1
Non-fundamental undervaluation of targets in takeovers:
asymmetric learning or negotiation effectiveness.
Abstract
Merger negotiations should safeguard efficiency in the takeover market. Previous literature finds that the information exchange leads to an inability of acquirers to exploit their own overpricing. In contrast, after a non-fundamental decrease in market value, the probability of becoming the target of a takeover increases. This study investigates whether this decrease gets accounted for in the takeover premia. No association between underpricing, instrumented by mutual fund-driven price pressure, and takeover premia. This suggests that although acquirers are not able to exploit their own overvaluation, they can profit from target underpricing. Therefore, the results are hinting at an inefficiency in the market for corporate control through asymmetric learning between targets and acquirers. Furthermore, alternative ways to profit from underpricing (with effective negotiations) are tested, such as toehold acquisitions and hostility. These mechanisms do not seem to provide significant explanations.
Keywords: Mispricing – Takeovers – Premia – Asymmetric Learning – Merger Negotiations
Master Thesis
University of Amsterdam
Msc Finance: Quantitative Finance
Student: Marius Birkenhäger
Supervisor: Florian Peters
2
Contents
1. Introduction ... 3
2. Literature Review ... 5
A rational market for corporate control? ... 5
Market mispricing and takeovers ... 7
Instrumenting exogenous variation of market prices ... 10
Merger negotiations ... 11
Circumventing management ... 12
Toeholds and underpricing ... 12
3. Data ... 13 Sample construction ... 13 Descriptive statistics ... 13 4. Methodology ... 15 Price pressure ... 16 Baseline model ... 16 Duration of negotiations ... 18 Absence of negotiations ... 18 Toeholds ... 19 4. Results ... 20 First stage ... 20 Instrument validity ... 22 Second stage... 23 Toeholds ... 26 5. Robustness analysis ... 27 Instrument robustness ... 27 Hostile takeovers ... 29
Censored price pressure ... 32
6. Conclusion ... 32
This document is written by Marius Birkenhäger who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
1. Introduction
Activity in the market for corporate control increased substantially over the last decades. The global transaction volume in this market averages 2,71 trillion dollars annually (Yilmaz, Tanyeri, 2016). Moreover, in each of the last four years, the volume of transactions even exceeded 3,3 trillion dollars (Financial Times, 2018). The size of M&A activity has spurred academic activity in the field as well, but there is remains a discrepancy between the impact of this market and the number of academic outcomes (Calipha, Tarba & Brock, 2015). This market should ensure that firms are run by the most efficient management. When a firm’s management disfunctions, the firm would be worth more under another, more capable, management. Subsequently, an acquisition by the firm with a more efficient management would be profitable and adds value to the wider economy. Whether an active market for corporate control creates value in practice remains a topic for discussion in the literature. On average, mergers seem to create value (Eckbo, 2014; Yilmaz, Tanyeri, 2016), but the variation between deals is high and there are a myriad of deals which do not turn out to be beneficial (Moeller, Schlingemann & Stulz, 2005; Calipha, Tarba & Brock, 2015).
The sale of a firm can have a significant impact on an economy. Often there are far-reaching consequences for all stakeholders, it can lead to substantial reorganizations and even alter the competitive landscape within an entire industry (Eckbo, 2014). The scale of single transactions in the market for corporate control is relatively large. One-third of the total market volume in 2017 involved M&A deals where the transaction value exceeded 5 billion dollars (Financial Times, 2018). The sheer size of transactions means that small discrepancies can have large consequences in terms of allocative efficiency. This feature warrants a relatively extensive mechanism in safeguarding accurate transaction prices. Generally, merger negotiations including extensive information exchange should ensure efficient transaction prices. Whether this is the case is a popular subject of study, but remains debated in the literature.
Merger negotiations involve extensive proprietary information exchange in the form of due diligence processes that should safeguard accurate transaction prices. Eckbo, Thompson and Thorburn (2014) assert that merger negotiations show to be a rather sophisticated method of safeguarding accurate transaction prices in the market for corporate control. Due to projected synergies, value creation due to the takeover, firms generally acquire a target at a mark-up (premium) over the market valuation. Furthermore, as rumors of a takeover spread, part of these projected synergies already become reflected in the market price. Takeover mark-ups (premia) account for these target run-ups due to takeover signals. Moreover, this relationship is sensitive to the informativeness of the run-up. In cases where the run-up is uninformative about true synergies, the bidder is compensated almost linearly through a decrease in the mark-up, or premium. Furthermore, Eckbo, Makaew and Thorburn (2017) research
4
whether payment methods indicate bidder opportunism. They conclude that bidders rationally pay with a certain proportion of stock to decrease adverse selection problems, but cannot profit from their own overvaluation. They argue that due to the extensive information exchange, valuation estimations and bargaining any informational asymmetry is resolved. This way, merger negotiations seem to safeguard the allocative efficiency of the market for corporate control.
However, Edmans, Goldstein and Jiang (2012) find that pricing of targets strongly affects takeover activity. As the market price of a firm decreases due to non-fundamental factors, this increases the probability of becoming a takeover target, the trigger effect. They infer this indicates profit opportunities for acquirers and conclude that secondary market prices have a significant effect on the market for corporate control. This effect is thought to be the consequence of asymmetric learning between an acquirer and the target’s shareholders. It could be that the target’s shareholders infer the firm’s value predominantly from its market price while potential acquirers have additional information about the potential value under their management. However, Edmans, Goldstein and Jiang (2012) are inconclusive as to why market prices should impact acquisition likelihood. They do not investigate the mechanism behind the trigger effect, or the way acquirers are able to profit from these decreases in market valuation of targets. The existence of the trigger effect, a higher takeover probability after a non-fundamental decline in target valuation, seems to be contradictory to the effectiveness of merger negotiations in safeguarding accurate transaction prices, as found by e.g. Eckbo, Thompson and Thorburn (2014) and Eckbo, Makaew and Thorburn (2017). In other words, if merger negotiations warrant accurate transaction prices, why would the takeover likelihood increase in case of non-fundamental target underpricing? This raises the question:
To what extent does target underpricing influence transaction premia in takeovers?
Through an investigation of how non-fundamental changes in target valuation influence transaction prices, this study attempts to reconcile the effectiveness of merger negotiations with the existence of profit opportunities for bidders in case of target mispricing. Following Eckbo, Thompson and Thorburn (2014) and Eckbo, Makaew and Thorburn (2017) there should be a correction in takeover premia proportional to the target underpricing. But if this is the case, there should be another way acquirers can profit from underpriced targets following the results of Edmans, Goldstein and Jiang (2012). Therefore, this study also considers situations where acquirers can gain from underpricing while merger negotiations are still effective, i.e. through toeholds and hostile bids.
Following Coval and Stafford (2007), Edmans, Goldstein and Jiang (2012) and Eckbo, Makaew and Thorburn (2017) price pressure following large mutual fund redemptions is used to instrument for non-fundamental declines in market valuation. The next section discusses prior literature on the subject. Explicit attention is payed to the difference in results between studies following different approaches to address endogeneity problems. The section is concluded by formulating several hypotheses. Section 3 describes the dataset. This is followed by an explanation of the methodology and
5
the various regression specifications. In the results section the main findings are discussed, which is followed by several robustness checks. These checks assess for example how the results change if hostile bids are excluded. Also, the effectiveness of the instrument is evaluated. Finally, section 6 draws some tentative conclusions, and discusses potential improvements for future research.
2. Literature Review
A rational market for corporate control?
According to the managerial competition model, the market for corporate control should ensure that firms are run by the most competent management (Eckbo, 2014). This model entails that the management of a firm can notice that another firm is run relatively inefficient. Consequently, the firm would be worth more under their management, which means that an acquisition would increase the value of the target firm. This mechanism theoretically clears any substantial discrepancies between the potential, and current value of firms, stimulating efficiency in the economy and maximizing welfare in society. Hence, the existence of the market for corporate control has a disciplinary role on any management (Eckbo, Thorburn & Makaew, 2017). The threat of a takeover prevents suboptimal actions by a firm’s management.
Rational actors are essential for this mechanism in the takeover market to stimulate efficiency in the economy. Actors, especially the management of firms, should know the value of a firm and the approximate potential value under a different management. To what extent actors in the takeover market are able to act rationally is a widely studied and debated topic (Baker, Ruback & Wurgler, 2013). Generally, two strands of research are distinguished within this field of research on rationality in the market for corporate control and governance, called behavioral corporate finance (Baker, Ruback & Wurgler, 2013). One is based on the irrationality of investors and the effects of their actions on market prices. This approach assumes that investors are less than rational, and that there are some limits to arbitrage. Consequently, investors’ irrational behavior has some influence on market prices which is independent of the fundamental value of the underlying firm. Another strand focuses on the irrationality of management. This approach studies the effects of biases and non-rational decision making by management. This strand focuses on situations where the manager believes he/she is making optimal decisions, but in fact some bias or lack of information impedes the manager from acting rational. This is qualitatively different from studies of corporate governance issues where management does not act in the interest of shareholders, but does act rational from a personal perspective (e.g. moral hazard). Most studies confirm that less than rational investors are a more reasonable approximation of reality (Coval & Stafford, 2007; Malcom, Ruback & Wurgler, 2007). However, the extent to which findings corroborate the notion of irrational management seems to vary.
Irrationality of investors, combined with some limits to arbitrage, causes mispricing (misvaluation) in secondary markets. Mispricing is defined as any deviation of the market price from
6
the informationally efficient value (Coval & Stafford, 2007). There is a myriad of examples of mispricing, but one that is particularly important for this study is the effect of mutual fund redemptions found by Coval and Stafford (2007). They found that forced mutual fund sales and purchases significantly affect market prices. Managers of mutual funds employ certain methods and strategies when choosing the allocation of the fund. However, in case of inflows or redemptions by clients, managers have limited discretion to expand or decrease positions in response due to time constraints. Therefore, they tend to decrease or expand all existing positions accordingly. This uninformed shock to demand in secondary markets creates a price pressure on these securities that is independent of the firm’s fundamental value. Coval and Stafford (2007) show that the price changes due to mutual fund flows are significant and persistent. Because these movements are independent of fundamental value, they can be classified as a market inefficiency.
The second strand of behavioral corporate finance assumes that secondary markets are efficient, but managers are not. Managers can be subject to certain biases that leads them to making suboptimal decisions, while believing to maximize firm value (Baker, Ruback & Wurgler, 2013). For these biases to have impact on the economy, corporate governance has to be limited. This is not a stringent assumption. Managers generally have some discretion with respect to their actions. Most research that assumes managers to be irrational focuses on the biases of overconfidence (hubris), optimism, reference points and bounded rationality (Baker, Ruback & Wurgler, 2013). This last form, bounded rationality, is somewhat different with respect to the assumptions about managerial behavior. Unlike with other biases, bounded rationality is defined as a situation where an actor behaves optimally, but where information is not complete (Dhir & Mital, 2012). In these situations, managers make optimal decisions given the specific subset of information that is available to them. Imperfect information and high cost of information gathering seem to be sensible assumptions (Baker, Ruback & Wurgler, 2013). Managers regularly seem to cope with the complexity of the real world using certain rules of thumb. Bounded rationality could explain why some managers infer firm value predominantly from secondary market prices.
Whether market inefficiencies potentially affect the real economy often combines these two strands of research. First, there have to be some limits to arbitrage and irrationality in investors for any secondary market inefficiency to persist. Secondly, inefficiencies affecting the real economy often requires a situation where managers are not able to fully identify or act on them. The extent to which such an inefficiency in secondary markets affects the real economy has been studied by several researchers. Generally, findings indicate that the real economy is to some extent affected by conditions in secondary markets (Eckbo, 2014). However, researchers seem to find varying magnitudes of these effects. Cremers, Nair and John (2009) find that firms that are exposed to potential takeover attempts have different rates of returns, which partly reflects a profitable exit strategy for shareholders. However, according to the authors the difference in returns also reflects an extra cyclical aspect due to the fact that takeover activity is in part determined by equity market conditions. The effect of secondary-market
7
prices on the real economy is usually ascribed to their informational role (Bond, Edmans & Goldstein, 2011). Market prices reflect information about a firm that can be useful in decision making. It seems that managers learn from secondary market prices and use this to guide their decisions. This is based on the assumption that managers do not have perfect information (in other words, bounded rationality) and that there exist third parties that are in possession of useful information (Bond, Edmans & Goldstein, 2011). At first glance, the assumptions that managers are boundedly rational and some third parties have useful information does not seem stringent. However, Jenter (2004) found that generally the management of a firm has superior information about their own firm compared to the market. Firm managers are able to generate significant abnormal returns. This superior information could render them able to recognize (significant) mispricing (Baker, Ruback & Wurgler, 2013).
Market mispricing and takeovers
For secondary markets to have significant influence on transaction prices in the takeover market, at least in negotiated takeovers, target or acquirer management should be less than rational (e.g. boundedly rational), not fully identifying mispricing. Moreover, there should be an asymmetry between target and acquirer management with respect to the way they value their firms and assess synergies. E.g. to what extent they infer firm value from prices in secondary markets. Therefore, studies that specifically research effects of secondary markets on the economy through the market for corporate control usually fall within both strands of behavioral corporate finance. Some investor irrationality and limits to arbitrage are required for mispricing to exist. However, for this mispricing to have an effect on takeovers, also some assumption involving less than perfect rationality regarding management or a significant corporate governance issue is required. Mispricing is combined with some bias or bounded rationality of managers.
Studies of this subject can be grouped with respect to the way mispricing is measured. Significant mispricing provides an arbitrage opportunity for investors, a chance to make a profit without being exposed to (large) risks. Therefore, in most instances deviations from fundamental value are corrected relatively quickly by market mechanics. Hence, mispricing is challenging to measure (Baker, Ruback & Wurgler, 2013). Most studies measure mispricing by some form of valuation of a firm compared with a group of peers, sometimes augmented with forecasts from analysts (e.g. Rhodes-Kopf, Robinson & Viswanathan, 2005; Dong, Hirshleifer, Richardson & Teoh, 2006; Coakley, Gazzaz & Thomas, 2017). However, there seem to be significant endogeneity issues when measuring mispricing this way. Several potential problems could arise, which include omitted variable bias and measurement error. In essence, the main problem arises due to the informational role of secondary market prices. A deviation of a firm’s market price from any group of peers, or previous market price, could be caused by a justifiable, firm-specific situation. There could be certain factors, e.g. technologies employed that render a firm relatively (un)attractive to a takeover (relative to its peers) and also influence its market valuation (Eckbo, Makaew & Thorburn, 2017). Technology shocks can increase the value of a firm and
8
potential synergies. It is not feasible to account for the variety of unobservable factors involved. Furthermore, raw market valuation is not the same as the valuation discount, the difference between potential value and current value (Edmans, Goldstein & Jiang, 2012). Eckbo (2014) argues that studies of premia that do not adjust for possible endogeneity should be interpreted with caution. He notes that more studies of the behavior of offer premia are warranted. Mutual fund redemptions prove to be a relatively appropriate way to isolate exogenous variation in market valuations. When clients of these funds demand (partial) redemption of their stake in a fund, the manager’s discretion is limited due to time pressure (Coval & Stafford, 2007). Due to this limited discretion, it is not likely that these trades reflect information regarding firm value. Moreover, because of these flows, prices of stocks deviate from their fundamental value for a prolonged period.
Studies that do not account for endogeneity generally find that mispricing in secondary markets significantly affects the real economy through the takeover market. In case of negotiations this requires that target or acquirer management engages in pursuing private benefits or is not able to act fully rational. Shleifer and Vishny (2003) develop a model where stock market valuations are the key drivers of merger activity. They argue that the central prediction of the managerial competition model, profitability of mergers, does not resonate with the empirical reality. Their model is based on the assumption that secondary markets are inefficient, but managers are fully rational. Overvalued firms acquire (less overvalued) targets with their own stock, which is a relatively profitable strategy for their shareholders. They substantiate these assertions with correlations between merger waves, valuations and payment methods. Rhodes-Kopf, Robinson and Viswanathan (2004) try to explain the correlation between valuations and merger activity where both target and acquirer management act on their specific private information. They assume that managers of acquirers have information on both the stand-alone value of their firm and the potential value of the target in case of a merger, whereas target management only has information on the stand-alone value of their firm. They also assume that some investor irrationality is present and that mispricing can be market-wide or firm specific. Because target management can only assess whether they are overvalued, but not the source of this overvaluation, they tend to view bids as relatively attractive in case of market-wide overvaluation. In case of market-wide overvaluation, bids tend to be higher, and targets ascribe a relatively large part to the potential synergies they infer from the bid. This idea reflects that bounded rationality leads to an asymmetry between the target and the acquirer with respect to their ability to assess synergies. They conclude that targets are more likely to overestimate synergies in overvalued markets due to limited information. This mechanism leads to a situation where misvaluation affects merger activity. Rhodes-Kopf, Robinson and Viswanathan (2004) present a mathematical model of their ideas, but do not test its implications. Dong, Hirshleifer, Richardson and Teoh (2006) conclude that managers value firms rationally while this is not the case with investors. They introduce the misvaluation hypothesis, where bidders undertake effort to buy undervalued targets or buy less overvalued firms with their own equity. Their proxies for misvaluation are the market-to-book (M/B) ratio, which entails (secondary) market value of equity
9
divided by the book value of equity, and the price-to-residual income value, which also includes income forecasts. The study claims to find evidence for both the misvaluation hypothesis and a variation of the managerial competition model, the Q hypothesis. They conclude that secondary market valuations significantly affect activity in the takeover market. When targets are undervalued, the manager has an incentive to drive up the price when facing a takeover, resulting in higher premia or tender offers (in cash) to bypass the management. In contrast, when targets are overvalued, they find that managers seem more willing to ‘cash out’, by accepting payment in relatively overpriced equity of the (overvalued) bidder. Therefore, they view (target) management as being rational but pursuing private benefits when possible. Ben-David, Drake and Roulstone (2013) assume that the short interest in a firm’s stock reflects beliefs about mispricing. Short sellers only take positions in overvalued markets. As short selling is costly, this provides useful information about the valuation of firms. They find that misvaluation is a significant determinant of decision making regarding mergers. Specifically, the payment method of overvalued firms (in top quintile of short interest) comprises significantly more stock. However, this mechanism does require, at least in negotiations, that speculative short sellers (in aggregate) have more information on the acquirer than target management gains in a due diligence process. Furthermore, the resulting price pressure from short interest has to be lower than the overvaluation, which means that these speculators have limited resources or information. It seems that in most studies that do not account for the endogeneity of market prices, misvaluation does affect the takeover market to some extent. The alleged reasons for this relationship vary from bounded rationality of target management (Rhodes-Kopf, Robinson & Viswanathan, 2004; Ben-David, Drake & Roulstone, 2013) to ‘cashing-out’ of management (Shleifer & Vishny, 2003; Dong, Hirshleifer, Richardson & Teoh, 2006).
In contrast to previous papers that do not take into account endogeneity, Coakley, Gazzaz and Thomas (2017) find that overvalued targets receive lower premia, suggesting a rational bidding strategy that takes into account misvaluation. They argue that misvaluation only affects the offer premium in case of overvaluation, hinting on effective corrections in negotiations by bidders. Furthermore, psychological reference points (the high price in the past weeks) do not seem to influence premia. Betton, Eckbo, Thompson and Thorburn (2014) study whether run-ups in stock price prior to mergers get accounted for in transactions. They argue that price run-ups are the consequence of rumors that increase market price reflecting the implied deal synergies in combination with the probability that the deal is completed. The authors conclude that these run-ups are identified as a consequence of the deal and are therefore accounted for in the mark-up (premium). Moreover, market driven changes in target value, which are arguably exogenous to the takeover synergy gains, are passed through to the target, therefore increase the mark-up. Betton, Eckbo, Thompson and Thorburn (2014) also argue that if bids are motivated by undervaluation, which is subsequently corrected for through a run-up, this run-up does not revert in case of failure. The paper suggests that managers (and investors) have a relatively
10
sophisticated grasp on the sources of changes in market price and fundamental firm value. Therefore, any mispricing is ought to be corrected in negotiations between target and acquirer.
Instrumenting variation of market prices
As mentioned, mutual fund redemptions provide a way to measure variation in prices that is exogenous to the merger process. Therefore, in case of manager rationality, any change in stock price due to mutual fund redemptions should be corrected for in transaction prices. Both Eckbo, Makaew and Thorburn (2017) and Edmans, Goldstein and Jiang (2012) exploit this feature to instrument variation in market valuations. The former find that there does not seem to be compelling empirical evidence of bidder opportunism. Unlike studies that do not account for endogeneity in valuations (Shleifer & Vishny, 2003; Rhodes-Kopf, Robinson & Viswanathan, 2005; Dong, Hirshleifer, Richardson & Teoh, 2006), bidders do not appear to adapt their payment method as a consequence of their own mispricing. If acquirers structurally exploit their overpricing to pay for targets, the proportion of stock payment should be relatively low when a bidder is underpriced. They conclude that the payment structure reflects rationality, and the proportion of stock is most likely determined by a desire to avoid adverse selection problems.
Edmans, Goldstein and Jiang (2012) find that the probability of a takeover increases with exogenous price pressure on the target’s shares. They conclude that this is an indication of an asymmetry in learning between an acquirer’s management and the shareholders of a target. This entails that acquirers have superior information on the firm’s potential value under their management, while target shareholders infer this value from the current market valuation. Consequently, target shareholders demand a takeover price that is close to the market valuation. Edmans, Goldstein and Jiang (2012) imply that target management does not try to correct for this underpricing as shareholders demand a price close to the market valuation. However, they do not specify what incentive the target management has to accept the offer, except for their own possible bounded rationality. It could be the case that secondary-market mispricing leads to a less efficient secondary-market for corporate control if also target management infers firm value predominantly from stock prices. However, the results from Eckbo, Makaew and Thorburn (2017) imply that bidder mispricing is taken into account in merger negotiations. It seems counter intuitive that target underpricing is not taken into account during the same due diligence process. Khan, Kogan and Serafeim (2012) do find evidence that firm managers are able to recognize and exploit their own overpricing. They use the reverse of the construct used by Eckbo, Makaew and Thorburn (2017), mutual fund purchases after significant inflows, to identify overvaluations. They find that there is a higher probability of SEO’s after such purchases. However, the instrument of purchases is arguably less exogenous that mutual fund sales following redemptions. In case of capital inflows, mutual fund managers have relatively less urgency in their actions, and probably more discretion with respect to choices in assets.
11
Merger negotiations
If there is an asymmetry in learning between the management of acquiring firm and the shareholders of a target, negotiations and the due diligence process should safeguard that the transaction price reflects the fundamental value of a firm and (at least a portion) the synergies between the companies. Targets can either set up an auction to stimulate competition between bidders or enter into negotiations with only one bidder immediately (Aktas, de Bodt & Roll, 2009; Eckbo, 2014). An auction is a situation where more than one bidder has started a due diligence process and places a non-binding offer (Chira & Volkov, 2015). The board of the target faces pressure from shareholders to obtain a high price in both situations. It turns out that entering into negotiations with only one bidder or setting up an auction provide similar returns for shareholders (Xie, 2010). In case of one-on-one negotiations, there are still potential offers from outside bidders. There is a latent auction in every takeover due to these potential offers (Betton, Eckbo & Thorburn, 2009; Eckbo, 2014). Setting up an auction is costly, and in case of low interest, a failed auction, premia are significantly lower (Chira & Volkov, 2015). A profit opportunity in case of significant underpricing is only plausible if there is low competition or asymmetry between bidders (Xie, 2010).
During negotiations, the target and bidder gather information about each other in a due diligence process that should provide a clear understanding of, among others, the true potential synergies between the firms. As the negotiation period increases, additional information on the companies is gathered. Consequently, the negotiating parties receive more information about the real synergies of the takeover (Calcagno, de Bodt & Demidova, 2017). It is therefore relatively probable that any valuation error in either of the companies is accounted for in the transaction price as the duration of negotiations increases. Earlier announcement, i.e. longer negotiation period, seems to also increase the premium independent of the valuation of the target (Aktas, Xu & Yurtoglu, 2018). Whether the due diligence process and negotiations are effective in safeguarding transaction prices in case of target underpricing, seems to depend on the rationality of the target’s management. A profit opportunity for bidders due to underpricing as described by Edmans, Goldstein and Jiang (2012) requires bounded rationality of target management (asymmetry in learning of managers) and low competition among bidders. These requirements do not follow from the results of Eckbo, Makaew and Thorburn (2017) and Betton, Eckbo & Thorburn (2009). Therefore, it is expected that merger negotiations have a significant effect in correcting for underpricing in the takeover premium. Research also suggests that any underpricing is more accurately filtered out as the negotiation duration increases.
Hypothesis 1: A non-fundamental decrease in stock price leads to a higher offer premium in friendly takeovers.
Hypothesis 1a: The relation between underpricing and premia becomes stronger as the negotiation duration increases.
12
Circumventing management
Besides the elaborate exchange of information in the form of a due diligence process during merger negotiations, target management already has access to relatively a lot of information on their own firm. On the contrary, shareholders generally seem to infer this from secondary market prices (Edmans, Goldstein & Jiang, 2012). If an acquirer is able to circumvent the relatively informed management of a target, they could exploit underpricing through this asymmetric learning. This could be managed through tender offers and hostile takeovers. It is theoretically possible to exploit underpricing through such a strategy. However, offer premia generally turn out to be higher in hostile takeovers (Betton, S., Eckbo, B. & Thorburn, 2008). Moreover, hostile takeovers are relatively uncommon after 2003. Nevertheless, if underpricing combined with a circumvention of target management leads to a lower offer premium this could align the findings of Eckbo, Makaew and Thorburn (2017) and Edmans, Goldstein and Jiang (2012). There could be (relative) profit opportunities in case of underpricing which increase takeover probability while the merger negotiation process generally safeguards efficient transaction prices.
Hypothesis 2: Underpricing does not affect offer premia substantially in takeovers where target management is circumvented.
Toeholds and underpricing
Possibly, the profit opportunity of acquirers in case of underpricing found by Edmans, Goldstein and Jiang (2012) exists while underpricing is accurately corrected in negotiations. The opportunity could arise not due to a lower final transaction price, but because acquirers are able to effectively acquire toeholds. Toehold acquisitions are purchases of a substantial amount of shares by a bidder prior to launching a bid (Betton, Eckbo & Thorburn, 2008). Consequently, when acquiring a toehold, the number of shares that need to be bought at full premium is lower. Furthermore, when losing the bid to a rival, the shares can be sold at a premium. Therefore, the bidder with the highest toehold has an advantage over tis competition (Eckbo, 2014). The potential gain from selling the toehold to a rival can facilitate raising the bid. Betton, Eckbo and Thorburn (2009) find a bimodal toehold distribution in takeovers, there seems to be either a large toehold acquisition or no toehold at all. Since toeholds decrease a target management’s claim on private benefits, they could antagonize target management, leading them to reject any bid. The benefits of toeholds need to weigh up to the potential rejection. Betton, Eckbo, Thompson and Thorburn (2014) show that increases in target price as a result of toehold acquisitions are identified by negotiating parties and do not increase the offer premia. It could be that exogenous underpricing of targets generally gets accounted for in the negotiation process, but that it decreases the threshold for which a toehold acquisition is profitable for the bidder. Furthermore, because a toehold creates a competitive advantage of a bidder over its competition, this strategy could be a way to profit from underpricing in case target management is able to identify underpricing in negotiations.
13
Hypothesis 3: higher price pressure increases the probability of a toehold acquisition.
When acquiring a significant toehold, price necessarily increases with the higher demand. If managers of the acquiring firm recognize underpricing, they will continue to buy more shares as the price increases. Besides increasing the probability of a toehold acquisition, the size of toeholds should, on average, be larger when targets are relatively underpriced.
Hypthesis 3a: toehold acquisitions are larger as price pressure decreases market valuation.
3. Data
Sample construction
The sample consists of 5.731 offer premia on U.S. non-financial public targets from 1985 to 2017 for which data on fundamentals and price pressure are also available. The analyses combine data on takeovers, fundamentals of the targets, mutual fund holdings of targets, mutual fund returns, and information on the stocks of the target sample. Data on takeovers and deal characteristics are extracted from the Thomson One database on mergers and acquisitions. The analyses are restricted to deals where the target is public, the database lists unique Cusip codes for the target and acquirer and there should be data available on target fundamentals and stock returns. Financial firms are excluded, just as deals that do not exceed a value of $ 10 million. Fundamental data on targets is retrieved from the combined Compustat/CRSP annual database. The CRSP database on securities is used to include data on trading volume and returns. Longitudinal data on the individual holdings of mutual funds and the returns of mutual funds are downloaded from the Thomson Reuters Mutual Fund Holdings database and the CRSP Mutual Fund database respectively. The longitudinal sample that is part of the first stage consists of 58.585 firm-years for which data is available on the market-to-book ratio. The instrument is constructed using the outflows of 6.282 mutual funds that are available in both the Thomson Reuters and CRSP databases on mutual funds.
Descriptive statistics
The descriptive statistics of the variables are shown in table 1. The Market to Book ratio is defined as the market value of equity divided by the difference between assets and liabilities (book value of equity). This ratio averages 2.57, which means that the average firm in the sample has a market value of equity more than double of the book value. This seems high, but is relatively common for publicly traded firms. Moreover, considering that the sample is made up of targets, which means there is generally a relatively high demand for their shares, it could be considered at the low end. This ratio is winsorized at the high end only (1%). The Total Assets of targets are specified in logarithmic terms. As any deal is excluded of the sample when its value is lower than $ 10 million, the log of assets (in $ million) should generally be higher than 2.3, which is the case. Leverage is made up of long-term and current liabilities
14
as a proportion of total assets, winsorized at the high end (1%). Cash is comprised of the cash holdings of a firm as percentage of total assets, winsorized at the 1% level. The Tangibility of a firm is defined as the proportion of Property, Plant and Equipment in total assets.
Significantly less observations are available for the cross sectional deal specific variables. As there are 5,731 deals where a premium is recorded, this is the number deals that is eligible for analysis of the second stage regression. The Premium is defined as the excess of the offer price over the target stock price four weeks before the announcement in percentage points. The average premium is around 42%, remarkably close to the 40% premium average stated by Edmans, Goldstein and Jiang (2012). As the 5th percentile shows, there are also targets in the sample who were sold at a discount. As this also happens regularly in practice, and there could be a variety of causes, these observations are kept in the analysis. Deal size is defined as the logarithm of the transaction value of the deal in $ millions. Merger
Duration is defined as the number of calendar days between the day that the deal becomes unconditional
or effective and the initial announcement date. Hostile is an indicator variable for when the target management did not approve of the takeover, 5.4% of the total sample. Unfriendly is an indicator variable for when the bid was unsolicited, regardless of approval of management (including hostile bids), which was in 9.7% of the bids in the sample. The variable Toehold is defined as the percentage of shares that the acquiring company held prior to the transaction. In line with Betton, Eckbo and Thorburn (2008), it seems that in most cases, a toehold is absent. However, if an acquirer already holds shares (864 observations), the number is substantial, averaging around 32%. The variable Price
Pressure comprises the hypothetical amount of shares sold by mutual funds in a certain period due to
redemptions, as a proportion of total trading volume of that period, a further elaboration on the construction of this variable is in the methodology section. From table 1 it can be inferred that there are a lot of observations without (significant) price pressure.
15 Table 1.
Sample characteristics
This table shows the mean values, standard deviations, and the number of observations of the variables. Furthermore, several percentiles are shown. Target fundamentals are normalized by total assets, unless otherwise specialized. The variable Merger Duration is specified in calendar days. Details on the calculation of Price Pressure are in the methodology section. Variable definitions are in the Appendix.
Percentiles
N Mean S.D. 5 25 50 75 95
Variable (1) (2) (3) (4) (5) (6) (7) (8)
Target fundamentals
Market to Book ratio 58585 2.567 2.065 0.605 1.178 1.852 3.172 8.372
Total Assets (ln) 58845 5.850 1.903 3.099 4.424 5.642 7.142 9.271 Leverage 58580 0.227 0.194 0 0.043 0.206 0.360 0.582 Cash 54104 0.111 0.145 0.002 0.015 0.051 0.152 0.425 Tangibility 57380 0.316 0.245 0.029 0.112 0.251 0.479 0.804 Deal specifics Premium (%) 5731 42.043 41.484 -7.630 17.220 34.940 58.140 118.060 Deal size (ln) 5745 5.794 1.857 2.988 4.377 5.653 7.100 9.001 Merger Duration 4661 121 102 29 60 97 148 301 Hostile (dummy) 5745 0.054 0.225 0 0 0 0 1 Unfriendly (dummy) 5745 0.097 0.296 0 0 0 0 1 Toehold (%) 5733 4.805 15.797 0 0 0 0 44.949 Price Pressure 58935 0.007 0.018 0 0 0.001 0.004 0.035
4. Methodology
In order to analyze whether exogenous underpricing leads to a higher transaction premium, a valid instrument for exogenous underpricing has to be constructed. Subsequently, a first stage regression analysis should show whether the instrument is relevant. In this study that entails whether hypothetical mutual fund redemptions significantly decrease the market valuation of a firm. Finally, a second stage analysis shows whether the variation in market valuation due to mutual fund price pressure significantly affects offer premia. Additionally, the effect of merger attitudes is included. Furthermore, IV Tobit and Probit models test whether, and by what magnitude, undervaluation lowers the threshold for toeholds found by Betton, Eckbo and Thorburn (2008).
16
Price pressure
The price pressure of mutual funds is calculated following the methodology of Eckbo, Makaew and Thorburn (2017). This entails aggregating significant negative fund flows over each quarter and calculating the hypothetical sales of a stock proportionately to the weight of that stock in each of the funds. Fund flows are defined as the change in total fund assets net of the realized fund return. For the fund flows by fund j in year t:
𝐹𝑗𝑡≡ 𝑇𝐴𝑗𝑡− 𝑇𝐴𝑗𝑡−1(1 + 𝑅𝑗𝑡) (1)
Only significant outflows should be considered. When managers face substantial outflows, their discretion regarding the choice of assets that are to be sold is the smallest. Therefore, higher flows will result in both more significant movements and movements independent of firm value. Only outflows larger than 5% of the fund are taken into account. Subsequently, the price pressure on the stock of target 𝑖 calculated as: 𝑃𝑟𝑖𝑐𝑒 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒𝑖𝑡 ≡ ∑ ∑ (−𝐹𝑗 𝑗𝜏∗𝑠𝑖𝑗,𝜏−1) 𝑉𝑜𝑙𝑢𝑚𝑒𝑖𝜏 𝑡 𝜏=𝑡−3 (2)
Where 𝑉𝑜𝑙𝑢𝑚𝑒𝑖𝜏 is the current trading volume, taking into account that more illiquid targets experience larger price pressure. The weight of a stock s is the lagged weight of stock i in fund j. Subsequently, quarterly pressure is aggregated per year to constitute annual price pressure per target. The negative of fund flows is used in the construction of the variable which means that more price pressure leads to a positive increase in the variable.
Baseline model
The market to book (M/B) ratio is used to proxy market valuation, following Eckbo, Makaew and Thorburn, (2017). Following Betton, Eckbo, Thompson and Thorburn (2014), the offer premium is defined as:
𝑃𝑟𝑒𝑚𝑖𝑢𝑚𝑖𝑡 ≡ 𝑂𝑃𝑖𝑡
𝑃𝑖𝑚−20− 1 (3)
Where OP denotes the final offer price on firm i in year t, P constitutes the price lagged by 20 trading days (four weeks), to exclude a run-up due to toehold acquisitions and merger rumors. The baseline model should show whether this premium is affected by underpricing due to mutual fund price pressure. This comprises a two-stage IV regression with industry and year fixed effects. In the first stage of this model (4), the variation in M/B due to price pressure is isolated.
𝑀/𝐵𝑖𝑡 = 𝜋0+ 𝜋1𝑃𝑟𝑖𝑐𝑒 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒𝑖𝑡+ ∑𝑗=2𝑛 𝜋𝑗𝑍𝑗𝑖𝑡+ 𝛼𝑘+ 𝛾𝑡+ 𝜀𝑖𝑡 (4)
17
𝑃𝑟𝑒𝑚𝑖𝑢𝑚𝑖𝑡 = 𝛽0+ 𝛽1𝑀/𝐵̂𝑖𝑡+ ∑𝑗=27 𝛽𝑗𝑍𝑗𝑖𝑡+ 𝛼𝑘+ 𝛾𝑡+ 𝜀𝑖𝑡 (5)
Where the Z’s comprise a series of 6 firm and deal characteristics following previous studies on takeovers and mispricing (Edmans, Goldstein & Jiang, 2012; Eckbo, Makaew & Thorburn, 2017). More specifically, the target’s capital structure is controlled for, as this can affect both valuation and premia. The capital structure variables include: total assets, leverage, cash & tangibility. It is expected that premia decrease with the size of the target (Betton, S., Eckbo, B. & Thorburn, 2008). On average, larger targets receive relatively smaller premia. Furthermore, firms with more leverage have a relatively small equity base, they therefore command a high relative premium over this equity e.g. for hypothetical synergy gains. Higher cash holdings are expected to decrease premia, just like high tangibility. Cash holdings do not provide potential synergies. On average, if a relatively large part of a firm’s assets is comprised of cash holdings, there are less assets that provide potential synergies. To a certain extent, the same holds for tangibility. Synergy potential for such assets with a relatively straightforward value is limited.
Furthermore, in additional analyses certain deal characteristics are also controlled for. Premia are expected to decline with deal size, largely an extension of the expectation of the general size effect. Smaller deals generally provide the acquirer with more discretion with regard to bidding. More often firms pay a high relative premium for a smaller firm as e.g. there could be more room for growth. Premia turn out to be larger for hostile or unsolicited bidders (Betton, S., Eckbo, B. & Thorburn, 2008). The direct effect of a larger negotiation period, or time from initial bid to deal conclusion, on a premium is expected to be mixed. From a competition perspective, it could mean that there is more time for other firms to compete in bidding. However, it could also mean a lack of interest after an initial, relatively low, bid. Therefore, no explicit expectation is formed with respect to the direct effect of negotiation duration. Premia and valuations vary significantly between industries and periods. Therefore, industry (SIC) and time fixed effects are included in the analyses (Edmans, Goldstein & Jiang, 2012; Eckbo, Makaew & Thorburn, 2017). These fixed effects are denoted by 𝛼𝑘 and 𝛾𝑡 respectively. Offer premiums were for example higher in the 1980s and lowest after 2003 (Betton, Eckbo, Thompson, & Thorburn, 2014). The coefficient of interest is 𝛽1, it is expected that non-fundamental decreases in market valuation lead to increases in the transaction premium, therefore a negative sign is to be expected.
18
Duration of negotiations
To test hypothesis 1a, whether negotiation duration strengthens the relationship between the premium and undervaluation, negotiation duration is included in the analysis as a moderator variable. If merger negotiations are effective in safeguarding efficient transaction prices, it is hypothesized that when these take longer, more information is exchanged, increasing the accuracy of transaction prices. As negotiations take longer, the true potential synergies between the firms become evident for both target and acquiring management (Calcagno, de Bodt & Demidova, 2017). The negotiation duration, D, for firm i is defined as the number of days between the merger announcement and completion:
𝐷𝑖≡ 𝐶𝑜𝑚𝑝𝑙𝑒𝑡𝑖𝑜𝑛 𝐷𝑎𝑡𝑒𝑖− 𝐴𝑛𝑛𝑜𝑢𝑛𝑐𝑒𝑚𝑒𝑛𝑡 𝐷𝑎𝑡𝑒𝑖 Which leads to the following model:
𝑃𝑟𝑒𝑚𝑖𝑢𝑚𝑖𝑡 = 𝛽0+ 𝛽1𝑀/𝐵̂𝑖𝑡+ 𝛽2𝐷𝑖𝑡+ 𝛽3𝑀/𝐵̂𝑖𝑡∗ 𝐷𝑖𝑡+ ∑9𝑗=4𝛽𝑗𝑍𝑗𝑖𝑡+ 𝛼𝑘+ 𝛾𝑡+ 𝜀𝑖𝑡 (6)
If mispricing gets corrected for accurately by negotiations, this correction should be more pronounced as the duration increases. The expected direct effect is mixed, but 𝛽3 is expected to be negative. The correction increases as more information is exchanged, up to a certain threshold. The relationship is likely to be non-linear, after a certain negotiation period the correction will be accounted for. However, this functional form will most likely still identify a possible relationship due to the fact that duration is relatively evenly distributed.
Absence of negotiations
In order to test the second hypothesis, the baseline model is extended to include a dummy variable in case the takeover is hostile, situations where negotiations with target management is circumvented. An interaction term is included to account for the hypothesized situation where exogenous price pressure provides a profit opportunity through asymmetric learning of bidders and target shareholders:
𝑃𝑟𝑒𝑚𝑖𝑢𝑚𝑖𝑡 = 𝛽0+ 𝛽1𝑀/𝐵̂𝑖𝑡+ 𝛽2𝐼𝑖𝑡+ 𝛽3𝑀/𝐵̂𝑖𝑡∗ 𝐼𝑖𝑡+ ∑9𝑗=4𝛽𝑗𝑍𝑗𝑖𝑡+ 𝛼𝑘 + 𝛾𝑡+ 𝜀𝑖𝑡 (7)
The directional expectations of the control variables are the same as in the baseline model. Furthermore, when takeovers are hostile, the offer premium generally turns out to be higher (Betton, Eckbo & Thorburn, 2008). This leads to a positive and significant expectation of coefficient 𝛽2. In case of underpricing, it could be that the relatively high premium of hostile takeovers is somewhat lower due to asymmetric learning. If this were to be the case, this leads to a positive and significant coefficient 𝛽1. A decrease in market valuation following price pressure, leads to a lower premium in case of a hostile takeover. The general expectation is that hostile acquirers, on average, pay a relatively high premium.
19
However, this ‘hostility’ premium is lower when the target is underpriced due to asymmetric learning of target shareholders.
Toeholds
Betton, Eckbo & Thorburn (2009) show that the distribution of toeholds is bimodal. Below a certain threshold the toeholds do not weigh up to the rejection costs and are therefore either 0 or positive. An assessment of the threshold using a Probit model analyzes whether the probability of a toehold acquisition increases in case of target underpricing. Additionally, to assess whether the size of toehold acquisitions increases with target underpricing, a Tobit model is specified. Here, the dependent variable is left censored at 0, and continuous when positive. The resulting model is specified as follows:
𝑇𝑜𝑒ℎ𝑜𝑙𝑑𝑖𝑡= 𝛽0+ 𝛽1𝑀/𝐵̂𝑖𝑡+ ∑𝑗=27 𝛽𝑗𝑍𝑗𝑖𝑡+ 𝛼𝑘+ 𝛾𝑡+ 𝜀𝑖𝑡 (8)
For the probability analysis the dependent variable, Toehold, is specified as an indicator variable (I). Taking on the value of 1 if there is a positive amount of target shares held by the acquirer previous to the transaction, and 0 otherwise. The series of capital structure control variables is identical to those used in previous models. However, the expectations regarding their effects vary somewhat. The chance of a toehold acquisition is expected to increase with target size. It is more likely that an acquirer is able to acquire a toehold without antagonizing the target management in case of a relatively large firm. Larger firms are likely to have a more dispersed ownership base. Furthermore, shares of larger firms are generally more liquid, this also means that it is more profitable to acquire a (larger) toehold. Expectations of the effects of the degree of leverage, cash holdings and tangibility of the target on toehold acquisitions are mixed. Generally, if a premium is expected to be relatively high, toeholds are more profitable. Therefore, the degree of leverage is expected to have a positive effect on toehold purchasing. The opposite holds for tangibility and cash holdings. Toehold bidding is generally more frequent in hostile bids (Betton, Eckbo & Thorburn, 2009). The coefficients on the control variables deal size and duration should be interpreted with caution as it is likely there are reversed causality issues with regard to toeholds. In case of significant toeholds, the actual transaction value will be lower by definition. If a larger amount of shares is already in possession, less have to be acquired to gain control. Therefore, toeholds lead to smaller deal sizes. Significant toeholds could antagonize management, which could, on average, increase the duration of negotiations. If the coefficient 𝛽1 is negative and significant, this would indicate that undervaluation increases the probability of a toehold. The nature of the Toehold acquisitions, bi-modal, left-censored, requires a Tobit model to assess the effect of mispricing on the size of toeholds. This model mixes a discrete and continuous distribution in order to account for an underlying (latent) linear relationship which is censored by the threshold. The model specification is equivalent to the probability analysis. However, the dependent variable, Toehold, is specified as continuous. Furthermore, the control variables are expected to have the same directional effects in the probability analysis as in the assessment of the size of toehold
20
acquisitions. In this case, if something is expected to increase the probability of a toehold, lowering the threshold, it is also expected to increase its absolute size. If the coefficient 𝛽1 is negative and significant, this would indicate that the size of toehold acquisitions is increasing with undervaluation.
4. Results
First stage
The results of the first stage regression are shown in table 2. The analysis is conducted both longitudinally, allowing for more observations, and cross-sectionally, including two control variables on deal characteristics. The longitudinal analysis incorporates all available years for the relevant targets. Therefore, it is likely to provide a more accurate estimation of the effect of price pressure on market valuation. Column 1 shows the coefficient estimates of this analysis. Price pressure due to mutual fund redemptions appears to significantly decrease the market valuation, in line with prior studies (Coval & Stafford, 2007; Edmans, Goldstein & Jiang, 2012; Eckbo, Makaew & Thorburn, 2017). If mutual funds sell 1% of a stock’s trading volume due to redemptions, on average, the market value of equity as a proportion of book value decreases around 6,8%, significant at the 1% level. This effect is also economically significant and in line with the magnitudes found in prior research. However, as the second stage analysis of the effect on premia only includes the year prior to the takeover, the effect should also be significant using only these observations. This is shown in Column 2 of table 2. Price pressure indeed also significantly reduces market valuation when only these observations are taken into account. The magnitude of the effect is somewhat lower, but roughly equal to the longitudinal analysis. If mutual funds sell 1% of a stock’s trading volume in a due to redemptions, on average, the market value of equity as a proportion of book value is expected to decrease around 6,3%, significant at the 1% level. This means that price pressure significantly influences secondary market prices.
The control variables generally have the predicted effect on the M/B ratio. Market valuation seems to decrease with firm size. On average, market valuation decreases with around 0,67% as a proportion of the book value of equity in case of a percentage point increase in total assets, significant at the 1% level. A possible explanation for this is that smaller firms have more growth opportunities and are therefore, on average, valued relatively high by the market. Market valuation is increasing with leverage. This is intuitive as higher leveraged firms share their income streams effectively with less owners. Therefore, the market value of the claims on these streams, the firm’s shares, increases. Although risks also increase, generally the market values these firms more. The effect is substantial, increasing leverage by 1% increases the market value by 3,2% as percentage of book value, significant at the 1% level. The coefficient estimate on cash holdings is relatively surprising. The analysis suggests that if cash holdings as percentage of total assets increase by 1%, the market value of equity increases by around 2,3% as percentage of book value. Generally, as can be seen in table 1, market to book ratio is above 1. Which means that equity is valued at a premium compared to its book value. This makes
21
sense if resources within the firm are put to a productive use, adding significant value to the resources. As this is not the case with cash holdings it seems counter intuitive that firms with more cash are valued relatively high. They should have an M/B ratio closer to par, which is 1. However, relatively successful firms may well generate substantial cash flows to the extent that they are not able to put them to use immediately. The tangibility of a firm does not seem to significantly influence a firm’s M/B ratio.
It seems that the transaction size increases with market valuation. On the one hand, this seems intuitive as deal size necessarily increases with an increase in market value. However, the result is surprising because smaller targets, on average, have higher M/B ratios. It seems that an increase in deal size of 1% is associated with in 0,75% higher market valuation compared to book value, significant at the 1% level. Furthermore, longer negotiations seem to be loosely associated with lower valued firms. An increase of negotiations with a single calendar day is associated with a decrease of 0,001 in the M/B ratio of the target.
22 Table 2
First stage regression of price pressure and controls on market to book ratio.
Coefficient estimates from the first stage regressions on market valuation, approximated by a firm's market to book ratio. Variables are defined in the Appendix. Column 1 shows the estimate of the effect of Price pressure on Market to Book ratio controlled for firm fundamentals, including fixed effects at the year and SIC level. This allows an annual longitudinal estimation of the effect. The second column adds deal characteristics, reducing the number of eligible observations. Robust t-statistics are reported in parentheses below coefficients. *, **, and *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.
M/B M/B Variable (1) (2) Price pressure -6.775*** -6.326*** (-5.03) (-4.04) Total assets (ln) -0.0249 -0.669*** (-1.03) (-11.09) Leverage 1.691*** 3.217*** (6.59) (12.64) Cash 2.716*** 2.299*** (11.12) (6.36) Tangibility -0.475** -0.346 (-2.44) (-1.33) Deal size 0.752*** (13.69) Merger duration -0.001* (-1.968) Observations 52,489 4,101 R-squared 0.272 0.418
SIC fixed effects Yes Yes
Year fixed effects Yes Yes
Instrument validity
Price pressure must satisfy several conditions to be classified as a valid instrument of the M/B ratio. First of all, the instrument should be relevant. It should hold enough information about the M/B ratio to be able to identify relationships between the M/B ratio and other variables. It turns out that price pressure is indeed a relevant instrument. Its effect on market valuation is very significant, has a t-statistic of more than four. Furthermore, the F-statistic of an additional analysis of only price pressure on the M/B ratio is 23, which is substantially higher than the rule-of-thumb value for weak instruments of 10 (Stock & Watson, 2011). Secondly, the exclusion restriction of instruments must be assessed. Price pressure should hold information about the M/B ratio that is completely independent of transaction premia in a takeover. Price pressure can only affect premia through market valuation. This entails a qualitative assessment of price pressure’s exogeneity. Arguably, significant redemptions of mutual funds are indeed independent of mergers. Even if significant redemptions are the cause or consequence of a merger wave, it should selectively affect the transaction premia of these takeovers in another way than through market valuation. Because there is only one instrument, and one endogenous variable, in
23
other words exact identification, the test of overidentifying restrictions is not possible to formally check for exogeneity of the instrument.
Second stage
The results for the second stage regressions of premia on (instrumented) market valuation are shown in table 3. The instrumented M/B ratio does not seem to have any significant effect on the transaction premia in takeovers. In fact, the effect has very low t-statistics in all four models and the direction of the effect even varies substantially between the models. Therefore, it is not possible to reject the null version of hypothesis 1 that underpricing has no effect on takeover premia. Moreover, the positive coefficient estimate in model 3 and 4 even suggest the opposite, that underpriced targets receive a relatively low premium. However, as the t-statistics are very low, these coefficient estimates cannot be interpreted in a meaningful way. This result is surprising in light of the results of Eckbo, Makaew and Thorburn, (2017). Their study suggests that merger negotiations should safeguard accurate prices. They find that acquirers are not able to exploit their own overpricing. As such, they also should not be able to exploit target’s underpricing. However, the results in Table 3 seem to suggest that any non-fundamental underpricing does not lead to significantly higher premia.
Most control variables also do not seem to have a significant influence on takeover premia. Most variables are not even close to attaining statistical significance. It seems that there is some indication that premia indeed tend to decrease with firm size. The first model suggests that a 1% increase in firm size, decreases the takeover premium, on average, by 0,1%, significant at the 5% level. However, as merger attitudes are included in the analysis this relationship loses its significance. In the fourth model, including a hostile indicator variable and moderator, firm size is not even significant at the 10% level.
Merger duration also does not seem to have a meaningful influence on premia. Moreover, there is no indication that as the time between the initial announcement and completion of the takeover increases, secondary market mispricing of the target is recognized and corrected for. It is not possible to reject the null hypothesis of 1a that lengthier merger negotiations do not lead to more accurate correction of mispricing. This result is curious as previous research found that more lengthy negotiations are associated with more information exchange and accurate picture of potential synergies (Calcagno, de Bodt & Demidova, 2017). However, regarding the estimates on the direct effect of M/B ratio, this result is not surprising. If there is no significant correction of mispricing in the first place, it is unlikely that this correction significantly increases with the length of merger negotiations. The direct effects of unsolicited or hostile attitudes in mergers on premia are in line with previous research. If merger bids are hostile or unsolicited, premia tend to be higher as found by Betton, Eckbo, and Thorburn (2008). The effect seems to be rather large, model 3 suggest that if a takeover is unsolicited or hostile, on average the premium tends to be 82,6% higher, significant at the 1% level. Furthermore, if the analysis is limited to hostile takeovers (model 4), the premium is on average 122,9%
24
higher, which is significant at the 1% level. These coefficients are relatively large as sample premia average 42% with a standard deviation of 41,5%. This means that hostile bidders on average pay a premium around two standard deviations above the average. The results regarding the effects of merger attitudes in light of underpricing are rather surprising. Not only do the results of the analysis suggest that hostile or unsolicited bidders pay premia that are substantially higher than friendly acquirers, but it appears that these differences significantly increase as targets are underpriced. If a bidder is hostile or unsolicited, a decrease in the market value of equity of 1% relative to its book value due to mutual fund redemptions increases offer premia by around 0,3%, which is significant at the 1% level. Moreover, in hostile takeovers, 1% underpricing due to fund redemptions can be expected to lead to an increase in the premium of around 0,46%, significant at the 1% level. These effects are substantial. It seems that in hostile takeovers there is a significant correction in the premium in case of underpricing. This is not in line with the hypothesized mechanism that acquirers are able to exploit underpricing through hostile takeovers. Apparently, some of the underpricing seems to be accounted for as a hostile acquirer is buying shares. Shares seem to be bid up more in a hostile acquiring process if a target is underpriced. It could be the case that shareholders indeed are able to identify certain mispricing, and do not fully infer firm value from secondary market prices. However, this result does not seem to be in line with asymmetric learning, as coined by Edmans, Goldstein and Jiang (2012).
25 Table 3.
Regression analyses of takeover premia on instrumented market valuation
The table reports estimates for the coefficients for IV regressions on premia. The general coefficient of interest is the (instrumented) market to book ratio. The regressions control for several fundamental characteristics of targets, as well as the size of the deal and the duration of merger negotiations. Models 2, 3 and 4 also include interaction variables of instrumented Market to book ratio with negotiations duration and attitudes, respectively. Variable definitions are in the Appendix. Robust t-statistics are reported in parentheses below coefficients. *, **, and *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.
Premium (%) Premium (%) Premium (%) Premium (%)
Variable (1) (2) (3) (4)
Market to book ratio -1.723 -2.308 0.981 2.131
(-0.23) (-0.36) (0.14) (0.30) Total assets (ln) -10.00** -10.10** -8.333* -7.671 (-2.04) (-2.16) (-1.82) (-1.67) Leverage 19.95 20.30 10.00 5.746 (0.786) (0.825) (0.431) (0.248) Tangibility -0.828 -1.037 0.555 0.762 (-0.113) (-0.147) (0.0807) (0.113) Cash 5.451 5.859 0.0181 -3.878 (0.316) (0.358) (0.00115) (-0.241) Deal size (ln) 7.106 7.184 5.415 4.688 (1.306) (1.374) (1.044) (0.894) Duration 0.00691 -0.00417 0.00643 0.00658 (0.802) (-0.105) (0.695) (0.711) Mispricing x Duration 0.00383 (0.248) Unfriendly 82.58*** (4.135) Hostile 122.9*** (3.854) Mispricing x Unfriendly -29.52*** (-3.423) Mispricing x Hostile -46.40*** (-3.298) Observations 4,092 4,092 4,092 4,092 R-squared 0.181 0.181 0.130 0.083
Industry FE Yes Yes Yes Yes
