last 4 quarters prior to the deal agreement. In 312 merger deals, both firms pay dividends, while no firm pays a dividend in 407 deals. The mean and median of the absolute differences in dividend yields, Diff.DivYield, are 0.404% and 0.077%, respectively, while those of Diff.Div/Book are 1.157% and 0.00%, respectively.

4. Effects of Differences in Dividend Policies on the Payment Method [Table 2 about here] Table 2 presents the mean and median differences in dividend policies for the three sub-groups in terms of the payment method. Tests for statistically significant differences between the Stock Only group and other groups are from t-tests and Wilcoxon rank-sum tests for each of the three measures of dividend differences: One is Payer, Diff.DivYield, and Diff.Div/Book. The overall results in Table 2 support the hypothesis that acquirers are more likely to pay with stock when dividend policies are similar. All 1,022 deals are examined in Panel A. Thirty-nine percent of the merger deals in the Cash Only group have a single dividend payer (i.e. are part of the One is Payer classification). In the Mixed Payment group, 24% of the deals have only one dividend payer. The percentages of One is Payer in both Cash Only and Mixed Payment groups are significantly higher than that in the Stock Only group, 19%. The mean and median Diff.DivYield of the Cash Only group are 0.495% and

11

Note that this is average quarterly dividend yield. One may calculate annual dividend yield by multiplying by

4.

20

0.122%, respectively, which are (weakly) significantly greater than those of the Stock Only group, 0.391% and 0%. The mean and median of Diff.Div/Book are also significantly higher for the Cash Only group (1.467% and 0.155%, respectively) than for the Stock Only group (0.472% and 0%). The Mixed Payment group also has greater dividend policy differences than the Stock Only group, while lower than the Cash Only group (these test results are not reported). In Panel B, we drop the deals where both targets and acquires do not pay any dividend in the previous quarters. As a result, the values of the percentage of One is Payer, Diff.DivYield and Diff.Div/Book are greater than those in Panel A. However, the results of univariate tests remain consistent with Panel A. [Table 3 about here] Table 3 reports estimates of multinomial logit models on the acquirers financing decisions as a function of the measures for dividend differences and the control variables. The table makes pair-wise comparisons between three categories of payment methods: Cash Only, Mixed Payment and Stock Only. Note that in a multinomial logit analysis, a regression coefficient indicates the effects on the log-odds between each of the groups and the reference group. In the first regression, where the reference choice is the Stock Only group, the coefficients of One is Payer, Diff.DivYield and Diff.Div/Book are all positive and statistically significant in the Cash Only group. Thus, if dividend differences between the target and acquirer are greater, the acquisition is more likely to be financed by cash than stock, which is consistent with our first hypothesis. In the second and third regressions, when Diff.DivYield (Diff.Div/Book) is used as the independent variable, we control for the DivYield (Div/Book) of acquiring firms. The negative coefficients on Acquirer DivYield and Acquirer Div/Book suggest that the higher the

21

acquirers dividend level, the more likely the acquisition is to be stock based. While we had no prior expectation regarding the sign of this variable, the empirical results are consistent with the proposition that dividend paying stock is more likely to be used as an acquisition vehicle. In each of our regressions, where the dependent variable is Mixed Payment and the reference group is Stock Only, the coefficients of our dividend difference measures are positive but not statistically significant. Also, in the third regression, where the dependent variable is Cash Only and reference group is Mixed Payment, the coefficients of the dividend difference measures are positive, but again insignificant. Therefore, the results suggest that even if the degree of dividend differences leads an acquirer to choose a cash deal rather than stock deal, it does not significantly affect a choice between mixed payment and stock only or a cash deal versus a mixed deal. Note that most of the coefficients of the control variables are signed in accordance with our expectations and prior literature. We find that tender offers are usually financed with cash (Jensen and Ruback, 1983). The positive coefficient of Cash/Deal Value implies that an acquirer maintaining more cash has a greater ability of cash financing and, therefore, is more likely to use cash as a mean of payment. Consistent with Faccio and Masulis (2005), the negative coefficient of Prereturns suggests that when acquirers stock price is overvalued at the announcement date the acquirer is more likely to use stock financing. Acquirer Leverage displays a positive sign, which suggests that acquirers may borrow early in anticipation of future cash acquisitions. The coefficient of Target Insider Ownership displays a significant quadratic relationship, first falling and then rising, implying higher managerial ownership in target firms at first decreases, but ultimately increases, the likelihood of the

22

method of payment being cash. Ghosh and Ruland (1998) have previously documented a negative relationship (i.e. a positive relationship between target inside ownership and stock payments). Relative Size is negatively correlated with the probability of cash payment and is weakly significant, implying a larger target size, which increases the risk of valuation miscalculation, results in the lower chance of cash financing rather than stock financing. The results support the asymmetric information hypothesis (Hansen, 1987; Martin, 1996). Target Institutional Ownership is positively related to the probability of a cash payment, consistent with the arguments of Baker, et al. (2007). Target Leverage is negatively related to the probability of a cash payment. [Table 4 about here] In Table 4, we alternatively employ a two-limit Tobit approach to examine the effect of dividend differences on the fraction of cash financing, %Cash PMT. The Q-MLE is used to maximize a log-likelihood function, in cases where that function is possibly misspecified due to the specification of the wrong density. The proxies for a difference in dividend policies (One is Payer, Diff.DivYield and Diff.Div/Book) have positive and significant coefficients, implying that the proportion of cash financing is increasing in these variables. The result is consistent with the first hypothesis that with a greater difference in the dividend policies between an acquirer and target, acquirer management is more likely to choose cash as the payment form. The effect of the level of the dividend yield or dividend to book ratio of an acquirer in the second and third regressions is negative and significant, consistent with the findings in the first regression in Table 3. As expected, other important determinants of the percentage of cash payment include Deal Premium, Tender Offer, Cash/Deal Value, Prereturns, Target Insider Ownership, Target Institutional Ownership, and Target Leverage.

23

They are correctly signed as discussed in the results for Table 3. Leverage is positive.

Once again Acquirer

Overall, the results of our univariate tests, multinomial regressions, and Tobit regressions are consistent with a notion that the dividend clientele effect significantly affects the choice of payment in acquisitions. Specifically, after controlling for the acquirers dividend levels, the closer the dividend policies are between the acquirers and the target firms, the more likely the acquirers pay for the acquisition with stock. This implies that, for a given acquirers dividend level, the likelihood of acquirers using stock as a means of payment in acquisitions increases with the degree of similarity in the dividend policies of acquirers and target firms.

5. Effects of Dividend Policies on Announcement Returns In this section, we test our second hypothesis that abnormal stock returns on the announcement date are lower for stock based mergers if the level of difference in dividend policies is greater. Consistent with the dividend clientele story, an acquirer with a dividend policy that differs from that of its target who uses stock may result in selling activities by target shareholders who do not prefer the acquirers dividend policy, which may cause a price drop in the target firms stock, and, ultimately, a drop in the acquiring firms stock. [Table 5 about here] Table 5 presents the average announcement CARs of both target and acquiring firms and the test results for mean differences. In this table, the sample is divided into three groups based on payment methods. The overall average CARs suggest that target firms are clearly winners in merger transactions with average three-day and five-day CARs of 23.81% and

24

24.60%, respectively. In contrast, takeovers are possibly wealth destroying for acquiring firms where average CARs are -1.45% and -1.41% in the three-day and five-day windows, respectively. The table also shows that the average CARs of the Cash Only group are significantly higher than the Stock Only group. The average CARs of acquiring firms are 0.41% in the three-day window and 0.57% in the five-day window for a cash deal, while they have negative values for a stock deal. Also, cash based takeovers are associated with greater target CARs. As widely documented in the literature, the lower announcement returns associated with stock deals are consistent with the adverse selection argument initially suggested by Myers and Majluf (1984). The mergers and acquisitions literature applies the adverse selection argument to suggest that acquiring firms pay with stock only when their shares are overvalued (Travlos, 1987; Amihud, Lev and Travlos, 1990). [Table 6 about here] Table 6 presents the results of OLS regressions that analyze the determinants of takeover announcement returns for both target and acquiring firms. The variable of interest in the regressions is the interaction variable of a dummy for a stock deal and the difference in dividend policies. As discussed above, our second hypothesis states that abnormal announcement returns are lower for a stock based deal with a greater difference in dividend policies. Whether this negative announcement effect is greater for target firms or acquiring firms depends upon whether merger arbitragers anticipate the extent of dividend related selling and the speed with which they shift its effects to acquirers. In Panel A, consistent with our hypothesis, the coefficients of the interaction variables between a stock deal dummy and the degree of the dividend differences are negatively correlated with target returns for all three measures and are statistically significant when One is Payer and Diff.DivYield are used

25

to measure the dividend differential. In stock based mergers target CARs decrease by, on average, 2.6% during the three-day window around the deal announcement and 5.9% during the five-day window if only one firm pays a dividend.12 Also, target CARs drop by 2.9% or 4.5% during the three-day or five-day windows, respectively, when the difference in dividend yield increases by one unit. In the acquirer CARs regressions in Panel B, the interaction terms again have negative coefficients for the three measures of dividend differences but are significant only when One is Payer and Diff.Div/Book (and the latter at the 10% level) are used as the dividend difference measure. In stock based takeovers, CARs of acquiring firms decrease by 0.6% (that is, -0.015 + 0.009) or 0.3% during the three-day or five-day windows, respectively, when only one firm pays a dividend. The evidence suggests that the market does not fully immediately shift the dividend effect in the acquirer stock price, but that some arbitrage activity does occur. In both panels, the coefficients of a dummy for a pure stock deal, Stock Only, are negative and significant at the 1% level in 11 of the 12 specifications (and at the 5% level in the other), supporting the adverse selection hypothesis. We include several control variables in both the target and acquirer CARs regressions as suggested by previous literature. Target and acquirer announcement returns are significantly higher if the ln(Market Value) of an acquirer is larger, while they are lower if the targets Relative Size is bigger. The positive coefficient on target Stock Return Volatility suggests that target firms with greater stock return volatility are more likely to benefit from takeovers. Higher Target Institutional Ownership depresses announcement returns for both targets and acquirers consistent with Baker, et al.s (2007) passive shareholder arguments.
12

Note that the coefficient of One is Payer is not statistically different from 0 in both the three-day and fiveday CARs regressions. Therefore, the marginal effect of One is Payer in stock based deals is -0.0261+0=0.026.

26

Finally, we find that Target Leverage is negatively associated with target announcement returns. [Table 7 about here] A potential concern in Table 6 is that a decision on a pure stock payment would not be randomly assigned, resulting in endogeneity bias. As discussed in Section I.C, we employ the two-stage least squares (2SLS) approach to control for the endogeneity of payment decisions in takeovers. Table 7 reports the results of both first and second stage 2SLS regressions. In Panel A, One is Payer is used for the measure of dividend difference, while Diff.DivYield and Diff.Div/Book are used in Panel B and Panel C, respectively. In the reduced form regressions in each panel, we estimate the determinants of the endogenous variables, %Stock PMT and %Stock PMTMeasure of Dividend Differences. In each panel, the instrumental variables, described in Section 3.2.5, are significantly and positively correlated with each of the endogenous variables they are intended to instrument, confirming that there is no weak instrument problem. The results of 2SLS in Table 7 are generally consistent with our findings from the OLS regressions in Table 6. In Panel A, the predicted value of the interaction variable between %Stock and One is Payer is negatively and significantly correlated with target CARs, confirming that our results in Table 6 are robust after controlling for endogeneity. For acquiring firms, it is negative and weakly significant when three-day CARs are used, but not significant when five-day CARs are used. In Panel B Diff.DivYield, is used for the measure of the dividend differential. Consistent with our previous findings, target CARs are negatively correlated with the predicted value of the interaction variable, %Stock Diff.DivYield. The coefficient of the interaction variable, however, is negative but not 27

significant for acquirer CARs. In Panel C Diff.Div/Book, is used for the measure of dividend differential. The effect of the predicted value of %StockDiff.Div/Book is negative but not significant on both target and acquirer abnormal announcement returns. The results are thus similar to the OLS results for target CARs but somewhat weaker, in terms of statistical significance, for acquirer CARs. In addition, the predicted values of the percentage of stock payment in each panel are negatively correlated with both target and acquirer announcement returns, consistent with adverse selection explanations. Overall, the results of the announcement returns analysis support our second hypothesis. In stock based takeovers, the difference in dividend policies has a negative effect on target returns, reflecting the possibility of selling activities by target shareholders who will experience a change in dividend policy from that of the target firm to that of the acquirer. The effect, however, is less significant in the acquirer CARs regressions, suggesting that the market does not immediately fully arbitrage the dividend related effect.

6. Conclusions The dividend clientele hypothesis suggests that shareholders are different in their preferences for payouts from the firms they invest in. Some shareholders prefer to receive a regular stream of income in the form of cash dividends, while others may prefer to forgo cash dividends in order to get a possibly better payout from a firm in the form of capital gains. Hence, different dividend policies attract different types of shareholders. Changes in dividend policies resulting from mergers of firms with dissimilar policies may result in portfolio rebalancing from target shareholders. Shareholders may reduce or totally liquidate their positions from the firms whose dividend policies have changed in a 28

way that is unfavorable to them. In anticipation of such a consequence, acquirers may at the margin select a method of payment based on how different the dividend policies are between the acquirers and the target firms. We offer two hypotheses. First, similar dividend policies between acquirers and targets increase the likelihood of the use of stock as a method of payment in acquisitions. Second, in stock based acquisitions abnormal announcement returns are lower if the level of difference in dividend policies is larger. Our empirical results generally support the hypothesis that the likelihood of acquirers using stock as the payment method in takeovers increases with the degree of similarity in dividend policies. To arrive at this conclusion, first we consider the case where acquirer management faces the qualitative decision to pay in the form of stock, cash, or a mix of the two. The results of our multinomial analysis show that the degree of difference in dividend policies is significantly higher for pure cash deals than pure stock deals. We, then, examine the determinants of the proportion of cash payment using a two-limit Tobit approach. The results show that a larger difference in dividend policies significantly increases the percentage of cash payment. We also find that the higher the dividend on the acquirer stock, the more likely it is to be used as an acquisition currency. To test our second hypothesis related to announcement returns, we conduct OLS and a 2-step procedure in order to control for endogeneity of payment decisions. Our results show that, for stock deals, target cumulative abnormal returns (CARs) around the takeover announcement date significantly decrease if only one firm pays a dividend and the other does not or if the difference in dividend yields between target and acquiring firms becomes larger. The negative effect for acquirer CARs, however, is only weakly significant and only when

29

one firm pays a dividend and the other does not. These results suggest that the market does not fully immediately incorporate the dividend effect in the acquirer stock price. In addition, our results are largely consistent with adverse selection explanations of the method of payment, which suggest that stock based acquisitions produce lower announcement returns because acquirers may use their stock to pay for acquisitions only when their stocks are overvalued. In sum, this study provides a new perspective regarding the takeover method of payment choice. We show that, in addition to factors related to traditional explanations of the method of payment, the dividend policies of target and acquiring firms are a key determinant of the payment choice.

30

References Allen, F., Bernardo, A., Welch, I., 2000. A theory of dividends based on tax clienteles. Journal of Finance 55, 24992536. Amihud, Y., Lev, B., Travlos, N.G., 1990. Corporate control and the choice of investment financing: The case of corporate acquisitions. Journal of Finance 45, 603-616. Bagwell, L., 1992. Dutch auction repurchases: An analysis of shareholders heterogeneity. Journal of Finance 47, 71-105. Baker, M., Coval, J., Stein, J.C., 2007. Corporate financing decisions when investors take the path of least resistance. Journal of Financial Economics 84, 266-298. Carleton, W., Guilkey, D., Harris, R., Stewart, J., 1983. An empirical analysis of the role of the medium of exchange in mergers. Journal of Finance 38, 813-826. Cameron, A. C., P.K. Trivedi, P.K., 2005. Microeconometrics. Cambridge University Press, New York. Chaney, P. K., Lovata, L.M., Philipich, K.L., 1991. Acquiring firm characteristics and the medium of exchange. Quarterly Journal of Business and Economics 30, 55-69. Eckbo, B. E., Giammarino, R.M., Heinkel, R.L., 1990. Asymmetric information and the medium of exchange in takeovers: Theory and tests. Review of Financial Studies 3, 651675. Elton, E. J., Gruber, M.J., 1970. Marginal stockholder tax rates and the clientele effect. Review of Economics and Statistics 52, 68-74. Faccio, M., Masulis, R., 2005. The choice of payment method in European mergers and acquisitions. Journal of Finance 60, 13451388. Fishman, M. J., 1988. Preemptive bidding and the role of the medium of exchange in acquisitions. Journal of Finance 44, 41-57. Ghosh, A., Ruland, W., 1998. Managerial ownership, the method of payment for acquisitions and executive job retention. Journal of Finance 53, 785-798. Greenwood, R., 2005. Short- and long-term demand curves for stocks: Theory and evidence on the dynamics of arbitrage. Journal of Financial Economics 75, 607-649. Grinstein, Y., Michaely, R., 2005. Institutional holdings and payout policy. Journal of Finance 60, 1389-1426.

31

Hansen, R. G., 1987. A theory for the choice of exchange medium in mergers and acquisitions. Journal of Business 60, 75-95. Harris, L., Gurel, E., 1986. Price and volume effects associated with changes in the S&P500 list: New evidence for the existence of price pressures. Journal of Finance 41, 815-829. Harris, M., Raviv, A., 1988. Corporate control contests and capital structure. Journal of Financial Economics 20, 55-86. Heron, R., Lie, E., 2002. Operating performance and the method of payment in takeovers. Journal of Financial and Quantitative Analysis 37, 137-155. Hodrick, L., 1999. Does stock price elasticity affect corporate financial decisions? Journal of Financial Economics 52, 225-256. Jensen, M. C., Ruback, R.S., 1983. The market for corporate control. The scientific evidence. Journal of Financial Economics 11, 5-50 Kalay, A., 1982. The ex-dividend day behavior of stock prices: A re-examination of the clientele effect. Journal of Finance 37, 1059-1070. Kaul, A., Mehrotra, V., Morck, R., 2000. Demand curves for stocks do slope down: New evidence from an index weights adjustment. Journal of Finance 55, 893-912. Li, K., Zhao, X., 2008. Asymmetric information and dividend policy. Financial Management 37, 673-694. Martin, K. J., 1996. The method of payment in corporate acquisitions, investment opportunities and management ownership. Journal of Finance 51, 1227-1246. Mayer, W. J., Walker, M.M. 1996. An empirical analysis of the choice of payment method in corporate acquisitions during 1980 to 1990. Quarterly Journal of Business and Economics 35, 48-65. Miller, E., 1977. Risk, uncertainty and divergence of opinion. Journal of Finance 32, 11511168. Mitchell, M, Pulvino, T., Stafford, E., 2004. Price pressure around mergers. Journal of Finance 59, 31-63. Moeller, S., Schlingemann, F., Stulz, R., 2004. Firm size and the gains from acquisitions. Journal of Financial Economics 73, 201228. Myers, S. C., 1984. The capital structure puzzle. Journal of Finance 39, 575-592.

32

Myers, S., Majluf, N., 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics 13, 187221. Officer, M. S., 2003. Termination fees in mergers and acquisitions. Journal of Financial Economics 69, 43167. Officer, M. S., Poulsen, A.B., Stegemoller, M., 2009. Target-firm information asymmetry and acquirer returns. Review of Finance 13, 467-493. Rhodes-Kropf, M., Viswanathan, S., 2004. Market valuation and merger waves. Journal of Finance 59, 2685-2718. Scholz, J. K., 1992. A direct examination of the dividend clientele hypothesis. Journal of Public Economics 49, 261-285. Shleifer, A., 1986. Do demand curves for stocks slope down? Journal of Finance 41, 579590. Shleifer, A., Vishny, R.W., 2003. Stock market driven acquisitions. Journal of Financial Economics 70, 295-311. Stulz, R. M., 1988. Managerial control of voting rights: Financing policies and the market for corporate control. Journal of Financial Economics 20, 25-54. Travlos, N., 1987. Corporate takeover bids, methods of payment, and bidding firms stock returns. Journal of Finance 42, 943963. Wooldridge, J. M., 2002. Econometric Analysis of Cross Section and Panel Data. The MIT Press, Cambridge, MA. Wurgler, J., Zhuravskaya, E., 2002. Does arbitrage flatten demand curves for stocks? Journal of Business 75, 583-608. Yook, K. C., 2003. Larger return to cash acquisitions: Signaling effect or leverage effect? Journal of Business 76, 477-498. Yook, K. C., Gangopadhyay, P., McCabe, G.M., 1999. Information asymmetry, management control, and method of payment in acquisitions. Journal of Financial Research 22, 413427.

33

Table 1. Descriptive Statistics

This table provides summary statistics regarding the method of payment and dividend policies of target and acquiring firms. The sample includes 1,022 merger agreements during the period January, 2001 to December, 2007. Cash Only includes deals where at least 90% of consideration is paid with cash. Stock Only includes deals where at least 90% of consideration is paid with acquirer stocks. Mixed Payment includes deals financed with both stock and cash. One is Payer, Diff.DivYield and Diff.Div/Book, based on quarterly dividends, are our measures of the degree of difference in dividend policies and are discussed in Section 2.2. Year N Payment Method Cash Only (N) Stock Only (N) Mixed Payment (N) % Cash Financing % Stock Financing Target Dividends Dividend Payer (N) DivYield (%) Div/Book (%) Acquirer Dividends Dividend Payer (N) DivYield (%) Div/Book (%) Dividend Differential One is Payer (N) Both are Payers (N) Neither is Payer (N) Diff.DivYield (%) Diff.Div/Book (%) 2001 Mean 208 48 87 73 34.708 55.101 57 0.881 0.895 79 0.779 1.191 59 57 92 0.195 0.754 2002 Mean 110 41 33 36 46.248 45.040 34 0.838 0.790 53 0.858 1.313 27 30 53 0.246 0.517 2003 Mean 150 50 40 60 47.950 45.578 54 0.812 0.269 75 1.576 1.163 31 49 70 0.736 0.563 2004 Mean 149 52 39 58 48.824 45.564 52 0.694 0.601 93 0.827 1.003 51 47 51 0.264 0.667 2005 Mean 148 53 29 66 52.799 40.402 53 0.814 1.692 82 0.983 2.752 49 43 56 0.510 1.845 2006 Mean 143 76 25 42 63.543 30.090 56 0.849 2.623 90 0.878 1.735 50 48 45 0.564 1.796 2007 Mean 114 59 18 37 64.574 31.282 49 0.716 1.579 63 0.798 1.740 36 38 40 0.348 2.240 Total Mean Median 1,022 [100%] 379 [37.1%] 271 [26.5%] 372 [36.4%] 49.937 46.135 42.945 43.725 355 0.802 1.233 535 0.956 1.555 303 312 407 0.404 1.157

0.570 0.270

0.652 0.569

0.077 0.000

34

Table 2. Univariate Tests

The sample is divided into three groups based on their payment methods. Cash Only includes deals where at least 90% of the consideration is paid with cash. Stock Only includes deals where at least 90% of the consideration is paid with acquirer stock. Mixed Payment includes deals financed with both stock and cash. Tests for statistically significant differences between the Stock Only group and other groups are from t-tests (for means) and Wilcoxon rank-sum (i.e. Mann-Whitney-Wilcoxon) tests for each of the three measures of differences in dividend policies, One is Payer, Diff.DivYield and Diff.Div/Book, which are discussed in Section 2.2. The Wilcoxon tests are technically for the equality of the distributions rather than medians per se. The symbols ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively. Panel A. Full Sample Cash Only N = 1022 Mean One is Payer (%) Diff.DivYield (%) Diff.Div/Book (%) 38.522 0.495 1.467
*** * ***

Mixed Payment 372 Median 0 0.122 0.155

*** *** ***

Stock Only 271 Mean

*** *** ***

379 Mean 23.656 0.423 1.341

Median
**

Median 0 0 0

0 0.106 0.006

19.188 0.391 0.472

**

Panel B. Subsample (When both targets and acquirers are dividend payers) Cash Only N = 588 Mean One is Payer (%) Diff.DivYield (%) Diff.Div/Book (%) 64.889 0.966 2.242
*** ** ***

Mixed Payment 231 Median 1 0.345 1.062

*** ** ***

Stock Only 132 Mean

* * ***

225

Mean 40.125 0.682 1.474

* *

Median 0 0.326 0.324

Median 0 0.295 0.165

38.393 0.602 0.662

35

Table 3. Determinants of the Payment Choice

Cash Only includes deals where at least 90% of consideration is paid with cash. Stock Only includes deals where at least 90% of consideration is paid with acquirer stock. Mixed Payment includes deals financed with both stock and cash. Multinomial logit regressions are estimated for the payment method categories. The measures of differences in dividend policies include One is Payer, Diff.DivYield, and Diff.Div/Book, which are discussed in Section 2.2. All tests use the QML robust standard errors and z statistics which are reported in brackets. The symbols ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively. Coefficients on Relative Size have been multiplied by 10,000 for presentation purposes. All of the control variables are discussed in Section 3.2.

Cash Only Reference group : Dividend Differential One is Payer Diff.DivYield Acquirer DivYield Diff.Div/Book Acquirer Div/Book Deal Characteristics Deal Premium Ln (Deal Value) Related Deal Hostile Tender Offer Acquirer Characteristics Insider Ownership Insider Ownership2 Leverage Leverage PPE/Book 0.614** [2.23] 33.308** [2.15] -30.697** [-2.05] 11.042** [2.32] -6.152* [-1.91] 0.592** [2.16] -0.296*** [-2.89] -0.112 [-0.54] -0.158 [-0.18] 1.752*** [4.96] -0.374 [-0.42] -0.010 [-0.03] 1.355*** [2.71] -1.082 [-1.08] 0.544 [1.59] 0.538 [1.25] 0.249* [1.76] -0.075 [-0.36] -0.076 [-0.09] 1.706*** [4.95] -0.478 [-0.53] 0.009 [0.03] 1.605*** [3.22] -1.258 [-1.23] 0.362 [1.07] 0.546 [1.34] 0.277* [1.71] -0.105 [-0.51] -0.131 [-0.15] 1.712*** [4.96] -0.304 [-0.34] -0.060 [-0.18] 1.444*** [2.84] -1.290 [-1.26] 0.213 [0.62] 0.227 [0.96] Stock Only

Mixed Payment Stock Only 0.387 [1.44] 17.753 [1.05] -15.562 [-0.95] 10.363 [0.87] 7.635 [0.72] 0.066 [0.25] 0.198** [2.13] 0.471** [2.58] 0.131 [0.15] 0.346 [0.96] 0.173 [0.21] -0.441 [-1.05] 0.253 [0.52] -1.456 [-1.43] 0.693** [2.26] 0.042 [0.16] 0.216** [2.34] 0.495*** [2.69] 0.153 [0.18] 0.337 [0.93] 0.141 [0.18] -0.440 [-1.04] 0.391 [0.81] -1.626 [-1.58] 0.613** [2.03] 0.075 [0.29] 0.200** [2.15] 0.464** [2.50] 0.105 [0.12] 0.333 [0.91] 0.311 [0.37] -0.510 [-1.18] 0.249 [0.49] -1.708 [-1.58] 0.449 [1.49] 0.526** [2.07] -0.494*** [-5.00] -0.583*** [-3.09] -0.289 [-0.45] 1.406*** [4.13] -0.546 [-0.62] 0.431 [1.13] 1.102** [2.13] 0.374 [0.41] -0.149 [-0.53]

Cash Only Mixed Payment

15.555 [1.01] -15.135 [-0.96] 0.679 [0.23] -1.483 [-0.46] 0.495 [1.35] -0.465 [-1.48] -0.570*** [-3.03] -0.228 [-0.35] 1.369*** [4.11] -0.620 [-0.70] 0.449 [1.16] 1.214** [2.41] 0.368 [0.41] -0.251 [-0.90] 0.470 [1.56] -0.477 [-1.49] -0.569*** [-3.03] -0.236 [-0.36] 1.378*** [4.13] -0.615 [-0.70] 0.449 [1.15] 1.194** [2.35] 0.418 [0.46] -0.236 [-0.84]

36

Market/Book Cash/Deal value Prereturns Target Characteristics Relative Size Insider Ownership Insider Ownership2 Institutional Ownership Leverage Market/Book Intercept

-0.029 [-0.31] 0.022** [2.14] -1.502*** [-2.59] -0.923* [-1.69] -4.121** [-2.58] 4.769*** [2.69] 2.835*** [5.49] -1.886*** [-3.97] 0.059 [0.63] 4.033** [2.24] 1,022 199.34 0.000 0.1449

-0.032 [-0.34] 0.022** [2.19] -1.587*** [-2.73] -0.895* [-1.69] -0.004** [-2.55] 0.047*** [2.72] 2.631*** [5.14] -1.630*** [-3.55] 0.038 [0.44] -3.205* [-1.79] 1,022 201.64 0.000 0.1443

-0.056 [-0.58] 0.020** [2.14] -1.535*** [-2.70] -0.858* [-1.66] -0.005*** [-2.60] 0.049*** [-2.81] 2.658*** [5.17] -1.610*** [-3.50] 0.037 [0.44] -3.769** [-2.10] 1,022 201.64 0.000 0.1443

0.069 [0.74] 0.002 [0.11] -0.582 [-1.04] -0.615*** [-2.70] 1.939 [1.22] -2.182 [-1.27] 0.314 [0.66] 0.501 [1.20] -0.173 [-1.54] -4.553*** [-2.73]

0.068 [0.72] 0.002 [0.09] -0.629 [-1.12] -0.611*** [-2.72] 1.938 [1.21] -2.227 [-1.29] 0.241 [0.51] 0.585 [1.38] -0.180 [-1.62] -4.890*** [-2.95]

0.046 [0.47] 0.000 [0.00] -0.580 [-1.03] -0.606*** [-2.71] 2.026 [1.26] -2.384 [-1.37] 0.232 [0.48] 0.611 [1.37] -0.188 [-1.60] -4.568*** [-2.72]

-0.098 [-1.20] 0.020 [1.15] -0.920** [-2.04] -0.308 [-0.55] 2.182 [1.48] -2.587 [-1.55] 2.522*** [5.34] -2.388*** [-5.17] 0.232* [1.84] 8.586*** [5.02]

-0.100 [-1.24] 0.020 [1.17] -0.958** [-2.11] -0.284 [-0.52] 2.089 [1.43] -2.508 [-1.53] 2.390*** [5.17] -2.215*** [-5.01] 0.219* [1.81] 8.095*** [4.81]

-0.102 [-1.22] 0.020 [1.15] -0.955** [-2.09] -0.252 [-0.48] 2.081 [1.43] -2.507 [-1.53] 2.426*** [5.23] -2.221*** [-5.07] 0.225* [1.83] 8.336*** [4.97]

No. observations Wald test P-value Pseudo R2

37

Table 4. Determinants of the Proportion of Cash Payment

Two-limit Tobit regressions are estimated where the dependent variable is the proportion of cash payment, %Cash PMT. The measures of differences in dividend policies include One is Payer, Diff.DivYield, and Diff.Div/Book, which are discussed in Section 2.2. All tests use the QML robust standard errors and t statistics which are reported in brackets. The symbols ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively. Coefficients on Relative Size have been multiplied by 10,000 for presentation purposes. All of the control variables are discussed in Section 3.2. Dependent variable: Dividend Differential

%Cash PMT
0.298 [3.29]*** 13.432 -11.387 [2.32]** [-2.11]** 1.668 -1.221 0.212 0.115 -0.006 -0.332 0.874 -0.212 -0.029 0.714 -0.427 0.053 -0.038 0.008 -0.679 -0.144 -1.557 1.723 1.112 -0.727 0.047 1.896 1,022 199.99 0.000 0.145 [2.01]** [1.97]** [-0.08] [-1.10] [6.37]*** [-0.58] [-0.16] [3.45]*** [-1.00] [0.45] [-1.10] [3.03]*** [-3.13]*** [-1.18] [-2.53]** [2.55]** [5.62]*** [-3.96]*** [1.32] [2.76]*** 0.232 0.056 -0.028 -0.318 0.876 -0.319 0.024 0.820 -0.492 0.078 -0.037 0.009 -0.701 -0.157 -0.016 0.171 1.168 -0.758 0.051 1.833 1,022 193.44 0.000 0.100 [2.19]** [1.45] [-0.35] [-1.04] [6.36]*** [-0.87] [0.13] [3.97]*** [-1.14] [0.65] [-1.07] [3.08]*** [-3.22]*** [-1.27] [-2.59]*** [2.53]** [5.87]*** [-4.12]*** [1.42] [2.64]*** 0.230 0.049 -0.041 -0.319 0.878 -0.268 0.007 0.794 -0.440 0.089 -0.039 0.009 -0.690 -0.151 -0.016 0.175 1.172 -0.757 0.051 1.902 1,022 190.29 0.000 0.099 [2.70]*** [-2.48]** [2.16]** [1.50] [-0.50] [-1.04] [6.36]*** [-0.73] [0.04] [3.81]*** [-1.03] [0.74] [-1.12] [3.11]*** [-3.16]*** [-1.23] [-2.61]*** [2.59]*** [5.90]*** [-4.10]*** [1.42] [2.75]***

One is Payer
Diff.DivYield Acquirer DivYield Diff.Div/Book Acquirer Div/Book Deal Characteristics Deal Premium ln (Deal Value) Related Deal Hostile Tender Offer Acquirer Characteristics Insider Ownership Insider Ownership2 Leverage Leverage PPE/Book Market/Book Cash/Deal value Prereturns Target Characteristics Relative Size Insider Ownership Insider Ownership2 Institutional Ownership Leverage Market/Book Intercept No. observations Wald test P-value Pseudo R2

38

Table 5. Univariate Tests for Announcement Returns

Announcement returns are measured as the cumulative abnormal returns (CARs) relative to the CRSP value-weighted market index over three day [-1, 1] and five day [-2, 2] horizons. The sample is divided into three groups based on the method of payment. Tests for statistically significant differences are from t-tests between the Stock Only group and other groups. The symbols ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively.

Cash Only Acquirer CARs (%) [-1, 1] [-2, 2] Target CARs (%) [-1, 1] [-2, 2] 0.41 30.71 31.37
***

Mixed Payment -2.08 19.93 20.86

**

Stock Only -3.06 -3.59 19.46 20.25

Overall -1.45 -1.41 23.81 24.60

0.57 ***
*** ***

-1.74 ***

39

Table 6. Determinants of Announcement Returns

This table reports the results of OLS regressions that test the determinants of takeover announcement returns, defined as the cumulative abnormal returns (CARs) over the three-day [-1, +1] and five-day [-2, +2] windows. The dependent variable in Panel A is target CARs, while it is acquirer CARs in Panel B. The measures of differences in dividend policies include One is Payer, Diff.DivYield, and Diff.Div/Book, which are discussed in Section 2.2. The t-statistics are based on White robust standard errors. The symbols ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively. Coefficients on Relative Size have been multiplied by 10,000 for presentation purposes. All of the control variables are discussed in Section 3.2.

Panel A. Target Announcement Returns Dependent Variables Dividend Differential Stock OnlyOne is Payer One is Payer Stock OnlyDiff.DivYield Diff.DivYield Stock OnlyDiff.Div/Book Diff.Div/Book Stock Only Acquirer Characteristics ln(Market Value) Market/Book Leverage PPE/Book Target Characteristics Relative Size Institutional Ownership Leverage Stock Return Volatility Market/Book Intercept -0.058*** [-3.55] 0.019*** [4.08] 0.008 [1.25] -0.071 [-1.64] -0.018 [-0.96] -0.044*** [-3.19] -0.120*** [-3.50] -0.140*** [-4.14] 0.150*** [4.55] -0.009 [-1.47] 0.258*** [5.87] 1,022 8.34 0.000 0.117 -0.059*** [-3.70] 0.019*** [3.82] 0.008 [1.25] -0.073* [-1.67] -0.020 [-1.06] -0.043*** [-3.12] -0.123*** [-3.60] -0.137*** [-4.10] 0.146*** [4.51] -0.009 [-1.50] 0.263*** [5.87] 1,022 11.57 0.000 0.116 Target CAR [-1,1] -0.026** [-2.27] -0.011 [-0.51] -0.029** [-2.02] -0.242 [-0.84] -0.079 [-1.08] -0.049 [-0.52] -0.059*** [-3.71] 0.019*** [3.85] 0.008 [1.27] -0.072* [-1.65] -0.019 [-1.02] -0.043*** [-3.14] -0.122*** [-3.57] -0.138*** [-4.12] 0.146*** [4.52] -0.009 [-1.49] 0.261*** [5.90] 1,022 8.42 0.000 0.116 Target CAR [-2,2] -0.059** [-2.16] -0.003 [-0.12] -0.045** [-2.35] -0.322 [-1.08] -0.509 [-0.85] -0.041 [-0.44] -0.062*** [-3.85] 0.021*** [4.22] 0.007 [1.13] -0.086* [-1.94] -0.018 [-0.95] -0.045*** [-3.23] -0.137*** [-3.96] -0.134*** [-3.99] 0.148*** [4.59] -0.007 [-1.15] 0.267*** [5.80] 1,022 9.11 0.000 0.228

-0.059*** [-3.50] 0.021*** [4.35] 0.007 [1.15] -0.086* [-1.96] -0.018 [-0.92] -0.044*** [-3.16] -0.136*** [-3.92] -0.134*** [-3.95] 0.149*** [4.56] -0.007 [-1.15] 0.264*** [5.77] 1,022 8.95 0.000 0.123

-0.060*** [-3.73] 0.021*** [4.20] 0.007 [1.13] -0.086* [-1.95] -0.018 [-0.94] -0.044*** [-3.15] -0.138*** [-4.00] -0.133*** [-3.96] 0.147*** [4.57] -0.007 [-1.16] 0.268*** [5.74] 1,022 12.37 0.000 0.123

No. observations F test P-value Adj. R2

40

Panel B. Acquirer Announcement Returns Dependent Variables Dividend Differential Stock OnlyOne is Payer One is Payer Stock OnlyDiff.DivYield Diff.DivYield Stock OnlyDiff.Div/Book Diff.Div/Book Stock Only Acquirer Characteristics ln(Market Value) Market/Book Leverage PPE/Book Target Characteristics Relative Size Institutional Ownership Leverage Stock Return Volatility Market/Book Intercept -0.015 ** [-2.41] 0.002 * [1.74] -0.003 [-1.46] 0.009 [0.65] 0.009 * [1.88] -0.011 ** [-2.54] -0.016 ** [1.98] 0.010 [0.96] 0.011 * [1.77] 0.002 [0.96] -0.035 * [-1.90] 1,022 5.57 0.000 0.056 -0.020 *** [-3.53] 0.003 ** [1.97] -0.004 [-1.63] 0.010 [0.71] 0.011 * [1.93] -0.013 *** [-3.07] -0.015 ** [-2.47] 0.006 [0.60] 0.014 ** [2.34] 0.003 [1.15] -0.041 ** [-2.21] 1,022 5.95 0.000 0.051 -0.015 ** [-2.23] 0.009 *** [2.66] -0.113 [-0.98] 0.126 [1.22] -0.008 * [-1.73] 0.060 [1.70] * -0.019 *** [-3.30] 0.003 * [1.90] -0.004 * [-1.78] 0.009 [0.62] 0.010 * [1.70] -0.013 *** [-3.00] -0.016 ** [-2.49] 0.007 [0.65] 0.013 * [1.68] 0.003 [1.14] -0.038 ** [-2.11] 1,022 4.92 0.000 0.049 -0.019 * [-1.68] 0.016 ** [2.20] -0.168 [-1.51] 0.041 [0.43] -0.237 * [-1.81] 0.045 [1.32] -0.029 *** [-4.59] 0.003 * [1.73] -0.005 ** [-2.36] 0.007 [0.46] 0.012 * [1.91] -0.012 ** [-2.28] -0.019 * [-1.73] 0.008 [0.70] 0.008 [1.45] 0.003 [1.57] -0.028 [-1.56] 1,022 4.09 0.000 0.059 Acquirer CAR [-1,1] Acquirer CAR [-2,2]

-0.024 *** [-3.64] 0.002 [1.07] -0.005 ** [-2.07] 0.007 [0.47] 0.012 * [1.77] -0.010 * [-1.91] -0.020 * [-1.81] 0.010 [0.92] 0.006 [1.24] 0.003 [1.38] -0.026 [-1.42] 1,022 4.96 0.000 0.063

-0.029 *** [-4.67] 0.003 * [1.83] -0.005 ** [-2.21] 0.008 [0.55] 0.014 ** [2.13] -0.012 ** [-2.23] -0.019 * [-1.75] 0.007 [0.67] 0.008 * [1.86] 0.003 [1.58] -0.031 * [-1.67] 1,022 4.34 0.000 0.059

No. observations F test P-value Adj. R2

41

Table 7. Two-Stage Analysis of Announcement Returns

This table reports the results of two-stage least squares (2SLS) regressions that test the determinants of takeover announcement returns of both target and acquiring firms. The two-step approach when the interaction term is an endogenous variable is discussed in Section 2.3. In Panel A, One is Payer is used as the measure of dividend differential, while Diff.DivYield and Diff.Div/Book are used in Panel B and Panel C, respectively. These measures are discussed in Section 2.2. The t-statistics based on robust standard errors are reported in parentheses below coefficient values. The symbols ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively. Coefficients on Relative Size have been multiplied by 10,000 for presentation purposes. All of the control variables are discussed in Section 3.2.

Panel A. One is Payer as a Measure of Dividend Differential 1st stage regressions 2nd stage regressions %Stock* Target CARs Acquirer CARs Dependent Variables %Stock [-1,1] [-2,2] [-1,1] [-2,2] One is Payer Dividend Differential %StockOne is Payer -0.012** -0.033* -0.008* -0.007 [-2.02] [-1.78] [-1.88] [-1.65] %Stock -0.133*** -0.138*** -0.060*** -0.070*** [-3.20] [-3.26] [-3.96] [-4.45] One is Payer (1) -0.062* -0.084*** -0.020 -0.020 0.005 0.002 [-1.71] [-3.35] [-0.57] [-0.59] [0.53] [0.19] Acquirer Characteristics ln(Market Value) -0.038*** -0.018*** 0.015*** 0.017*** 0.000 0.000 [-5.42] [-4.41] [3.00] [3.34] [0.12] [-0.09] Market/Book -0.001 0.004 0.009 0.008 -0.003 -0.004* [-0.13] [0.76] [1.46] [1.31] [-1.28] [-1.82] Leverage -0.014 0.010 -0.064 -0.080* 0.011 0.010 [-0.25] [0.43] [-1.49] [-1.83] [0.78] [0.65] PPE/Book -0.040 -0.011 -0.026 -0.025 0.005 0.008 [-1.30] [-0.65] [-1.36] [-1.29] [0.85] [1.18] Target Characteristics Relative Size 0.135*** 0.042*** -0.025* -0.026* -0.002 -0.001 [5.72] [3.25] [-1.67] [-1.68] [-0.45] [-0.14] Institutional Ownership -0.029 0.044* -0.146*** -0.023** -0.026** -0.129*** [-0.62] [1.68] [-3.71] [-4.17] [-2.18] [-2.39] Leverage 0.147*** 0.064*** -0.113*** -0.108*** 0.021* 0.023* [3.00] [2.70] [-3.27] [-3.12] [1.93] [1.92] Stock Return Volatility -0.159*** -0.014 0.135*** 0.134*** 0.003 -0.002 [-3.25] [-0.59] [3.98] [3.99] [0.29] [-0.22] Market/Book 0.024*** -0.001 -0.008 -0.006 0.003 0.004 [3.18] [-0.17] [-1.25] [-0.93] [1.18] [1.53] Instrumental Variables Average Acquirer 0.502*** 0.004 Industry %Stock (2) [5.69] [0.40] Average Target 0.499*** -0.028** Industry %Stock (3) [5.66] [-2.19] (1) (2) -0.016 0.447*** [-0.10] [3.40] (1) (3) 0.073 0.672*** [0.50] [5.87] Intercept 0.260*** 0.082** 0.326*** 0.334*** -0.001 0.010 [3.62] [2.56] [6.52] [6.47] [-0.04] [0.53] No. observations F test P-value Adj. R2 1,022 88.34 0.000 0.372 1,022 29.29 0.000 0.561 1,022 8.50 0.000 11.890 1,022 9.13 0.000 12.420 1,022 5.80 0.000 0.069 1,022 5.52 0.000 0.071

42

Panel B. Diff.DivYield as a Measure of Dividend Differential

1st stage regressions %Stock* %Stock Diff.DivYield 2nd stage regressions Target CARs Acquirer CARs [-1,1] [-2,2] [-1,1] [-2,2] -0.327** [-2.31] -0.129*** [-3.42] -0.126 [-0.36] 0.014*** [2.71] 0.009 [1.45] -0.069 [-1.59] -0.030 [-1.53] -0.023 [-1.57] -0.133*** [-3.84] -0.110*** [-3.18] 0.129*** [3.82] -0.008 [-1.33] -0.260** [-2.08] -0.130*** [-3.40] -0.205 [-0.56] 0.015*** [3.06] 0.008 [1.32] -0.082* [-1.86] -0.028 [-1.40] -0.024 [-1.61] -0.148*** [-4.24] -0.106*** [-3.05] 0.130*** [3.90] -0.007 [-1.01] -0.007 [-1.32] -0.062*** [-4.94] 0.092 [0.90] 0.001 [0.29] -0.003 [-1.27] 0.012 [0.84] 0.006 [1.01] -0.003 [-0.52] -0.022** [-2.12] 0.020* [1.85] 0.004 [0.42] 0.003 [1.25] -0.016 [-1.07] -0.071*** [-5.47] -0.024 [-0.21] 0.000 [-0.02] -0.004* [-1.81] 0.010 [0.71] 0.008 [1.30] -0.001 [-0.16] -0.026** [-2.36] 0.022* [1.90] -0.001 [-0.15] 0.004 [1.59]

Dependent Variables Dividends Differences %StockDiff.DivYield %Stock Diff.DivYield (1) Acquirer Characteristics ln(Market Value) Market/Book Leverage PPE/Book Target Characteristics Relative Size Institutional Ownership Leverage Stock Return Volatility Market/Book Instrumental Variables Average Acquirer Industry %Stock (2) Average Target Industry %Stock (3) (1) (2) (1) (3) Intercept

-0.514 [-0.69] -0.041*** [-6.03] -0.001 [-0.11] -0.019 [-0.32] -0.044 [-1.45] 0.140*** [5.80] -0.035*** [-2.75] 0.154*** [3.12] -0.171*** [-3.56] 0.024*** [3.08] 0.516*** [6.80] 0.511*** [6.78] -2.623 [-0.98] 2.525 [1.02] 0.266*** [3.71] 1,022 84.17 0.000 0.371

-0.157*** [-4.58] 0.000** [-2.15] 0.000 [1.11] -0.001 [-0.63] 0.000 [0.10] 0.001*** [2.72] -0.001 [-1.32] 0.000 [-0.41] 0.000 [-0.18] 0.000 [-0.69] -0.001 [-0.61] -0.001 [-1.48] 0.466** [2.56] 1.044*** [3.18] 0.002** [2.54] 1,022 19.74 0.000 0.857

0.333*** [6.51] 1,022 9.02 0.000 0.118

0.338*** [6.42] 1,022 9.70 0.000 0.125

-0.002 [-0.12] 1,022 6.48 0.000 0.069

0.009 [0.46] 1,022 5.57 0.000 0.070

No. observations F test P-value Adj. R2

43

Panel C. Diff.Div/Book as a Measure of Dividend Differential

1st stage regressions %Stock* %Stock Diff.Div/Book 2nd stage regressions Target CARs Acquirer CARs [-1,1] [-2,2] [-1,1] [-2,2] -0.365 [-1.40] -0.128*** [-3.42] 0.020 [0.07] 0.014*** [2.76] 0.009 [1.51] -0.068 [-1.55] -0.028 [-1.43] -0.023 [-1.55] -0.132*** [-3.82] -0.110*** [-3.18] 0.129*** [3.83] -0.008 [-1.34] -0.286 [-1.22] -0.130*** [-3.42] 0.016 [0.06] 0.016*** [3.11] 0.009 [1.37] -0.082* [-1.83] -0.027 [-1.33] -0.024 [-1.59] -0.147*** [-4.22] -0.106*** [-3.05] 0.130*** [3.91] -0.007 [-1.01] -0.001 [-1.01] -0.061*** [-4.79] 0.033 [0.50] 0.000 [0.25] -0.003 [-1.35] 0.011 [0.79] 0.005 [0.90] -0.002 [-0.50] -0.022** [-2.14] 0.020* [1.88] 0.004 [0.40] 0.003 [1.24] -0.006 [-1.35] -0.070*** [-5.34] 0.003 [0.04] 0.000 [-0.04] -0.004* [-1.86] 0.010 [0.66] 0.008 [1.21] -0.001 [-0.17] -0.026** [-2.36] 0.022* [1.91] -0.001 [-0.15] 0.004 [1.59]

Dependent Variables Dividend Differential %StockDiff.Div/Book %Stock Diff.Div/Book (1) Acquirer Characteristics ln(Market Value) Market/Book Leverage PPE/Book Target Characteristics Relative Size Institutional Ownership Leverage Stock Return Volatility Market/Book Instrumental Variables Average Acquirer Industry %Stock (2) Average Target Industry %Stock (3) (1) (2) (1) (3) Intercept No. observations F test P-value Adj. R2

0.015 [0.05] -0.041*** [-5.96] 0.000 [0.06] -0.017 [-0.30] -0.042 [-1.33] 0.139*** [5.84] -0.033 [-0.72] 0.153*** [3.11] -0.169*** [-3.51] 0.024*** [3.07] 0.542*** [6.62] 0.490*** [5.96] -3.144 [-1.24] 2.488 [1.08] 0.255*** [3.59] 1,022 88.41 0.000 0.372

0.028 [0.28] 0.000*** [-2.62] 0.000 [-0.90] -0.002 [-1.10] 0.000 [-0.25] 0.002*** [3.33] 0.000 [-0.08] 0.003** [2.08] -0.002** [-2.05] 0.000 [-0.45] 0.006** [2.01] -0.005* [-1.78] 0.226*** [2.67] 1.016*** [3.63] 0.003 [1.10] 1,022 106.26 0.000 0.819

0.330*** [6.50] 1,022 8.60 0.000 0.119

0.334*** [6.42] 1,022 9.27 0.000 0.124

-0.001 [-0.05] 1,022 5.61 0.000 0.068

0.010 [0.50] 1,022 5.26 0.000 0.070

44