- Email: [email protected]
The takeover game
Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Contents lists available at ScienceDirect
Journal of Behavioral and Experimental Finance journal homepage: www.elsevier.com/locate/jbef
Full length article
The takeover game Sascha Füllbrunn a,1 , Ernan Haruvy b,∗ a
Radboud University Nijmegen, Institute for Management Research, Department of Economics, Thomas van Aquinostraat 5, 6525 GD Nijmegen, The Netherlands b
University of Texas at Dallas, Jindal School of Management, Richardson, TX 75093, United States
article
info
Article history: Received 6 November 2013 Received in revised form 24 January 2014 Accepted 27 January 2014 Available online 17 February 2014
abstract The takeover game is an experimental asset market characterized by three important features: (1) manager-determined dividend payments to shareholders, (2) a takeover offer to shareholders, and (3) a double auction market in which shareholders trade shares. This market mechanism essentially allows shareholders to price their trust in management. Despite the unique subgame perfect equilibrium outcome in which no takeover offer is ever rejected and no dividends are ever paid out, we find that the market often survives takeover offers. Managers pay positive dividends and appear to do so strategically to signal trustworthiness to shareholders, especially in periods in which takeover is to be made. While prices are not directly responsive to dividends, we find that market price is a good predictor of shareholder’s intent to accept takeover offers. © 2014 Elsevier B.V. All rights reserved.
1. Introduction One of shareholders’ primary powers lies in their ability to accept outside takeover offers (e.g. Dumaine, 1987). In the present work, we focus on the standard tender offer which takes place when an acquiring company makes a public offer for shares. The takeover scheme protects the interest of small non-controlling shareholders (Manne, 1965). Martin and McConnell (1991) find that ‘‘disciplinary takeovers’’ are effective in ridding the firm of managers that do not meet shareholder expectations.2 Liu (2012) suggests that a failed takeover attempt acts as a ‘‘wake-up call’’ to managers to make value-increasing improvements or as mechanisms to replace the incumbent managers. Thus, the takeover threat is an indirect yet powerful mechanism for shareholders to renegotiate their contract with
management (see also Grossman and Hart, 1980; Jensen, 1986; Scharfstein, 1988; Gompers et al., 2003; Bertrand and Mullainathan, 2003; Bebchuk et al., 2009).3 In that sense, Kerschbamer (1998) observes that a takeover threat pressures not only a single firm but also the industry as a whole to align with shareholder interests. Dividend payment to shareholders is another mechanism used to align manager interests with those of investors since a dividend payment reduces the resources controlled by managers (see Rozeff, 1982; Easterbrook, 1984; Jensen, 1986). The dividend mechanism is related to takeover offers in that the takeover threat can persuade management to pay out high dividends. Indeed, dividend payout ratios and propensities typically fall when managers are insulated from takeovers (Francis et al., 2011). A natural benchmark for this situation is the investment game introduced by Berg et al. (1995). In the investment
∗
Corresponding author. Tel.: +1 972 883 4865; fax: +1 972 883 5819. E-mail addresses: [email protected] (S. Füllbrunn), [email protected] (E. Haruvy). 1 Tel.: +31 0 24 36 15474. 2 A shareholder class action is another instrument to discipline management (Humphery-Jenner, 2012). http://dx.doi.org/10.1016/j.jbef.2014.01.002 2214-6350/© 2014 Elsevier B.V. All rights reserved.
3 Some even argue that making firms more vulnerable to hostile takeovers would benefit shareholders (e.g. Bebchuk, 2005). However, others argue that reducing the takeover threat leads to better operating and stock performance (Cen et al., 2011).
86
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
game, an investor (the proposer) invests in a company which triples the value of his investment. The manager (the responder) then decides how much to remit to the investor in a manner akin to a dividend payment, though the manager has the dominant strategy of extracting all the profit. A large literature on evidence on investment game experiments finds that investors place substantial trust in managers, from about 30% to 80% of endowments (see for example Füllbrunn et al. (2011) for an overview of results). Managers reward that trust with a small return, making the expected return on this investment around zero. While the investment game is useful in studying investor–management interaction, the main incentive for dividend payment in the investment game is regard for investors on the part of managers. While it would be nice to believe that managers feel regard for investors and therefore reward them with dividends, we examine the view that dividend payments may be related to managers behaving in their own self-interest. We examine the relationship between dividend payments and the takeover threat using experimental methods. In this environment investors are initially endowed with shares and cash. The shares generate income which can only be distributed to shareholders by the manager, who may also keep much of the income for himself. Shares may be traded on the open market. Investors who do not trust the manager to distribute sufficient income to shareholders via dividends may attempt to sell their shares or may accept an outside tender offer. When a majority of shareholders vote to accept a tender offer, the firm is sold. Thus, while trust between investor and manager is an important element in this environment, direct reciprocity is not as salient as in the standard investment game. Following three periods of trading and income distribution, investors receive their first tender offer. Two more tender offers would be received in future periods if investors reject the preceding offers. Investors face a number of periods in which they can trade, receive dividends and accept or reject tender offers. Thus, there are repeated game interactions. While the subgame perfect equilibrium is zero investment in the finitely repeated game (by backward induction), we know from other finitely repeated games that players are able to sustain some periods of cooperation (e.g. Chaudhuri, 2011). The ability of investors to trade shares gives us insights about their beliefs and intentions that are not available in the standard investment game. Ideally, it would be useful to be able to predict the outcome of a hostile takeover based on share prices and this is one of the issues under investigation. We are interested in several important questions of managerial relevance. First, we would like to know whether dividends are outcomes of strategic considerations by managers. Unlike other settings, here the manager’s decision has personal consequences and we would like to know whether and how the manager incorporates these consequences into his decision. Indeed, we find that managers strategically raise dividends prior to tender offers, and that dividends strategically decline over time. We next want to know how investors react to managers’
actions and whether they adhere to the equilibrium prediction of accepting the first takeover offer. We find that investors are indeed responsive to dividends, and while they do not accept the first takeover offer, they do so with increasing frequency as they gain experience. Lastly, we are interested in the ability of prices to be leading indicators of voting outcomes. We find that prices are informative about future outcomes. 2. Setting A firm has k outstanding shares owned by n traders. Shares are finitely lived productive assets that pay a dividend dt at the end of each trading period and mature in period T . Traders can buy and sell shares in every trading period t 6 T . Each period, the firm realizes randomly distributed earnings Et > 0 with an expectation of µ. A manager determines the allocation of earnings each period through two policy decisions (Oprea, 2008) regarding the dividend amount Dt , the amount of earnings paid out to the investors,4 and the self-dealing amount St , the amount of earnings kept by the manager.5 Et = D t + S t .
(1)
While shareholders know only the expected earnings, µ, only the manager observes realized earnings. Note that in our setting the manager has no influence on earnings. Given the market ends in period T , the value of a share in period t equals the sum of the expected dividend payments E (dt ) = E (Dt )/k until period T as given in (2).
vt =
T −t
E (dτ ).
(2)
τ =1
Apart from the generation of dividends, this setting of the fundamental value is not necessarily different to experiments from Smith et al. (1988). However, prior to the beginning of trading shareholders are able to accept or reject a tender offer to tender their shares at a fixed common known price pˆ t per share. Accepting the offer in period t ′ ends the market immediately, i.e. the manager is fired, and Dt = 0 and St = 0 for t > t ′ . When do traders vote to accept the takeover offer? Theoretically, this is the case if pˆ t > vt . As for a fixed dividend payment and due to the finite nature of the market pt < pt −1 holds, the tender offer price is chosen to decrease over time. Note that we allow for the first tender offer not before period 1 such that the manager is able to signal good will. In the laboratory setting, we set the first takeover decision to occur later than the first period in order for managers to be able to signal their willingness to pay high dividends. Indeed, this signaling or strategic raising of the dividends before the vote is an important question in the present research, and this is one of the primary reasons
4 We use the term dividends to refer to payouts to investors. Payouts to investors may be made in a variety of other ways, including share buybacks. The word dividend is used broadly to refer to all such methods of paying out a portion of the profit to investors. 5 In Oprea (2008) the manager receives additionally a fixed wage in each period. In our setting, the manager’s wage is a fixed total sum.
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
we selected fixed periods for tender offers.6 To summarize: (1) in fixed periods traders are given a takeover offer they may accept or reject before the trading starts. (2) Traders can buy and sell shares in a double auction. (3) A manager decides on dividend payments given realized earnings. Depending on other-regarding preferences, there are two possible equilibria to this market. We consider two manager types. We would expect a selfish equilibrium in which the market cannot be sustained past the first takeover offer for a manager type with only selfish preferences. This behavior may be associated with ‘‘managerial self-dealing’’ such as outright theft from the firm (Liu and Magnan, 2011), excessive compensation (Murphy, 2012), or issues of additional securities (such as equity) awarded to management (Shleifer and Vishny, 1997). We would expect an equitable equilibrium in which the market sustains until T due to the fact that the manager pays sufficiently high dividends as he has other regarding preferences. Proposition 1 (The Selfish Equilibrium). If managers are selfish, they pay no dividends; traders accept the first takeover offer received. The manager maximizes his total earnings by deciding on dividend payments. Thus, the expected payoff in period t equals the sum of kept earnings until period T or until acceptance of a tender offer as given by (3). The indicator variable 1t equals one if pˆ t 6 vt and zero, otherwise.
πtM =
T −t
(Et − Dm ) × 1m .
(3)
m=t
As the manager’s payoff function equals πTM = ET − DT in period T , the manager’s dominant strategy is to choose DT = 0, leaving the shares with zero value. Shareholders are then better off accepting the tender offer in that period as long as pˆ t > 0. As now T − 1 is the last trading period, the same argument holds as for T . Shareholders would not vote to enter period T − 1. This unravels to the very first tender offer. That is, the traders will always accept the first takeover offer and the manager would never pay positive dividends. This backward induction argument yields total payoffs7
π M = E1 ,
π i = pˆ 2 ki1 .
(4)
Proposition 2 (The Equitable Equilibrium). If the manager has sufficient regard for traders, the manager pays a positive dividend and the traders do not accept early takeover offers.
6 In reality, tender offers do not arrive at regular fixed intervals, although one may think of them as looming in every period. The periods of tender offers in the present setting are fixed and known in advance in order to simplify the analysis. Knowing the period of the vote with certainty in advance makes it possible to compute fundamental values and to conduct fairly straightforward tests of manager signaling. 7 Note that we allow the manager to show good will in the first period before the shareholders are able to vote.
87
Assume a ‘‘fair distribution’’ of earnings across market participants to be an equal distribution as in (5). Df =
µn n+1
,
Sf =
µ n+1
.
(5)
Given shareholders belief the manager to have fairness preferences and, thus, to pay fair dividend payments f dt = df the corresponding market value equals vt = (T − t f f +1)d . Note a necessary condition is pˆ t < vt ∀ t. The shareholder payoff equals then just πti = vt kit and a takeover never occurs. The manager should even pay df in the case of zero earnings in a period. This dividend policy is in line with Lintner (1956) who finds that managers are hesitant to make dividend changes that may later need to be reversed. Managers also try to stabilize dividends and avoid dividend cuts. Prices in this scenario equal the fair value, f pt = vt . The corresponding payoffs given equal distribution of shares are
πM = πi =
T ×µ n+1
.
(6)
Given a manager is able to decide between being selfish or ‘‘generous’’ then being selfish yields a higher income if T > n + 1. 3. Experimental implementation The experimental setting reflects the theoretical discussion in that we implement parameters T = 8, n = 3, k = 20, and Et ∼ U {0, 1000, 2000} with µ = 1000. During trading periods, traders can buy and sell shares in a double auction. At the end of every period, the manager decides on how much to keep for himself (St ) and the dividend per share dt = Dt /20 is publicly announced. The manager alone knows the realization of ET . The following limitations for Dt hold for the manager: the manager is not allowed to set Dt > 2000 or Dt > Et + 1000 but otherwise Dt > Et is allowed. However, negative balances will be paid from next period’s earnings as long as the next period’s earnings are positive and thereafter managerial salary can be paid. At the beginning of period 1, all traders are endowed with money and shares, such that there exists an equitable dividend payment that would ensure all participants receive the same final payoff. The manager starts with 4000 francs, while the traders begin with one of three endowment profiles: One trader begins with 4 shares and 4800 francs, one trader begins with 7 shares and 3900 francs, and one trader begins with 9 shares and 3300 francs.8
8 The expected dividend equals 1000 and a fair distribution would be 250 francs to the manager and 750 francs to the traders in every period. This makes a dividend per share of 37.5 francs, i.e. a fair value of 8 × 37.5 = 300 francs per share at the beginning of the market. For an overall fair allocation the initial endowment plus the dividend payments or salary, respectively, must be the same for all traders. Thus, given the final allocation is 6000 francs to all participants every trader receives an initial money endowment of 6000 − 300 × # shares. The manager receives an initial endowment of 6000 − 250 × 8.
88
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
(a) Average dividend per share, by period.
(b) Average dividend per share, by block.
(c) Dividend per share: difference between period 3 and period 2, by market. Fig. 1. Dividend payouts over time. (a) Depicts the average dividends per share by period. (b) Depicts the average dividends per share averaged by block. We divide the periods into three blocks: block 1 = periods 1–3, block 2 = periods 4–5, and block 3 = periods 6–7. (c) Depicts average differences between dividends per share in period 3 and period 2 across markets. The dividend payments (Dt ) can also be seen in the appendix in Table 5.
At the beginning of Periods 4, 6, and 8 traders can vote to either END MARKET, i.e. to accept the tender offer, or CONTINUE. We start with the first takeover offer in period 4 as we want to give the manager an opportunity to signal other regarding preferences. Each share entitles its owner to one vote. If the majority of shares vote to accept, the market ends immediately without further payments to the manager or the shareholders. Each share in that period pays a final price per share to its owner: 150 francs in period 4, 90 in period 6, and 30 in period 8.9 Sixteen subjects participated in each session. Participants were randomly and anonymously arranged in 2 groups of 8 participants. This group constellation remained fixed during the entire experiment. We allow for four repetitions. Before each market, participants within a group were rearranged into 2 market groups of 4 participants such that everyone was in the role of a manager in 1 of the 4 markets and in the role of a trader in the other 3 markets.10
9 The final price equals 80% of the fair value in the corresponding period. 10 Participants never met again as traders in a different market.
4. Experimental results The experiments were conducted at the university of Texas at Dallas with undergraduate business students who had never previously participated in an asset market experiment. The experiment was computerized using zTree (Fischbacher, 2007). In total, there were seven independent cohorts, for a total of 56 participants. Sessions lasted one and a half hours. Participants earned on average about $25. Instructions were read aloud and participants were trained in using the interface before the experiment started. In Appendix A we provide the instructions and Appendix B reports Period-by-period data on price, volume, shares held, and dividends. We now summarize the findings by player role. 4.1. Manager’s strategy Ideally, rational forward looking investors should anticipate the manager’s behavior and price their shares and vote accordingly. Thus, we begin our analysis by looking at the manager’s strategy. From Fig. 1, we see that dividend payments increase in period 3, right before the takeover
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
vote of period 4, and to a lesser extent in period 5, right before the takeover vote of period 6. Thus, the manager appears to be strategically raising dividends in these instances. We find the second instance to be statistically insignificant, but the first instance (immediately prior to the first takeover vote) is marginally significant — onesided pair-wise t-test11 p-value = 0.055 (t = 1.63, df = 27, 7 sessions × 4 markets = 28 observations). We also see this tendency rising across markets (Fig. 1(c)), meaning that as managers accumulate experience they increase the tendency to raise dividends immediately prior to voting. We state this finding as Observation 1. Observation 1. Managers raise dividends prior to the first voting period. Note in Fig. 1 that aside from the periods immediately preceding voting, there seems to be a clear downward trend in dividend payment over time. This is reasonable in that the manager should realistically expect investors are more likely to accept the takeover offer towards the end of the game.12 Thus the manager has a greater incentive to ‘‘defect’’ to the selfish equilibrium as the end game approaches since his ability to sway the vote through generous dividends decreases. We divide the periods into three blocks, corresponding to a vote at the end of each block (periods 1–3, 4–5, and 6–7) to look at the dividend observation prior to tender offers over time. We look at the differences in average dividend per share across voting blocks and find that dividends decline with marginal significance over voting blocks. Significance, using a two-sided Jonckheere–Terpstra-Test is marginal, with p-value of 0.057.13 We state Observation 2. Observation 2. Dividend payments decrease over blocks of time. Combined, Observations 1 and 2 suggest that dividend payments decrease over time. Peaks in periods 3 and 5 indicate some pre-voting strategic increases. The random effects regression in Table 1 jointly accounts for these two observations. The significantly positive pre-vote-dummy coefficient for Period 3 is in line with the t-tests findings reported in Observation 1. That is, managers appear to strategically raise dividends prior to the first vote, and this effect is significant. The regression estimates of Table 1 do not support Observation 2. That is, no significance is indicated for the Market Period coefficient. Even though the Market Period coefficient is negative, it is not significantly different from zero. It is apparent from Fig. 1 why we got a marginally significant observation when comparing blocks of periods (corresponding to the right-hand side of Fig. 1) but not over periods (the left hand side of Fig. 1, as well the regression
11 7 sessions × 4 markets = 28 observations. 12 Although participants may not see the backward induction argument at the beginning of the experiment it may become clearer, at least in period 8. 13 We had 15 markets with 3 blocks to consider in these tests.
89
Table 1 Random effects regression: performance of dividends. Coefficient Market Market period Pre-vote-dummy period 3 Pre-vote-dummy period 5 Pre-vote-dummy period 7 Constant R2
Estimate (Std. error) 0.223 (0.989)
−0.637 (0.583) 5.882∗∗ (2.776) 2.775 (3.400) −1.090 (4.285) 10.09∗∗ (3.471) 0.0499
Random effects panel regression accounting for Session ID (1–7) and Period ID (1–32). The dependent variable is the average dividend per share (DpS ) in each market. Market and Market Period provide information on the time trend while the pre-vote-dummy 3 (5, 7) equals 1 if the period is a pre-vote-period to the first takeover vote (second, third). Standard errors in parentheses; ∗ p < 0.1, ∗∗ p < 0.05.
of Table 1). Despite the decline over blocks of periods, in terms of period-by-period observations, dividends are actually increasing in the first three periods, drop prior to the first vote, and then increase again from period 4 to period 5. By the nature of the design having a takeover possibility, we have more observations in early periods than in later periods, and in those early periods, dividends are increasing, so the lack of significance of negative trend in dividends over time is not surprising. The takeover threat is supposed to be an indirect but effective mechanism for shareholders to renegotiate their contract with management (see also Grossman and Hart, 1980; Jensen, 1986; Scharfstein, 1988). If so, managers should increase their dividends over time to avoid the takeover. However, the results suggest that dividends do not increase over time. Since managers do not learn to increase dividends over time, it seems to call into question the disciplinary role of takeover threats. 4.2. Investor behavior It is important to first characterize what rational investors ought to do with regards to voting. Recall that the takeover offer entitles each share to 150 francs in period 4, 90 in period 6, and 30 in period 8. Given these takeover offers, a rational investor should require a dividend per share of at least 30 per period in order to rationally reject takeover offers. In looking at the manager’s actions the average dividend in any period never exceeds 15. This is nowhere near the threshold dividend of 30 as discussed above. Offers of 30 or higher are nearly never observed. Therefore, theoretically it should never be rational for selfish profit maximizing investors to reject any takeover offer. It is possible that investors have naïve or unrealistic expectations that cause them to reject takeover offers. However, since dividends sharply decline over time, investors should be more likely to reject takeover offers in earlier periods than those in later periods. Thus, we make two predictions: 1. Earlier period takeover offers are more likely to be rejected than later ones. 2. Early takeover offers are less likely to be rejected as investors gain experience.
90
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
(a) Frequency of rejected takeover offers by takeover period.
(b) Average number of rejected takeover offers by market.
Fig. 2. Number of rejected takeovers. (a) We took the average of the frequency of rejected offers. Note that the number of offers depends on whether a takeover period occurs. (b) Average number of rejected takeover offers by market. Investors can reject three times (period 4, 6, 8), 2 times (period 6, 8), one time (period 8) or never.
As Fig. 2 shows, we find some, albeit statistically insignificant, evidence for the first conjecture. We suspect that the lack of significance is due to investor inertia: Conditional on having rejected a takeover offer, investors are more likely to reject again. So if we look at acceptance conditional on having reached a node, simple inertia generates the opposite trend to what we conjectured. Unconditional on reaching the node, however, we observe the downward trend because the majority of investors never reach period 6 and even fewer investors reach period 8. The second claim is strongly supported as can be seen in Fig. 2. The number of rejected takeover offers significantly declines over markets using a Jonckheere–TerpstraTest (2-sided p = 0.018) and a Cuzick Test (one-sided, p = 0.015). We state this as Observation 3.14 Note that Observation 3 calls into question the disciplinary role of takeover threats. If managers respond to takeover threats with increasing dividends, the rejection rate of takeovers should be expected to increase with experience.
Fig. 3. Relationship between dividends per share and subsequent takeover decisions. The figure shows the average dividend payout prior to rejection and prior to acceptance.
Observation 4. Investor’s propensity to accept takeover offers is responsive to dividends.
for vote-by-vote count report of shares held by the top investor) a single shareholder has the majority of shares and is therefore responsible for the voting outcome. However, only three traders followed the strategy to always be the majority shareholder in all markets. While we do not know what the average shareholder expectation is (nor can we infer it from the outcomes due to this selection bias), we can safely say that a property of markets in general is that even if expectations are on average unbiased, the shareholders who end up with the shares are those with the more optimistic expectations (this is the well-known winner’s curse). This bias is likely responsible for the reluctance to accept takeover offers that should be accepted with unbiased expectations.
Finally, it is important to emphasize that in the majority of the cases (65% of all voting instances, see Appendix B
4.3. Market performance
14 7 sessions × 4 markets = 28 observations. 15 The data consists of 56 takeover situations (4 markets × 7 sessions ×
Fig. 4 depicts the average number of executed transactions before the first takeover vote. The figure shows that trades decrease within and across markets. When the average number of executed trades per session and markets is the dependent variable (87 observations), the coefficients
Observation 3. Early takeover offers are less likely to be rejected as investors gain experience. As Fig. 3 suggests, investors are responsive to dividends. That is, higher dividend payments reduce the probability of a takeover offer being accepted. The spearman rank correlation between the average block 1 dividend and the vote to accept the first takeover offer is −0.25 and is significant.15 Thus, we state Observation 4:
2 groups). We correlate dividend payments in periods 1–3 with the vote to accept that is either 1 or 0.
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
91
Fig. 4. Trade volume. The figure shows the average number of executed trades in periods 1 to 3.
Fig. 5. Median prices and fair value (averaged over markets and sessions).
for market ID (coefficient = −0.450, p-value = 0.0003) and period ID (coefficient = −0.527, p-value 0.012) are significant. We state this as Observation 5. Observation 5. Trade volume declines within and across markets. The large number of trades early on could be an indicator of heterogeneous trader beliefs in the early part of a market (in theory, two shareholders need to have opposing beliefs in order to trade with each other). The declining trade volume over time in a market could thus be a consequence of converging trader beliefs over time. To gauge whether the market reflects reasonable valuations, we need to determine a benchmark market value for an asset that reflects some reasonable expectations for dividends. We call this assessed market value the ‘‘fair market value’’. The fair market value equals the ‘‘fair dividend per share’’ times the number of remaining dividend payments. We begin with the assumption that the fair dividend per share is the dividend amount that would equally divide all
firm profits among the four participants (the three traders and the manager). The expected earnings in a period equals 1000: dividing it four ways means that traders would get three-fourths of that amount. Since there are 20 shares, the fair dividend per share would be df = 1000 × 3/4 × 1/20 = 37.50. Hence, the fair market value in period t f equals vt = (9 − t ) × 37.5. Fig. 5 provides the median prices along with the fair price. Overall we find that the market initially overprices assets relative to fair market value. The overpriced markets do not appear to correct themselves. However after about two repetitions, the market closely tracks the fair market value in the first four periods. Thus it appears that markets exhibit a reasonable degree of confidence in future dividend payments. When price is the dependent variable, neither the lag dividend (p-value= 0.431) nor a dummy for a positive lag dividend (p-value= 0.932) are significant as explanatory variables using a random effects model regression. We obtain similar results if we replace price with investor willingness to pay, i.e. bids (p-values 0.766 and 0.917 for
92
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Fig. 6. Relationship between prices and the takeover decisions. The figure shows the average price prior to takeover offers, separately for the two voting outcomes.
vations) suggests significantly higher prices when an offer is about to be rejected (one sided, p = 0.010). The same test for period 6 with the average price of all first 5 periods (in total 21 observations) leads to no significant differences. However, testing the same for the takeover in period 8 using average price from periods 1 to 7 (in total 15 observations) again yields a significant difference (one sided, p = 0.072). The results are in line with evidence shown in the Fig. 6. They are also in line with recent empirical observations in the literature. For example, Edmans et al. (forthcoming) found that an inter-quartile decrease in valuation leads to a 7% point increase in acquisition likelihood, relative to a 6% unconditional takeover probability. Thus, we can state the last finding: Observation 7. Prices indicate investors’ voting outcomes.
lag dividends and lag dummy, respectively). Thus, we can state Observation 6. Observation 6. Prices are not responsive to the manager’s recent dividend payment. Observation 6 is consistent with the extant literature. Shachat and Srinivasan (2011), for example, examined asset markets where a publically observed signal indicated the future value of the share. They found that investors were sensitive to this signal, but the impact of the signal was small relative to the theoretical prediction. In treatments where the signal was observed by only a fraction of the market, its effect dropped further. Thus, to the extent that recent dividends serve as signals about the manager’s intent to pay future dividends, our results are consistent with Shachat and Srinivasan (2011).16 While price may not be fully responsive to dividend payments, we find that price is a good predictor of the investors’ intention to accept the takeover target. Fig. 6 shows the average prices before takeover decisions. It seems that the market can anticipate the result of the voting before the voting takes place. At least in period 3, prices should be at about 150 (plus expected dividend in period 3) if the takeover is predicted to be successful. Otherwise, if traders predict the dividend stream to be sufficiently high, prices should exceed the tender offer price assuming the takeover is not successful.17 The hypothesis is that higher prices reflect the traders’ willingness to reject the takeover. Indeed, prices are higher if the voting is about to end in a rejected takeover offer. The t-test comparing average periods 1–3 prices when accepted (35 observations) to average periods 1–3 prices when rejected (21 obser-
16 There are other examples in the literature documenting investor over-responsiveness to recent information about value (e.g. Shafran et al., 2009). 17 Samualeson and Rosenthal (1986) consider price movements before tender offers. Tender offers are conditional on current market prices. Since they are usually above current market prices, prices tend to increase to the tender offer if a successful takeover is predicted. In our case it is the other way around since takeover offers are even below the fair market valuation.
5. Conclusions We characterized the Takeover Game— a market setting with takeover offers, in the presence of misalignment in interests between managers and shareholders. The takeover feature, along with the open market for shares, allows us to quantify trust using market prices. Despite a unique subgame perfect equilibrium outcome in which no takeover offer is ever rejected and no dividends are ever paid out, we found that the market survived takeover offers. Specifically, 37.5% of the sessions survived the first takeover offer and 27% survived the second. These did not appear to be due to investor mistakes in that they were shown to be sensitive to dividend payments. Managers in turn were cognizant of the relationship between dividends and investor acceptance of takeover offers. They paid positive dividends and appeared to do so for strategic reasons. This is manifested in a dramatic increase in dividends immediately before voting periods. Lastly, we found that market prices were both remarkably rational and informative. Prices appeared to track fair market value and appear to indicate investors’ intentions. While prices were not directly responsive to dividends, although takeover acceptance was responsive, we find that market price was a good predictor of when investors intend to accept the takeover offer. Thus the effect of dividends on prices was indirect and mediated by the effect of dividends on voter intentions which then translated to market prices. Combined, these results suggest that, as conjectured in the literature but not previously shown, the takeover threat seems to translate into potential investor power over management. Specifically, managers anticipate investors’ responsiveness to dividends as evidenced by their strategic pre-voting peaks in dividend payments and investors seem to be responding to such actions. Thus, investors could potentially exercise some disciplinary power over management. However, we found that investors did not fully utilize this potential disciplinary power over management to their advantage in a manner that would result in management taking actions more favorable to investors over time. That is, managers did not learn to increase dividends across markets (Observation 2) and early takeover
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
offers were less likely to be rejected with experience (Observation 3). The implication is that even in the presence of the takeover’s disciplinary power, it does not seem to be the case that investors fully exploit this threat to their advantage. The present study was largely an investigation of a particular setting – the takeover game – rather than a comparison of institutions. Nevertheless, broader implications about the takeover threat as a mechanism for investor control could be drawn. The takeover threat is not the most attractive mechanism for shareholder control because takeovers can be inefficient. In theory, dividends are the primary mechanism by which investors retrieve surplus from the firm, but to generate positive dividend payments one would require a mechanism through which investors can pressure management to award dividends. In the present context, the pressure mechanism was the threat of accepting a takeover offer. Investor revolts that result in a management shake-up are another potential mechanism. This mechanism received much attention following the recent financial meltdown. Attention has also been paid to regulatory curbs on management compensation following the financial meltdown, but we know of no evidence that such regulatory practices have been effective. Before these alternative institutions could be compared in the laboratory, one would need to understand the fundamental issues in conducting such research and the present study is intended to establish some of these benchmarks. Specifically, we highlighted the degree of strategic behavior on the part of managers, responsiveness to managerial actions on the part of investors and the ability of prices to serve as leading indicators for takeover outcomes. Appendix A. Instructions 1. General instructions This is an experiment in the economics of market decision making. If you follow the instructions and make good decisions, you might earn a considerable amount of money, which will be paid to you in cash at the end of the experiment. The experiment will consist of a sequence of trading periods in which you will have the opportunity to buy and sell shares. Money in this experiment is expressed in francs (1000 francs = 80 cents). 2. How to use the computerized market The goods that can be bought and sold in the market are called Shares. On the top panel of your computer screen you can see the Money you have available to buy shares and the number of shares you currently have. If you would like to offer to sell a share, use the text area entitled ‘‘Enter Ask price’’. In that text area you can enter the price at which you are offering to sell a share, and then select ‘‘Submit Ask Price’’. Please do so now. You will notice that eight numbers, one submitted by each participant, now appear in the column entitled ‘‘Ask Price’’. The lowest ask price will always be on the top of that list and will be highlighted. If you press ‘‘BUY’’, you
93
will buy one share for the lowest current ask price. You can also highlight one of the other prices if you wish to buy at a price other than the lowest. Please purchase a share now by highlighting a price and selecting ‘‘BUY’’. Since each of you had put a share for sale and attempted to buy a share, if all were successful, you all have the same number of shares you started out with. This is because you bought one share and sold one share. When you buy a share, your Money decreases by the price of the purchase. When you sell a share, your Money increases by the price of the sale. If you would like to offer to buy a share, use the text area entitled ‘‘Enter Bid price’’. In that text area you can enter the price at which you are offering to buy a share, and then select ‘‘Submit Bid Price’’. Please do so now. You will notice that 8 numbers, one submitted by each participant, now appear in the column entitled ‘‘Bid Price’’. The highest price will always be on the top of that list and will be highlighted. If you press ‘‘SELL’’, you will sell one share for the highest current bid price. You can also highlight one of the other prices if you wish to sell at a price other than the highest. Please sell a share now by highlighting a price and selecting ‘‘SELL’’. Since each of you had put a share for purchase and attempted to sell a share, if all were successful, you all have the same number of shares you started out with. This is because you sold one share and bought one share. You will now have a practice period. Your actions in the practice period do not count towards your earnings and do not influence your position later in the experiment. The goal of the practice period is only to master the use of the interface. Please be sure that you have successfully submitted bid prices and ask prices. Also be sure that you have accepted both bid and ask prices. You are free to ask questions, by raising your hand, during the practice period. 3. Traders and manager The experiment consists of four markets, with each market lasting 8 trading periods. Participant roles. In each market, participants in this room will be organized into trading groups of 4 participants. Three participants in a group are in the role of Trader. The fourth is in the role of Manager. Your role will be indicated shortly on your screen. The groups and the roles change each market. You will be in the role of a Manager in 1 of the 4 Markets and in the role of a Trader in the other 3 Markets. In each market, the groups are reshuffled and you will meet another trader only in one market. Traders. In each period, there will be 90 s for trading, in which Traders may buy and sell Shares, using the mechanics we just practices. There are 20 Shares in the Market. Shares are assets with a life of 8 periods. Shares in your inventory carry over from one period to the next. Traders may receive Dividends for each Share in their inventory at the end of each Period. Manager. The Manager decides on Dividend Payments but cannot trade. At the beginning of each Period, the
94
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Table 2 Closing prices. Group
Subgroup
Market 1 1
1
2
3
4
5
6
7
Group
2
3
4
5
6
7
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
650
550
300
500
500
400
450
450
450
500
300
400
200
250
30
/
2
300
250
250
/
/
/
/
/
350
130
120
360
250
200
150
100
1
400
500
170
/
/
/
/
/
300
300
350
-
500
450
500
-
2
1200
750
600
700
-
500
650
300
350
250
-
/
/
/
/
/
1
500
300
200
/
/
/
/
/
400
400
400
500
550
400
400
500
2
10
50
2000
/
/
/
/
/
250
60
50
/
/
/
/
/
1
-
150
89
/
/
/
/
/
-
100
300
/
/
/
/
/
2
-
700
625
/
/
/
/
/
350
250
-
175
180
120
55
20
1
10
299
149
/
/
/
/
/
105
104
30
/
/
/
/
/
2
200
500
800
500
500
250
750
250
211
190
-
111
101
554
75
/
1
100
200
230
-
250
/
/
/
380
200
200
225
195
190
-
100
2
200
198
100
/
/
/
/
/
100
200
150
/
/
/
/
/
1
750
75
200
-
170
-
99
/
700
900
700
/
/
/
/
/
2
500
-
-
-
-
/
/
/
200
210
280
/
/
/
/
/
Subgroup
Market 3 1
1
Market 2 2
Market 4 2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
200
200
200
200
-
100
100
/
25
300
200
/
/
/
/
/
2
200
105
185
/
/
/
/
/
550
750
500
/
/
/
/
/
1
600
-
-
/
/
/
/
/
50
400
250
/
/
/
/
/
2
200
300
-
200
175
150
190
200
500
-
70
55
-
/
/
/
1
80
250
175
/
/
/
/
/
250
350
150
100
175
/
/
/
2
350
495
350
250
200
350
150
/
450
350
330
/
/
/
/
/
1
200
-
150
/
/
/
/
/
140
155
140
/
/
/
/
/
2
-
80
-
/
/
/
/
/
-
25
60
/
/
/
/
/
/
/
/
/
/
100
85
78
/
/
/
/
/
60
-
/
/
/
100
225
175
/
/
/
/
/
1
200
175
99
2
130
175
110
1
100
100
80
/
/
/
/
/
120
90
-
/
/
/
/
/
2
280
250
170
/
/
/
/
/
100
90
150
/
/
/
/
/
1
500
100
890
/
/
/
/
/
225
70
70
/
/
/
/
/
2
300
200
200
155
110
/
/
/
600
250
300
291
-
250
-
/
The table shows closing prices in each trading period. The symbol ‘‘/’’ indicates that there was no market open in the period since a tender offer had been accepted in some earlier period. The symbol ‘‘-’’ indicates the absence of trade in the period.
Manager receives Earnings that are randomly determined by the computer. Earnings are 2000, 1000, or 0, with equal likelihood for each. Thus, the average Earnings in each period is 1000. The Manager decides on how much of the Earnings in a period to pay as Dividends. The remaining Earnings are the Manager’s Salary and are added to the Manager’s Balance. The manager may choose to pay Dividends that are higher than the Earnings. If he chooses to do so, he will not get a salary for the current period in which dividends exceed earnings. The outstanding balance will be paid from next period’s earnings as long as next period’s earnings are positive and thereafter managerial
salary can be paid. Total dividend payments cannot exceed 2000 and can be no more than 1000 francs above earnings. At the end of the Period, each Share pays a Dividend per Share. The Dividend per Share equals the amount the Manager has decided to pay as Dividends divided by the number of Shares in the Market (i.e. 20 Shares). The Dividend per Share is announced by the manager after trading is finished for the period and is between 0 and 100 francs per share. 4. Vote To END MARKET At the beginning of Period 4, Traders can vote to END MARKET. Each share entitles its owner to one vote. Only if the majority of shares (more than 10 shares) votes to END
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
95
Table 3 Volume. Group
Subgroup
Market 1 1
1
2
3
4
5
6
7
Group
2
3
4
5
6
7
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
9
3
6
8
7
4
6
10
7
6
6
3
2
1
2
/
2
4
5
4
/
/
/
/
/
5
8
9
3
12
5
3
4
1
6
6
4
/
/
/
/
/
9
3
6
0
1
1
1
0
2
5
3
4
4
0
2
2
1
1
1
0
/
/
/
/
/
1
1
3
2
/
/
/
/
/
6
4
3
3
2
1
3
1
2
8
3
5
/
/
/
/
/
6
4
10
/
/
/
/
/
1
0
2
2
/
/
/
/
/
0
1
2
/
/
/
/
/
2
0
1
1
/
/
/
/
/
4
3
0
1
1
3
3
2
1
1
1
1
/
/
/
/
/
1
5
4
/
/
/
/
/
2
3
10
6
4
4
2
6
6
3
3
0
1
1
2
1
/
1
4
4
2
0
1
/
/
/
3
5
2
2
3
1
0
2
2
12
11
7
/
/
/
/
/
17
10
4
/
/
/
/
/
1
5
3
3
0
1
0
1
/
3
3
5
/
/
/
/
/
2
6
0
0
0
0
/
/
/
1
3
2
/
/
/
/
/
4
5
6
7
8
Subgroup
Market 3 1
1
Market 2
2
Market 4
2
3
4
5
6
7
8
1
2
3
1
2
3
3
1
0
1
1
/
8
2
7
/
/
/
/
/
2
3
7
3
/
/
/
/
/
2
1
1
/
/
/
/
/
1
8
0
0
/
/
/
/
/
1
1
2
/
/
/
/
/
2
3
2
0
1
1
2
1
2
4
0
2
1
0
/
/
/
1
9
6
8
/
/
/
/
/
5
7
4
3
6
/
/
/
2
2
5
2
6
3
5
1
/
4
2
2
/
/
/
/
/
1
1
0
2
/
/
/
/
/
2
1
4
/
/
/
/
/
2
0
2
0
/
/
/
/
/
0
1
2
/
/
/
/
/
1
2
2
3
/
/
/
/
/
1
4
3
/
/
/
/
/
2
3
3
1
2
0
/
/
/
1
2
1
/
/
/
/
/
1
6
4
2
/
/
/
/
/
1
4
0
/
/
/
/
/
2
3
2
8
/
/
/
/
/
10
8
4
/
/
/
/
/
1
9
6
1
/
/
/
/
/
2
2
2
/
/
/
/
/
2
7
4
2
1
3
/
/
/
1
3
2
2
0
1
0
/
The table shows the number of trades in each trading period. The symbol ‘‘/’’ indicates that there was no market open in the period since a tender offer had been accepted in some earlier period. The symbol ‘‘-’’ indicates the absence of trade in the period.
MARKET the market ends immediately after the voting. Each share then pays a Final Price of 150 francs to its owner.
No further Dividends are paid to the traders and no further salary is paid to the manager. A similar voting also takes
96
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Table 4 Shares held by the voter with the highest number of shares pre-voting. Group
Subgroup
Market 1 3
1
2
3
4
5
6
7
Market 2 5
Market 3
Market 4
7
3
5
7
3
5
7
3
5
7
1
9
10
9
13
15
16
17
18
18
14
/
/
2
8
/
/
13
13
14
15
/
/
9
/
/
1
12
/
/
13
14
15
12
/
/
10
/
/
2
13
13
13
/
/
10
12
12
7
7
/
1
8
/
/
16
17
19
/
/
10
15
/
2
18
/
/
8
/
/
11
13
16
/
/ /
9
8 11
1
11
/
/
9
/
/
8
/
/
10
/
2
8
/
/
13
12
12
10
/
/
9
/
/
1
8
/
/
16
/
/
10
/
/
17
/
/
2
7
8
11
10
12
14
14
14
/
10
/
/
1
12
13
/
11
15
15
11
/
/
9
/
/
2
12
/
/
14
/
/
12
/
/
12
/
/
1
9
10
9
11
/
/
13
/
/
15
/
/
2
14
14
/
12
/
/
15
15
/
10
11
10
50%
50%
50%
71%
100%
100%
64%
100%
100%
36%
67%
0%
If one pivotal shareholder
The table shows the number of shares held by the voter with the highest number of shares pre voting. A number above 10 indicates that this trader is able to unilaterally accept or reject the tender offer. ‘‘/’’ indicates that a tender offer was accepted in an earlier period. The bottom line shows the frequency of pivotal shareholders.
Table 5 Dividend payout. Group
Subgroup
Market 1 1
1
2
3
4
5
6
7
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
0
0
500
0
0
0
0
0
10
250
140
0
100
0
0
/
100
10
10
/
/
/
/
/
50
100
400
100
250
0
0
0
1
200
200
200
/
/
/
/
/
1000
500
250
250
250
1000
250
0
2
700
800
0
200
500
0
300
0
12
10
10
/
/
/
/
/
1
0
200
0
/
/
/
/
/
200
500
400
1000
1000
300
500
0
2
500
750
500
/
/
/
/
/
0
100
200
/
/
/
/
/
1
100
150
100
/
/
/
/
/
0
50
100
/
/
/
/
/
2
0
0
0
/
/
/
/
/
250
0
500
400
750
500
0
500
1
0
0
1000
/
/
/
/
/
0
0
100
/
/
/
/
/
2
300
50
20
30
25
0
0
0
75
77
22
32
44
26
0
/
1
800
700
700
750
250
/
/
/
20
10
10
10
10
10
10
1
/
/
/
/
/
1000
500
500
/
/
/
/
/
200
100
0
0
/
0
5
3
/
/
/
/
/
0
0
/
/
/
0
0
0
/
/
/
/
/
4
5
6
7
8
2
50
0
50
1
200
100
200
50
0
0
Subgroup
Market 3 1
1
3
2
2 Group
Market 2
2
Market 4
2
3
4
5
6
7
8
1
2
3
1
50
50
150
100
70
75
200
/
10
0
5
/
/
/
/
/
2
1000
500
0
/
/
/
/
/
2
3
20
/
/
/
/
/
/
/
/
/
/
200
0
400
/
/
/
/
/
300
500
100
100
100
100
1900
1900
100
0
/
/
/
/
/
/
/
/
10
0
2
0
3
/
/
/
1
20
0
0
2
100
200
200
1
0
0
0
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
97
Table 5 (continued) Group
Subgroup
Market 1 1
2 4
5
6
7
Market 2
2 150
3 20
15
4
5
6
7
8
40
0
5
4
/
1
2 5
3 0
4
5
6
7
8
500
/
/
/
/
/
1
0
0
0
/
/
/
/
/
5
10
1
/
/
/
/
/
2
700
700
700
/
/
/
/
/
50
1000
2000
/
/
/
/
/
1
1
0
1
/
/
/
/
/
0
0
2
/
/
/
/
/
2
500
800
500
0
200
/
/
/
1
0
1
/
/
/
/
/
1
5
1
1
/
/
/
/
/
0
500
0
/
/
/
/
/
2
2
1
3
/
/
/
/
/
1000
1000
2000
/
/
/
/
/
1
0
0
2000
/
/
/
/
/
1
5
0
/
/
/
/
/
2
0
0
0
0
0
/
/
/
55
45
50
50
45
40
35
/
The table shows the dividend payout to shareholders for each period. ‘‘/’’ indicates that a tender offer was accepted in an earlier period.
place at the beginning of Periods 6 and 8. The Final Price equals 90 francs per share in period 6 and 30 francs per share in period 8. Thus, the expected value of one share is the sum of expected Dividends per Share until the end of the market (either 4, 6, or 8 periods) plus Final Price given a majority vote to END MARKET.
Note again, in every market the composition of traders is new and you never meet the same trader again in another market.
5. Endowment and market payoffs Summing up, a period consists of (1) the trader’s voting to END MARKET or CONTINUE (only in Periods 4, 6 and 8). If the market does not end, then (2) trading follows and (3) the Manager decides on Dividends. At the beginning of each Market, Money and Shares are provided to the Traders such that one Trader has 4 Shares and 4800 francs, another Trader has 7 Shares and 3900 francs, and the third trader has 9 Shares and 3300 francs. At the beginning of each Market, the Manager’s Balance equals 4000 francs. The Trader’s payoff in one market will be the Money at the end of the market, i.e. Money at the beginning, plus Dividends per Share received in each period, plus Money received from sales of Shares, minus Money spent on purchases of Shares, plus the Final Price if the majority votes to END MARKET. The Manager’s payoff in one market will be the Balance at the beginning (4000 francs) plus the sum of all Manager’s salaries from each occurred period in the market. The manager’s per period salary is:
Appendix C. Supplementary data
If there is no outstanding balance Manager Salary in this period = Earnings in this period − Dividends in this period. If there is outstanding balance Manager Salary in this period = 0. If there is a vote to END MARKET Manager Salary in this period and in all future periods in this market = 0. Your final earnings of the experiment will be the sum of your payoff from being a Trader (in 3 markets) and from being a Manager (in 1 market).
Appendix B. Period level data See Tables 2–4.
Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.jbef.2014.01. 002. References Bebchuk, L.A., 2005. The case for increasing shareholder power. Harv. Law Rev. 118 (3), 833–913. Bebchuk, L.A., Cohen, A., Ferrell, A., 2009. What matters in corporate governance? Rev. Financ. Stud. 22 (2), 783–827. Berg, J., Dickhaut, J., McCabe, K., 1995. Trust, reciprocity, and socialhistory. Games Econ. Behav. 10, 122–142. Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and managerial preferences. J. Political Economy 111 (5), 1043–1075. Cen, L., Dasgupta, S., Sen, R., 2011. Discipline or Disruption? Stakeholder relationships and the effect of takeover threat. http://ssrn.com/abstract=1746422 or http://dx.doi.org/10.2139/ssrn.1746422. Chaudhuri, A., 2011. Sustaining cooperation in laboratory public goods experiments: a selective survey of the literature. Exp. Econ. 1, 47–83. Dumaine, B., 1987. A Shareholder revolt at telecom. Fortune www. money.cnn.com/magazines/fortune/fortune_archive/1987/03/02/ 68732/index.htm. Easterbrook, F.H., 1984. Two agency-cost explanations of dividends. Amer. Econ. Rev. 74, 650–659. Edmans, A., Goldstein, I., Jiang, W., 2012. The real effects of financial markets: the impact of prices on takeovers. J. Financ. (forthcoming). Fischbacher, U., 2007. z-Tree — Zurich toolbox for readymade economic experiments. Exp. Econ. 10, 171–178. Francis, B., Hasan, I., John, K., Song, L., 2011. Corporate governance and dividend payout policy: a test using antitakeover legislation. Financ. Manage. 1, 83–112. Füllbrunn, S., Richwien, K., Sadrieh, A., 2011. Trust and trustworthiness in anonymous virtual worlds. J. Media Econ. 24 (1), 48–63. Gompers, P.A., Ishii, J.L., Metrick, A., 2003. Corporate governance and equity prices. Q. J. Econ. 118 (1), 107–155. Grossman, S.J., Hart, O.D., 1980. Takeover bids, the free-rider problem, and the theory of the corporation. Bell J. Econ. 42–64. Humphery-Jenner, M.L., 2012. Internal and external discipline following securities class actions. J. Financ. Intermediation 21 (1), 151–179. Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeover. Amer. Econ. Rev. 76, 323–329. Kerschbamer, R., 1998. Disciplinary takeovers and industry effects. J. Econ. Manage. Strategy 7 (2), 265–306.
98
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Lintner, J., 1956. Distribution of incomes of corporations among dividends, retained earnings and taxes. Amer. Econ. Rev. 46, 97–113. Liu, B., 2012. Revisiting the disciplinary role of failed takeover attempts. http://ssrn.com/abstract=1999499 or http://dx.doi.org/10.2139/ssrn.1999499. Liu, M., Magnan, M., 2011. Self-dealing regulations, ownership wedge, and corporate valuation: international evidence. Corp. Governance: Int. Rev. 19, 99–115. Manne, H.G., 1965. Mergers and the market for corporate control. J. Political Economy 73 (2), 110–120. Martin, K., McConnell, J., 1991. Corporate performance, corporate takeovers, and management turnover. J. Financ. 46, 671–687. Murphy, K.J., 2012. Executive compensation: where we are, and how we got there. In: Harris, Milton, Stulz, René (Eds.), George Constantinides. In: Handbook of the Economics of Finance, Elsevier. Oprea, R., 2008. Free cash flow and takeover threats: an experimental study. South. Econ. J. 75 (2), 351–366.
Rozeff, M., 1982. Growth, beta and agency costs as determinants of dividend payout ratios. J. Financ. Res. 5, 249–259. Samualeson, W., Rosenthal, L., 1986. Price movements as indicators of tender offer success. J. Financ. 41 (2), 481–499. Scharfstein, D.S., 1988. The disciplinary role of takeovers. Rev. Econom. Stud. 55 (2), 185–200. Shachat, J., Srinivasan, A., 2011. Informational price cascades and non-aggregation of asymmetric information in experimental asset markets. Available at SSRN: http://ssrn.com/abstract=1813383 or http://dx.doi.org/10.2139/ssrn.1813383. Shafran, S., Benzion, U., Shavit, T., 2009. Investors’ decision to trade stocks—an experimental study. J. Behav. Financ. 10, 81–88. Shleifer, A., Vishny, R., 1997. A survey of corporate governance. J. Financ. 52, 737–783. Smith, V.L., Suchanek, G.L., Williams, A.W., 1988. Bubbles, crashes, and endogenous expectations in experimental spot asset markets. Econometrica 1119–1151.
Contents lists available at ScienceDirect
Journal of Behavioral and Experimental Finance journal homepage: www.elsevier.com/locate/jbef
Full length article
The takeover game Sascha Füllbrunn a,1 , Ernan Haruvy b,∗ a
Radboud University Nijmegen, Institute for Management Research, Department of Economics, Thomas van Aquinostraat 5, 6525 GD Nijmegen, The Netherlands b
University of Texas at Dallas, Jindal School of Management, Richardson, TX 75093, United States
article
info
Article history: Received 6 November 2013 Received in revised form 24 January 2014 Accepted 27 January 2014 Available online 17 February 2014
abstract The takeover game is an experimental asset market characterized by three important features: (1) manager-determined dividend payments to shareholders, (2) a takeover offer to shareholders, and (3) a double auction market in which shareholders trade shares. This market mechanism essentially allows shareholders to price their trust in management. Despite the unique subgame perfect equilibrium outcome in which no takeover offer is ever rejected and no dividends are ever paid out, we find that the market often survives takeover offers. Managers pay positive dividends and appear to do so strategically to signal trustworthiness to shareholders, especially in periods in which takeover is to be made. While prices are not directly responsive to dividends, we find that market price is a good predictor of shareholder’s intent to accept takeover offers. © 2014 Elsevier B.V. All rights reserved.
1. Introduction One of shareholders’ primary powers lies in their ability to accept outside takeover offers (e.g. Dumaine, 1987). In the present work, we focus on the standard tender offer which takes place when an acquiring company makes a public offer for shares. The takeover scheme protects the interest of small non-controlling shareholders (Manne, 1965). Martin and McConnell (1991) find that ‘‘disciplinary takeovers’’ are effective in ridding the firm of managers that do not meet shareholder expectations.2 Liu (2012) suggests that a failed takeover attempt acts as a ‘‘wake-up call’’ to managers to make value-increasing improvements or as mechanisms to replace the incumbent managers. Thus, the takeover threat is an indirect yet powerful mechanism for shareholders to renegotiate their contract with
management (see also Grossman and Hart, 1980; Jensen, 1986; Scharfstein, 1988; Gompers et al., 2003; Bertrand and Mullainathan, 2003; Bebchuk et al., 2009).3 In that sense, Kerschbamer (1998) observes that a takeover threat pressures not only a single firm but also the industry as a whole to align with shareholder interests. Dividend payment to shareholders is another mechanism used to align manager interests with those of investors since a dividend payment reduces the resources controlled by managers (see Rozeff, 1982; Easterbrook, 1984; Jensen, 1986). The dividend mechanism is related to takeover offers in that the takeover threat can persuade management to pay out high dividends. Indeed, dividend payout ratios and propensities typically fall when managers are insulated from takeovers (Francis et al., 2011). A natural benchmark for this situation is the investment game introduced by Berg et al. (1995). In the investment
∗
Corresponding author. Tel.: +1 972 883 4865; fax: +1 972 883 5819. E-mail addresses: [email protected] (S. Füllbrunn), [email protected] (E. Haruvy). 1 Tel.: +31 0 24 36 15474. 2 A shareholder class action is another instrument to discipline management (Humphery-Jenner, 2012). http://dx.doi.org/10.1016/j.jbef.2014.01.002 2214-6350/© 2014 Elsevier B.V. All rights reserved.
3 Some even argue that making firms more vulnerable to hostile takeovers would benefit shareholders (e.g. Bebchuk, 2005). However, others argue that reducing the takeover threat leads to better operating and stock performance (Cen et al., 2011).
86
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
game, an investor (the proposer) invests in a company which triples the value of his investment. The manager (the responder) then decides how much to remit to the investor in a manner akin to a dividend payment, though the manager has the dominant strategy of extracting all the profit. A large literature on evidence on investment game experiments finds that investors place substantial trust in managers, from about 30% to 80% of endowments (see for example Füllbrunn et al. (2011) for an overview of results). Managers reward that trust with a small return, making the expected return on this investment around zero. While the investment game is useful in studying investor–management interaction, the main incentive for dividend payment in the investment game is regard for investors on the part of managers. While it would be nice to believe that managers feel regard for investors and therefore reward them with dividends, we examine the view that dividend payments may be related to managers behaving in their own self-interest. We examine the relationship between dividend payments and the takeover threat using experimental methods. In this environment investors are initially endowed with shares and cash. The shares generate income which can only be distributed to shareholders by the manager, who may also keep much of the income for himself. Shares may be traded on the open market. Investors who do not trust the manager to distribute sufficient income to shareholders via dividends may attempt to sell their shares or may accept an outside tender offer. When a majority of shareholders vote to accept a tender offer, the firm is sold. Thus, while trust between investor and manager is an important element in this environment, direct reciprocity is not as salient as in the standard investment game. Following three periods of trading and income distribution, investors receive their first tender offer. Two more tender offers would be received in future periods if investors reject the preceding offers. Investors face a number of periods in which they can trade, receive dividends and accept or reject tender offers. Thus, there are repeated game interactions. While the subgame perfect equilibrium is zero investment in the finitely repeated game (by backward induction), we know from other finitely repeated games that players are able to sustain some periods of cooperation (e.g. Chaudhuri, 2011). The ability of investors to trade shares gives us insights about their beliefs and intentions that are not available in the standard investment game. Ideally, it would be useful to be able to predict the outcome of a hostile takeover based on share prices and this is one of the issues under investigation. We are interested in several important questions of managerial relevance. First, we would like to know whether dividends are outcomes of strategic considerations by managers. Unlike other settings, here the manager’s decision has personal consequences and we would like to know whether and how the manager incorporates these consequences into his decision. Indeed, we find that managers strategically raise dividends prior to tender offers, and that dividends strategically decline over time. We next want to know how investors react to managers’
actions and whether they adhere to the equilibrium prediction of accepting the first takeover offer. We find that investors are indeed responsive to dividends, and while they do not accept the first takeover offer, they do so with increasing frequency as they gain experience. Lastly, we are interested in the ability of prices to be leading indicators of voting outcomes. We find that prices are informative about future outcomes. 2. Setting A firm has k outstanding shares owned by n traders. Shares are finitely lived productive assets that pay a dividend dt at the end of each trading period and mature in period T . Traders can buy and sell shares in every trading period t 6 T . Each period, the firm realizes randomly distributed earnings Et > 0 with an expectation of µ. A manager determines the allocation of earnings each period through two policy decisions (Oprea, 2008) regarding the dividend amount Dt , the amount of earnings paid out to the investors,4 and the self-dealing amount St , the amount of earnings kept by the manager.5 Et = D t + S t .
(1)
While shareholders know only the expected earnings, µ, only the manager observes realized earnings. Note that in our setting the manager has no influence on earnings. Given the market ends in period T , the value of a share in period t equals the sum of the expected dividend payments E (dt ) = E (Dt )/k until period T as given in (2).
vt =
T −t
E (dτ ).
(2)
τ =1
Apart from the generation of dividends, this setting of the fundamental value is not necessarily different to experiments from Smith et al. (1988). However, prior to the beginning of trading shareholders are able to accept or reject a tender offer to tender their shares at a fixed common known price pˆ t per share. Accepting the offer in period t ′ ends the market immediately, i.e. the manager is fired, and Dt = 0 and St = 0 for t > t ′ . When do traders vote to accept the takeover offer? Theoretically, this is the case if pˆ t > vt . As for a fixed dividend payment and due to the finite nature of the market pt < pt −1 holds, the tender offer price is chosen to decrease over time. Note that we allow for the first tender offer not before period 1 such that the manager is able to signal good will. In the laboratory setting, we set the first takeover decision to occur later than the first period in order for managers to be able to signal their willingness to pay high dividends. Indeed, this signaling or strategic raising of the dividends before the vote is an important question in the present research, and this is one of the primary reasons
4 We use the term dividends to refer to payouts to investors. Payouts to investors may be made in a variety of other ways, including share buybacks. The word dividend is used broadly to refer to all such methods of paying out a portion of the profit to investors. 5 In Oprea (2008) the manager receives additionally a fixed wage in each period. In our setting, the manager’s wage is a fixed total sum.
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
we selected fixed periods for tender offers.6 To summarize: (1) in fixed periods traders are given a takeover offer they may accept or reject before the trading starts. (2) Traders can buy and sell shares in a double auction. (3) A manager decides on dividend payments given realized earnings. Depending on other-regarding preferences, there are two possible equilibria to this market. We consider two manager types. We would expect a selfish equilibrium in which the market cannot be sustained past the first takeover offer for a manager type with only selfish preferences. This behavior may be associated with ‘‘managerial self-dealing’’ such as outright theft from the firm (Liu and Magnan, 2011), excessive compensation (Murphy, 2012), or issues of additional securities (such as equity) awarded to management (Shleifer and Vishny, 1997). We would expect an equitable equilibrium in which the market sustains until T due to the fact that the manager pays sufficiently high dividends as he has other regarding preferences. Proposition 1 (The Selfish Equilibrium). If managers are selfish, they pay no dividends; traders accept the first takeover offer received. The manager maximizes his total earnings by deciding on dividend payments. Thus, the expected payoff in period t equals the sum of kept earnings until period T or until acceptance of a tender offer as given by (3). The indicator variable 1t equals one if pˆ t 6 vt and zero, otherwise.
πtM =
T −t
(Et − Dm ) × 1m .
(3)
m=t
As the manager’s payoff function equals πTM = ET − DT in period T , the manager’s dominant strategy is to choose DT = 0, leaving the shares with zero value. Shareholders are then better off accepting the tender offer in that period as long as pˆ t > 0. As now T − 1 is the last trading period, the same argument holds as for T . Shareholders would not vote to enter period T − 1. This unravels to the very first tender offer. That is, the traders will always accept the first takeover offer and the manager would never pay positive dividends. This backward induction argument yields total payoffs7
π M = E1 ,
π i = pˆ 2 ki1 .
(4)
Proposition 2 (The Equitable Equilibrium). If the manager has sufficient regard for traders, the manager pays a positive dividend and the traders do not accept early takeover offers.
6 In reality, tender offers do not arrive at regular fixed intervals, although one may think of them as looming in every period. The periods of tender offers in the present setting are fixed and known in advance in order to simplify the analysis. Knowing the period of the vote with certainty in advance makes it possible to compute fundamental values and to conduct fairly straightforward tests of manager signaling. 7 Note that we allow the manager to show good will in the first period before the shareholders are able to vote.
87
Assume a ‘‘fair distribution’’ of earnings across market participants to be an equal distribution as in (5). Df =
µn n+1
,
Sf =
µ n+1
.
(5)
Given shareholders belief the manager to have fairness preferences and, thus, to pay fair dividend payments f dt = df the corresponding market value equals vt = (T − t f f +1)d . Note a necessary condition is pˆ t < vt ∀ t. The shareholder payoff equals then just πti = vt kit and a takeover never occurs. The manager should even pay df in the case of zero earnings in a period. This dividend policy is in line with Lintner (1956) who finds that managers are hesitant to make dividend changes that may later need to be reversed. Managers also try to stabilize dividends and avoid dividend cuts. Prices in this scenario equal the fair value, f pt = vt . The corresponding payoffs given equal distribution of shares are
πM = πi =
T ×µ n+1
.
(6)
Given a manager is able to decide between being selfish or ‘‘generous’’ then being selfish yields a higher income if T > n + 1. 3. Experimental implementation The experimental setting reflects the theoretical discussion in that we implement parameters T = 8, n = 3, k = 20, and Et ∼ U {0, 1000, 2000} with µ = 1000. During trading periods, traders can buy and sell shares in a double auction. At the end of every period, the manager decides on how much to keep for himself (St ) and the dividend per share dt = Dt /20 is publicly announced. The manager alone knows the realization of ET . The following limitations for Dt hold for the manager: the manager is not allowed to set Dt > 2000 or Dt > Et + 1000 but otherwise Dt > Et is allowed. However, negative balances will be paid from next period’s earnings as long as the next period’s earnings are positive and thereafter managerial salary can be paid. At the beginning of period 1, all traders are endowed with money and shares, such that there exists an equitable dividend payment that would ensure all participants receive the same final payoff. The manager starts with 4000 francs, while the traders begin with one of three endowment profiles: One trader begins with 4 shares and 4800 francs, one trader begins with 7 shares and 3900 francs, and one trader begins with 9 shares and 3300 francs.8
8 The expected dividend equals 1000 and a fair distribution would be 250 francs to the manager and 750 francs to the traders in every period. This makes a dividend per share of 37.5 francs, i.e. a fair value of 8 × 37.5 = 300 francs per share at the beginning of the market. For an overall fair allocation the initial endowment plus the dividend payments or salary, respectively, must be the same for all traders. Thus, given the final allocation is 6000 francs to all participants every trader receives an initial money endowment of 6000 − 300 × # shares. The manager receives an initial endowment of 6000 − 250 × 8.
88
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
(a) Average dividend per share, by period.
(b) Average dividend per share, by block.
(c) Dividend per share: difference between period 3 and period 2, by market. Fig. 1. Dividend payouts over time. (a) Depicts the average dividends per share by period. (b) Depicts the average dividends per share averaged by block. We divide the periods into three blocks: block 1 = periods 1–3, block 2 = periods 4–5, and block 3 = periods 6–7. (c) Depicts average differences between dividends per share in period 3 and period 2 across markets. The dividend payments (Dt ) can also be seen in the appendix in Table 5.
At the beginning of Periods 4, 6, and 8 traders can vote to either END MARKET, i.e. to accept the tender offer, or CONTINUE. We start with the first takeover offer in period 4 as we want to give the manager an opportunity to signal other regarding preferences. Each share entitles its owner to one vote. If the majority of shares vote to accept, the market ends immediately without further payments to the manager or the shareholders. Each share in that period pays a final price per share to its owner: 150 francs in period 4, 90 in period 6, and 30 in period 8.9 Sixteen subjects participated in each session. Participants were randomly and anonymously arranged in 2 groups of 8 participants. This group constellation remained fixed during the entire experiment. We allow for four repetitions. Before each market, participants within a group were rearranged into 2 market groups of 4 participants such that everyone was in the role of a manager in 1 of the 4 markets and in the role of a trader in the other 3 markets.10
9 The final price equals 80% of the fair value in the corresponding period. 10 Participants never met again as traders in a different market.
4. Experimental results The experiments were conducted at the university of Texas at Dallas with undergraduate business students who had never previously participated in an asset market experiment. The experiment was computerized using zTree (Fischbacher, 2007). In total, there were seven independent cohorts, for a total of 56 participants. Sessions lasted one and a half hours. Participants earned on average about $25. Instructions were read aloud and participants were trained in using the interface before the experiment started. In Appendix A we provide the instructions and Appendix B reports Period-by-period data on price, volume, shares held, and dividends. We now summarize the findings by player role. 4.1. Manager’s strategy Ideally, rational forward looking investors should anticipate the manager’s behavior and price their shares and vote accordingly. Thus, we begin our analysis by looking at the manager’s strategy. From Fig. 1, we see that dividend payments increase in period 3, right before the takeover
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
vote of period 4, and to a lesser extent in period 5, right before the takeover vote of period 6. Thus, the manager appears to be strategically raising dividends in these instances. We find the second instance to be statistically insignificant, but the first instance (immediately prior to the first takeover vote) is marginally significant — onesided pair-wise t-test11 p-value = 0.055 (t = 1.63, df = 27, 7 sessions × 4 markets = 28 observations). We also see this tendency rising across markets (Fig. 1(c)), meaning that as managers accumulate experience they increase the tendency to raise dividends immediately prior to voting. We state this finding as Observation 1. Observation 1. Managers raise dividends prior to the first voting period. Note in Fig. 1 that aside from the periods immediately preceding voting, there seems to be a clear downward trend in dividend payment over time. This is reasonable in that the manager should realistically expect investors are more likely to accept the takeover offer towards the end of the game.12 Thus the manager has a greater incentive to ‘‘defect’’ to the selfish equilibrium as the end game approaches since his ability to sway the vote through generous dividends decreases. We divide the periods into three blocks, corresponding to a vote at the end of each block (periods 1–3, 4–5, and 6–7) to look at the dividend observation prior to tender offers over time. We look at the differences in average dividend per share across voting blocks and find that dividends decline with marginal significance over voting blocks. Significance, using a two-sided Jonckheere–Terpstra-Test is marginal, with p-value of 0.057.13 We state Observation 2. Observation 2. Dividend payments decrease over blocks of time. Combined, Observations 1 and 2 suggest that dividend payments decrease over time. Peaks in periods 3 and 5 indicate some pre-voting strategic increases. The random effects regression in Table 1 jointly accounts for these two observations. The significantly positive pre-vote-dummy coefficient for Period 3 is in line with the t-tests findings reported in Observation 1. That is, managers appear to strategically raise dividends prior to the first vote, and this effect is significant. The regression estimates of Table 1 do not support Observation 2. That is, no significance is indicated for the Market Period coefficient. Even though the Market Period coefficient is negative, it is not significantly different from zero. It is apparent from Fig. 1 why we got a marginally significant observation when comparing blocks of periods (corresponding to the right-hand side of Fig. 1) but not over periods (the left hand side of Fig. 1, as well the regression
11 7 sessions × 4 markets = 28 observations. 12 Although participants may not see the backward induction argument at the beginning of the experiment it may become clearer, at least in period 8. 13 We had 15 markets with 3 blocks to consider in these tests.
89
Table 1 Random effects regression: performance of dividends. Coefficient Market Market period Pre-vote-dummy period 3 Pre-vote-dummy period 5 Pre-vote-dummy period 7 Constant R2
Estimate (Std. error) 0.223 (0.989)
−0.637 (0.583) 5.882∗∗ (2.776) 2.775 (3.400) −1.090 (4.285) 10.09∗∗ (3.471) 0.0499
Random effects panel regression accounting for Session ID (1–7) and Period ID (1–32). The dependent variable is the average dividend per share (DpS ) in each market. Market and Market Period provide information on the time trend while the pre-vote-dummy 3 (5, 7) equals 1 if the period is a pre-vote-period to the first takeover vote (second, third). Standard errors in parentheses; ∗ p < 0.1, ∗∗ p < 0.05.
of Table 1). Despite the decline over blocks of periods, in terms of period-by-period observations, dividends are actually increasing in the first three periods, drop prior to the first vote, and then increase again from period 4 to period 5. By the nature of the design having a takeover possibility, we have more observations in early periods than in later periods, and in those early periods, dividends are increasing, so the lack of significance of negative trend in dividends over time is not surprising. The takeover threat is supposed to be an indirect but effective mechanism for shareholders to renegotiate their contract with management (see also Grossman and Hart, 1980; Jensen, 1986; Scharfstein, 1988). If so, managers should increase their dividends over time to avoid the takeover. However, the results suggest that dividends do not increase over time. Since managers do not learn to increase dividends over time, it seems to call into question the disciplinary role of takeover threats. 4.2. Investor behavior It is important to first characterize what rational investors ought to do with regards to voting. Recall that the takeover offer entitles each share to 150 francs in period 4, 90 in period 6, and 30 in period 8. Given these takeover offers, a rational investor should require a dividend per share of at least 30 per period in order to rationally reject takeover offers. In looking at the manager’s actions the average dividend in any period never exceeds 15. This is nowhere near the threshold dividend of 30 as discussed above. Offers of 30 or higher are nearly never observed. Therefore, theoretically it should never be rational for selfish profit maximizing investors to reject any takeover offer. It is possible that investors have naïve or unrealistic expectations that cause them to reject takeover offers. However, since dividends sharply decline over time, investors should be more likely to reject takeover offers in earlier periods than those in later periods. Thus, we make two predictions: 1. Earlier period takeover offers are more likely to be rejected than later ones. 2. Early takeover offers are less likely to be rejected as investors gain experience.
90
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
(a) Frequency of rejected takeover offers by takeover period.
(b) Average number of rejected takeover offers by market.
Fig. 2. Number of rejected takeovers. (a) We took the average of the frequency of rejected offers. Note that the number of offers depends on whether a takeover period occurs. (b) Average number of rejected takeover offers by market. Investors can reject three times (period 4, 6, 8), 2 times (period 6, 8), one time (period 8) or never.
As Fig. 2 shows, we find some, albeit statistically insignificant, evidence for the first conjecture. We suspect that the lack of significance is due to investor inertia: Conditional on having rejected a takeover offer, investors are more likely to reject again. So if we look at acceptance conditional on having reached a node, simple inertia generates the opposite trend to what we conjectured. Unconditional on reaching the node, however, we observe the downward trend because the majority of investors never reach period 6 and even fewer investors reach period 8. The second claim is strongly supported as can be seen in Fig. 2. The number of rejected takeover offers significantly declines over markets using a Jonckheere–TerpstraTest (2-sided p = 0.018) and a Cuzick Test (one-sided, p = 0.015). We state this as Observation 3.14 Note that Observation 3 calls into question the disciplinary role of takeover threats. If managers respond to takeover threats with increasing dividends, the rejection rate of takeovers should be expected to increase with experience.
Fig. 3. Relationship between dividends per share and subsequent takeover decisions. The figure shows the average dividend payout prior to rejection and prior to acceptance.
Observation 4. Investor’s propensity to accept takeover offers is responsive to dividends.
for vote-by-vote count report of shares held by the top investor) a single shareholder has the majority of shares and is therefore responsible for the voting outcome. However, only three traders followed the strategy to always be the majority shareholder in all markets. While we do not know what the average shareholder expectation is (nor can we infer it from the outcomes due to this selection bias), we can safely say that a property of markets in general is that even if expectations are on average unbiased, the shareholders who end up with the shares are those with the more optimistic expectations (this is the well-known winner’s curse). This bias is likely responsible for the reluctance to accept takeover offers that should be accepted with unbiased expectations.
Finally, it is important to emphasize that in the majority of the cases (65% of all voting instances, see Appendix B
4.3. Market performance
14 7 sessions × 4 markets = 28 observations. 15 The data consists of 56 takeover situations (4 markets × 7 sessions ×
Fig. 4 depicts the average number of executed transactions before the first takeover vote. The figure shows that trades decrease within and across markets. When the average number of executed trades per session and markets is the dependent variable (87 observations), the coefficients
Observation 3. Early takeover offers are less likely to be rejected as investors gain experience. As Fig. 3 suggests, investors are responsive to dividends. That is, higher dividend payments reduce the probability of a takeover offer being accepted. The spearman rank correlation between the average block 1 dividend and the vote to accept the first takeover offer is −0.25 and is significant.15 Thus, we state Observation 4:
2 groups). We correlate dividend payments in periods 1–3 with the vote to accept that is either 1 or 0.
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
91
Fig. 4. Trade volume. The figure shows the average number of executed trades in periods 1 to 3.
Fig. 5. Median prices and fair value (averaged over markets and sessions).
for market ID (coefficient = −0.450, p-value = 0.0003) and period ID (coefficient = −0.527, p-value 0.012) are significant. We state this as Observation 5. Observation 5. Trade volume declines within and across markets. The large number of trades early on could be an indicator of heterogeneous trader beliefs in the early part of a market (in theory, two shareholders need to have opposing beliefs in order to trade with each other). The declining trade volume over time in a market could thus be a consequence of converging trader beliefs over time. To gauge whether the market reflects reasonable valuations, we need to determine a benchmark market value for an asset that reflects some reasonable expectations for dividends. We call this assessed market value the ‘‘fair market value’’. The fair market value equals the ‘‘fair dividend per share’’ times the number of remaining dividend payments. We begin with the assumption that the fair dividend per share is the dividend amount that would equally divide all
firm profits among the four participants (the three traders and the manager). The expected earnings in a period equals 1000: dividing it four ways means that traders would get three-fourths of that amount. Since there are 20 shares, the fair dividend per share would be df = 1000 × 3/4 × 1/20 = 37.50. Hence, the fair market value in period t f equals vt = (9 − t ) × 37.5. Fig. 5 provides the median prices along with the fair price. Overall we find that the market initially overprices assets relative to fair market value. The overpriced markets do not appear to correct themselves. However after about two repetitions, the market closely tracks the fair market value in the first four periods. Thus it appears that markets exhibit a reasonable degree of confidence in future dividend payments. When price is the dependent variable, neither the lag dividend (p-value= 0.431) nor a dummy for a positive lag dividend (p-value= 0.932) are significant as explanatory variables using a random effects model regression. We obtain similar results if we replace price with investor willingness to pay, i.e. bids (p-values 0.766 and 0.917 for
92
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Fig. 6. Relationship between prices and the takeover decisions. The figure shows the average price prior to takeover offers, separately for the two voting outcomes.
vations) suggests significantly higher prices when an offer is about to be rejected (one sided, p = 0.010). The same test for period 6 with the average price of all first 5 periods (in total 21 observations) leads to no significant differences. However, testing the same for the takeover in period 8 using average price from periods 1 to 7 (in total 15 observations) again yields a significant difference (one sided, p = 0.072). The results are in line with evidence shown in the Fig. 6. They are also in line with recent empirical observations in the literature. For example, Edmans et al. (forthcoming) found that an inter-quartile decrease in valuation leads to a 7% point increase in acquisition likelihood, relative to a 6% unconditional takeover probability. Thus, we can state the last finding: Observation 7. Prices indicate investors’ voting outcomes.
lag dividends and lag dummy, respectively). Thus, we can state Observation 6. Observation 6. Prices are not responsive to the manager’s recent dividend payment. Observation 6 is consistent with the extant literature. Shachat and Srinivasan (2011), for example, examined asset markets where a publically observed signal indicated the future value of the share. They found that investors were sensitive to this signal, but the impact of the signal was small relative to the theoretical prediction. In treatments where the signal was observed by only a fraction of the market, its effect dropped further. Thus, to the extent that recent dividends serve as signals about the manager’s intent to pay future dividends, our results are consistent with Shachat and Srinivasan (2011).16 While price may not be fully responsive to dividend payments, we find that price is a good predictor of the investors’ intention to accept the takeover target. Fig. 6 shows the average prices before takeover decisions. It seems that the market can anticipate the result of the voting before the voting takes place. At least in period 3, prices should be at about 150 (plus expected dividend in period 3) if the takeover is predicted to be successful. Otherwise, if traders predict the dividend stream to be sufficiently high, prices should exceed the tender offer price assuming the takeover is not successful.17 The hypothesis is that higher prices reflect the traders’ willingness to reject the takeover. Indeed, prices are higher if the voting is about to end in a rejected takeover offer. The t-test comparing average periods 1–3 prices when accepted (35 observations) to average periods 1–3 prices when rejected (21 obser-
16 There are other examples in the literature documenting investor over-responsiveness to recent information about value (e.g. Shafran et al., 2009). 17 Samualeson and Rosenthal (1986) consider price movements before tender offers. Tender offers are conditional on current market prices. Since they are usually above current market prices, prices tend to increase to the tender offer if a successful takeover is predicted. In our case it is the other way around since takeover offers are even below the fair market valuation.
5. Conclusions We characterized the Takeover Game— a market setting with takeover offers, in the presence of misalignment in interests between managers and shareholders. The takeover feature, along with the open market for shares, allows us to quantify trust using market prices. Despite a unique subgame perfect equilibrium outcome in which no takeover offer is ever rejected and no dividends are ever paid out, we found that the market survived takeover offers. Specifically, 37.5% of the sessions survived the first takeover offer and 27% survived the second. These did not appear to be due to investor mistakes in that they were shown to be sensitive to dividend payments. Managers in turn were cognizant of the relationship between dividends and investor acceptance of takeover offers. They paid positive dividends and appeared to do so for strategic reasons. This is manifested in a dramatic increase in dividends immediately before voting periods. Lastly, we found that market prices were both remarkably rational and informative. Prices appeared to track fair market value and appear to indicate investors’ intentions. While prices were not directly responsive to dividends, although takeover acceptance was responsive, we find that market price was a good predictor of when investors intend to accept the takeover offer. Thus the effect of dividends on prices was indirect and mediated by the effect of dividends on voter intentions which then translated to market prices. Combined, these results suggest that, as conjectured in the literature but not previously shown, the takeover threat seems to translate into potential investor power over management. Specifically, managers anticipate investors’ responsiveness to dividends as evidenced by their strategic pre-voting peaks in dividend payments and investors seem to be responding to such actions. Thus, investors could potentially exercise some disciplinary power over management. However, we found that investors did not fully utilize this potential disciplinary power over management to their advantage in a manner that would result in management taking actions more favorable to investors over time. That is, managers did not learn to increase dividends across markets (Observation 2) and early takeover
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
offers were less likely to be rejected with experience (Observation 3). The implication is that even in the presence of the takeover’s disciplinary power, it does not seem to be the case that investors fully exploit this threat to their advantage. The present study was largely an investigation of a particular setting – the takeover game – rather than a comparison of institutions. Nevertheless, broader implications about the takeover threat as a mechanism for investor control could be drawn. The takeover threat is not the most attractive mechanism for shareholder control because takeovers can be inefficient. In theory, dividends are the primary mechanism by which investors retrieve surplus from the firm, but to generate positive dividend payments one would require a mechanism through which investors can pressure management to award dividends. In the present context, the pressure mechanism was the threat of accepting a takeover offer. Investor revolts that result in a management shake-up are another potential mechanism. This mechanism received much attention following the recent financial meltdown. Attention has also been paid to regulatory curbs on management compensation following the financial meltdown, but we know of no evidence that such regulatory practices have been effective. Before these alternative institutions could be compared in the laboratory, one would need to understand the fundamental issues in conducting such research and the present study is intended to establish some of these benchmarks. Specifically, we highlighted the degree of strategic behavior on the part of managers, responsiveness to managerial actions on the part of investors and the ability of prices to serve as leading indicators for takeover outcomes. Appendix A. Instructions 1. General instructions This is an experiment in the economics of market decision making. If you follow the instructions and make good decisions, you might earn a considerable amount of money, which will be paid to you in cash at the end of the experiment. The experiment will consist of a sequence of trading periods in which you will have the opportunity to buy and sell shares. Money in this experiment is expressed in francs (1000 francs = 80 cents). 2. How to use the computerized market The goods that can be bought and sold in the market are called Shares. On the top panel of your computer screen you can see the Money you have available to buy shares and the number of shares you currently have. If you would like to offer to sell a share, use the text area entitled ‘‘Enter Ask price’’. In that text area you can enter the price at which you are offering to sell a share, and then select ‘‘Submit Ask Price’’. Please do so now. You will notice that eight numbers, one submitted by each participant, now appear in the column entitled ‘‘Ask Price’’. The lowest ask price will always be on the top of that list and will be highlighted. If you press ‘‘BUY’’, you
93
will buy one share for the lowest current ask price. You can also highlight one of the other prices if you wish to buy at a price other than the lowest. Please purchase a share now by highlighting a price and selecting ‘‘BUY’’. Since each of you had put a share for sale and attempted to buy a share, if all were successful, you all have the same number of shares you started out with. This is because you bought one share and sold one share. When you buy a share, your Money decreases by the price of the purchase. When you sell a share, your Money increases by the price of the sale. If you would like to offer to buy a share, use the text area entitled ‘‘Enter Bid price’’. In that text area you can enter the price at which you are offering to buy a share, and then select ‘‘Submit Bid Price’’. Please do so now. You will notice that 8 numbers, one submitted by each participant, now appear in the column entitled ‘‘Bid Price’’. The highest price will always be on the top of that list and will be highlighted. If you press ‘‘SELL’’, you will sell one share for the highest current bid price. You can also highlight one of the other prices if you wish to sell at a price other than the highest. Please sell a share now by highlighting a price and selecting ‘‘SELL’’. Since each of you had put a share for purchase and attempted to sell a share, if all were successful, you all have the same number of shares you started out with. This is because you sold one share and bought one share. You will now have a practice period. Your actions in the practice period do not count towards your earnings and do not influence your position later in the experiment. The goal of the practice period is only to master the use of the interface. Please be sure that you have successfully submitted bid prices and ask prices. Also be sure that you have accepted both bid and ask prices. You are free to ask questions, by raising your hand, during the practice period. 3. Traders and manager The experiment consists of four markets, with each market lasting 8 trading periods. Participant roles. In each market, participants in this room will be organized into trading groups of 4 participants. Three participants in a group are in the role of Trader. The fourth is in the role of Manager. Your role will be indicated shortly on your screen. The groups and the roles change each market. You will be in the role of a Manager in 1 of the 4 Markets and in the role of a Trader in the other 3 Markets. In each market, the groups are reshuffled and you will meet another trader only in one market. Traders. In each period, there will be 90 s for trading, in which Traders may buy and sell Shares, using the mechanics we just practices. There are 20 Shares in the Market. Shares are assets with a life of 8 periods. Shares in your inventory carry over from one period to the next. Traders may receive Dividends for each Share in their inventory at the end of each Period. Manager. The Manager decides on Dividend Payments but cannot trade. At the beginning of each Period, the
94
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Table 2 Closing prices. Group
Subgroup
Market 1 1
1
2
3
4
5
6
7
Group
2
3
4
5
6
7
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
650
550
300
500
500
400
450
450
450
500
300
400
200
250
30
/
2
300
250
250
/
/
/
/
/
350
130
120
360
250
200
150
100
1
400
500
170
/
/
/
/
/
300
300
350
-
500
450
500
-
2
1200
750
600
700
-
500
650
300
350
250
-
/
/
/
/
/
1
500
300
200
/
/
/
/
/
400
400
400
500
550
400
400
500
2
10
50
2000
/
/
/
/
/
250
60
50
/
/
/
/
/
1
-
150
89
/
/
/
/
/
-
100
300
/
/
/
/
/
2
-
700
625
/
/
/
/
/
350
250
-
175
180
120
55
20
1
10
299
149
/
/
/
/
/
105
104
30
/
/
/
/
/
2
200
500
800
500
500
250
750
250
211
190
-
111
101
554
75
/
1
100
200
230
-
250
/
/
/
380
200
200
225
195
190
-
100
2
200
198
100
/
/
/
/
/
100
200
150
/
/
/
/
/
1
750
75
200
-
170
-
99
/
700
900
700
/
/
/
/
/
2
500
-
-
-
-
/
/
/
200
210
280
/
/
/
/
/
Subgroup
Market 3 1
1
Market 2 2
Market 4 2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
200
200
200
200
-
100
100
/
25
300
200
/
/
/
/
/
2
200
105
185
/
/
/
/
/
550
750
500
/
/
/
/
/
1
600
-
-
/
/
/
/
/
50
400
250
/
/
/
/
/
2
200
300
-
200
175
150
190
200
500
-
70
55
-
/
/
/
1
80
250
175
/
/
/
/
/
250
350
150
100
175
/
/
/
2
350
495
350
250
200
350
150
/
450
350
330
/
/
/
/
/
1
200
-
150
/
/
/
/
/
140
155
140
/
/
/
/
/
2
-
80
-
/
/
/
/
/
-
25
60
/
/
/
/
/
/
/
/
/
/
100
85
78
/
/
/
/
/
60
-
/
/
/
100
225
175
/
/
/
/
/
1
200
175
99
2
130
175
110
1
100
100
80
/
/
/
/
/
120
90
-
/
/
/
/
/
2
280
250
170
/
/
/
/
/
100
90
150
/
/
/
/
/
1
500
100
890
/
/
/
/
/
225
70
70
/
/
/
/
/
2
300
200
200
155
110
/
/
/
600
250
300
291
-
250
-
/
The table shows closing prices in each trading period. The symbol ‘‘/’’ indicates that there was no market open in the period since a tender offer had been accepted in some earlier period. The symbol ‘‘-’’ indicates the absence of trade in the period.
Manager receives Earnings that are randomly determined by the computer. Earnings are 2000, 1000, or 0, with equal likelihood for each. Thus, the average Earnings in each period is 1000. The Manager decides on how much of the Earnings in a period to pay as Dividends. The remaining Earnings are the Manager’s Salary and are added to the Manager’s Balance. The manager may choose to pay Dividends that are higher than the Earnings. If he chooses to do so, he will not get a salary for the current period in which dividends exceed earnings. The outstanding balance will be paid from next period’s earnings as long as next period’s earnings are positive and thereafter managerial
salary can be paid. Total dividend payments cannot exceed 2000 and can be no more than 1000 francs above earnings. At the end of the Period, each Share pays a Dividend per Share. The Dividend per Share equals the amount the Manager has decided to pay as Dividends divided by the number of Shares in the Market (i.e. 20 Shares). The Dividend per Share is announced by the manager after trading is finished for the period and is between 0 and 100 francs per share. 4. Vote To END MARKET At the beginning of Period 4, Traders can vote to END MARKET. Each share entitles its owner to one vote. Only if the majority of shares (more than 10 shares) votes to END
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
95
Table 3 Volume. Group
Subgroup
Market 1 1
1
2
3
4
5
6
7
Group
2
3
4
5
6
7
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
9
3
6
8
7
4
6
10
7
6
6
3
2
1
2
/
2
4
5
4
/
/
/
/
/
5
8
9
3
12
5
3
4
1
6
6
4
/
/
/
/
/
9
3
6
0
1
1
1
0
2
5
3
4
4
0
2
2
1
1
1
0
/
/
/
/
/
1
1
3
2
/
/
/
/
/
6
4
3
3
2
1
3
1
2
8
3
5
/
/
/
/
/
6
4
10
/
/
/
/
/
1
0
2
2
/
/
/
/
/
0
1
2
/
/
/
/
/
2
0
1
1
/
/
/
/
/
4
3
0
1
1
3
3
2
1
1
1
1
/
/
/
/
/
1
5
4
/
/
/
/
/
2
3
10
6
4
4
2
6
6
3
3
0
1
1
2
1
/
1
4
4
2
0
1
/
/
/
3
5
2
2
3
1
0
2
2
12
11
7
/
/
/
/
/
17
10
4
/
/
/
/
/
1
5
3
3
0
1
0
1
/
3
3
5
/
/
/
/
/
2
6
0
0
0
0
/
/
/
1
3
2
/
/
/
/
/
4
5
6
7
8
Subgroup
Market 3 1
1
Market 2
2
Market 4
2
3
4
5
6
7
8
1
2
3
1
2
3
3
1
0
1
1
/
8
2
7
/
/
/
/
/
2
3
7
3
/
/
/
/
/
2
1
1
/
/
/
/
/
1
8
0
0
/
/
/
/
/
1
1
2
/
/
/
/
/
2
3
2
0
1
1
2
1
2
4
0
2
1
0
/
/
/
1
9
6
8
/
/
/
/
/
5
7
4
3
6
/
/
/
2
2
5
2
6
3
5
1
/
4
2
2
/
/
/
/
/
1
1
0
2
/
/
/
/
/
2
1
4
/
/
/
/
/
2
0
2
0
/
/
/
/
/
0
1
2
/
/
/
/
/
1
2
2
3
/
/
/
/
/
1
4
3
/
/
/
/
/
2
3
3
1
2
0
/
/
/
1
2
1
/
/
/
/
/
1
6
4
2
/
/
/
/
/
1
4
0
/
/
/
/
/
2
3
2
8
/
/
/
/
/
10
8
4
/
/
/
/
/
1
9
6
1
/
/
/
/
/
2
2
2
/
/
/
/
/
2
7
4
2
1
3
/
/
/
1
3
2
2
0
1
0
/
The table shows the number of trades in each trading period. The symbol ‘‘/’’ indicates that there was no market open in the period since a tender offer had been accepted in some earlier period. The symbol ‘‘-’’ indicates the absence of trade in the period.
MARKET the market ends immediately after the voting. Each share then pays a Final Price of 150 francs to its owner.
No further Dividends are paid to the traders and no further salary is paid to the manager. A similar voting also takes
96
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Table 4 Shares held by the voter with the highest number of shares pre-voting. Group
Subgroup
Market 1 3
1
2
3
4
5
6
7
Market 2 5
Market 3
Market 4
7
3
5
7
3
5
7
3
5
7
1
9
10
9
13
15
16
17
18
18
14
/
/
2
8
/
/
13
13
14
15
/
/
9
/
/
1
12
/
/
13
14
15
12
/
/
10
/
/
2
13
13
13
/
/
10
12
12
7
7
/
1
8
/
/
16
17
19
/
/
10
15
/
2
18
/
/
8
/
/
11
13
16
/
/ /
9
8 11
1
11
/
/
9
/
/
8
/
/
10
/
2
8
/
/
13
12
12
10
/
/
9
/
/
1
8
/
/
16
/
/
10
/
/
17
/
/
2
7
8
11
10
12
14
14
14
/
10
/
/
1
12
13
/
11
15
15
11
/
/
9
/
/
2
12
/
/
14
/
/
12
/
/
12
/
/
1
9
10
9
11
/
/
13
/
/
15
/
/
2
14
14
/
12
/
/
15
15
/
10
11
10
50%
50%
50%
71%
100%
100%
64%
100%
100%
36%
67%
0%
If one pivotal shareholder
The table shows the number of shares held by the voter with the highest number of shares pre voting. A number above 10 indicates that this trader is able to unilaterally accept or reject the tender offer. ‘‘/’’ indicates that a tender offer was accepted in an earlier period. The bottom line shows the frequency of pivotal shareholders.
Table 5 Dividend payout. Group
Subgroup
Market 1 1
1
2
3
4
5
6
7
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
0
0
500
0
0
0
0
0
10
250
140
0
100
0
0
/
100
10
10
/
/
/
/
/
50
100
400
100
250
0
0
0
1
200
200
200
/
/
/
/
/
1000
500
250
250
250
1000
250
0
2
700
800
0
200
500
0
300
0
12
10
10
/
/
/
/
/
1
0
200
0
/
/
/
/
/
200
500
400
1000
1000
300
500
0
2
500
750
500
/
/
/
/
/
0
100
200
/
/
/
/
/
1
100
150
100
/
/
/
/
/
0
50
100
/
/
/
/
/
2
0
0
0
/
/
/
/
/
250
0
500
400
750
500
0
500
1
0
0
1000
/
/
/
/
/
0
0
100
/
/
/
/
/
2
300
50
20
30
25
0
0
0
75
77
22
32
44
26
0
/
1
800
700
700
750
250
/
/
/
20
10
10
10
10
10
10
1
/
/
/
/
/
1000
500
500
/
/
/
/
/
200
100
0
0
/
0
5
3
/
/
/
/
/
0
0
/
/
/
0
0
0
/
/
/
/
/
4
5
6
7
8
2
50
0
50
1
200
100
200
50
0
0
Subgroup
Market 3 1
1
3
2
2 Group
Market 2
2
Market 4
2
3
4
5
6
7
8
1
2
3
1
50
50
150
100
70
75
200
/
10
0
5
/
/
/
/
/
2
1000
500
0
/
/
/
/
/
2
3
20
/
/
/
/
/
/
/
/
/
/
200
0
400
/
/
/
/
/
300
500
100
100
100
100
1900
1900
100
0
/
/
/
/
/
/
/
/
10
0
2
0
3
/
/
/
1
20
0
0
2
100
200
200
1
0
0
0
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
97
Table 5 (continued) Group
Subgroup
Market 1 1
2 4
5
6
7
Market 2
2 150
3 20
15
4
5
6
7
8
40
0
5
4
/
1
2 5
3 0
4
5
6
7
8
500
/
/
/
/
/
1
0
0
0
/
/
/
/
/
5
10
1
/
/
/
/
/
2
700
700
700
/
/
/
/
/
50
1000
2000
/
/
/
/
/
1
1
0
1
/
/
/
/
/
0
0
2
/
/
/
/
/
2
500
800
500
0
200
/
/
/
1
0
1
/
/
/
/
/
1
5
1
1
/
/
/
/
/
0
500
0
/
/
/
/
/
2
2
1
3
/
/
/
/
/
1000
1000
2000
/
/
/
/
/
1
0
0
2000
/
/
/
/
/
1
5
0
/
/
/
/
/
2
0
0
0
0
0
/
/
/
55
45
50
50
45
40
35
/
The table shows the dividend payout to shareholders for each period. ‘‘/’’ indicates that a tender offer was accepted in an earlier period.
place at the beginning of Periods 6 and 8. The Final Price equals 90 francs per share in period 6 and 30 francs per share in period 8. Thus, the expected value of one share is the sum of expected Dividends per Share until the end of the market (either 4, 6, or 8 periods) plus Final Price given a majority vote to END MARKET.
Note again, in every market the composition of traders is new and you never meet the same trader again in another market.
5. Endowment and market payoffs Summing up, a period consists of (1) the trader’s voting to END MARKET or CONTINUE (only in Periods 4, 6 and 8). If the market does not end, then (2) trading follows and (3) the Manager decides on Dividends. At the beginning of each Market, Money and Shares are provided to the Traders such that one Trader has 4 Shares and 4800 francs, another Trader has 7 Shares and 3900 francs, and the third trader has 9 Shares and 3300 francs. At the beginning of each Market, the Manager’s Balance equals 4000 francs. The Trader’s payoff in one market will be the Money at the end of the market, i.e. Money at the beginning, plus Dividends per Share received in each period, plus Money received from sales of Shares, minus Money spent on purchases of Shares, plus the Final Price if the majority votes to END MARKET. The Manager’s payoff in one market will be the Balance at the beginning (4000 francs) plus the sum of all Manager’s salaries from each occurred period in the market. The manager’s per period salary is:
Appendix C. Supplementary data
If there is no outstanding balance Manager Salary in this period = Earnings in this period − Dividends in this period. If there is outstanding balance Manager Salary in this period = 0. If there is a vote to END MARKET Manager Salary in this period and in all future periods in this market = 0. Your final earnings of the experiment will be the sum of your payoff from being a Trader (in 3 markets) and from being a Manager (in 1 market).
Appendix B. Period level data See Tables 2–4.
Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.jbef.2014.01. 002. References Bebchuk, L.A., 2005. The case for increasing shareholder power. Harv. Law Rev. 118 (3), 833–913. Bebchuk, L.A., Cohen, A., Ferrell, A., 2009. What matters in corporate governance? Rev. Financ. Stud. 22 (2), 783–827. Berg, J., Dickhaut, J., McCabe, K., 1995. Trust, reciprocity, and socialhistory. Games Econ. Behav. 10, 122–142. Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and managerial preferences. J. Political Economy 111 (5), 1043–1075. Cen, L., Dasgupta, S., Sen, R., 2011. Discipline or Disruption? Stakeholder relationships and the effect of takeover threat. http://ssrn.com/abstract=1746422 or http://dx.doi.org/10.2139/ssrn.1746422. Chaudhuri, A., 2011. Sustaining cooperation in laboratory public goods experiments: a selective survey of the literature. Exp. Econ. 1, 47–83. Dumaine, B., 1987. A Shareholder revolt at telecom. Fortune www. money.cnn.com/magazines/fortune/fortune_archive/1987/03/02/ 68732/index.htm. Easterbrook, F.H., 1984. Two agency-cost explanations of dividends. Amer. Econ. Rev. 74, 650–659. Edmans, A., Goldstein, I., Jiang, W., 2012. The real effects of financial markets: the impact of prices on takeovers. J. Financ. (forthcoming). Fischbacher, U., 2007. z-Tree — Zurich toolbox for readymade economic experiments. Exp. Econ. 10, 171–178. Francis, B., Hasan, I., John, K., Song, L., 2011. Corporate governance and dividend payout policy: a test using antitakeover legislation. Financ. Manage. 1, 83–112. Füllbrunn, S., Richwien, K., Sadrieh, A., 2011. Trust and trustworthiness in anonymous virtual worlds. J. Media Econ. 24 (1), 48–63. Gompers, P.A., Ishii, J.L., Metrick, A., 2003. Corporate governance and equity prices. Q. J. Econ. 118 (1), 107–155. Grossman, S.J., Hart, O.D., 1980. Takeover bids, the free-rider problem, and the theory of the corporation. Bell J. Econ. 42–64. Humphery-Jenner, M.L., 2012. Internal and external discipline following securities class actions. J. Financ. Intermediation 21 (1), 151–179. Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeover. Amer. Econ. Rev. 76, 323–329. Kerschbamer, R., 1998. Disciplinary takeovers and industry effects. J. Econ. Manage. Strategy 7 (2), 265–306.
98
S. Füllbrunn, E. Haruvy / Journal of Behavioral and Experimental Finance 1 (2014) 85–98
Lintner, J., 1956. Distribution of incomes of corporations among dividends, retained earnings and taxes. Amer. Econ. Rev. 46, 97–113. Liu, B., 2012. Revisiting the disciplinary role of failed takeover attempts. http://ssrn.com/abstract=1999499 or http://dx.doi.org/10.2139/ssrn.1999499. Liu, M., Magnan, M., 2011. Self-dealing regulations, ownership wedge, and corporate valuation: international evidence. Corp. Governance: Int. Rev. 19, 99–115. Manne, H.G., 1965. Mergers and the market for corporate control. J. Political Economy 73 (2), 110–120. Martin, K., McConnell, J., 1991. Corporate performance, corporate takeovers, and management turnover. J. Financ. 46, 671–687. Murphy, K.J., 2012. Executive compensation: where we are, and how we got there. In: Harris, Milton, Stulz, René (Eds.), George Constantinides. In: Handbook of the Economics of Finance, Elsevier. Oprea, R., 2008. Free cash flow and takeover threats: an experimental study. South. Econ. J. 75 (2), 351–366.
Rozeff, M., 1982. Growth, beta and agency costs as determinants of dividend payout ratios. J. Financ. Res. 5, 249–259. Samualeson, W., Rosenthal, L., 1986. Price movements as indicators of tender offer success. J. Financ. 41 (2), 481–499. Scharfstein, D.S., 1988. The disciplinary role of takeovers. Rev. Econom. Stud. 55 (2), 185–200. Shachat, J., Srinivasan, A., 2011. Informational price cascades and non-aggregation of asymmetric information in experimental asset markets. Available at SSRN: http://ssrn.com/abstract=1813383 or http://dx.doi.org/10.2139/ssrn.1813383. Shafran, S., Benzion, U., Shavit, T., 2009. Investors’ decision to trade stocks—an experimental study. J. Behav. Financ. 10, 81–88. Shleifer, A., Vishny, R., 1997. A survey of corporate governance. J. Financ. 52, 737–783. Smith, V.L., Suchanek, G.L., Williams, A.W., 1988. Bubbles, crashes, and endogenous expectations in experimental spot asset markets. Econometrica 1119–1151.