1
The effect of the increase in intangible assets on acquisition premia following the Sixth merger wave.
Master’s thesis
Name: Thomas Butt Student number: 10826610
Supervisor: Florian Peters Track: Corporate Finance Amsterdam Business School
2
Statement of Originality
This document is written by Thomas Butt who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is 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.
Abstract
This paper analyses potential causes for the rise in acquisition premia in the last decade. Previously a downward trend in acquisition premia had been attributed to improvements in corporate governance throughout the 1990s and early 2000s. However, this trend deviated from theory following the end of the Sixth merger wave (2007) and has risen since. This paper tests a range of hypotheses using event studies, IV regressions and multivariate analysis. I find that the rise in intangible assets is the main cause of increased acquisition premia. Furthermore, through examining the channels through which intangible assets affect acquisition premia, it is shown that acquirers typically overvalue intangible assets which results in a premium. This contributes to prior literature by showing that, contrary to previous findings that intangible assets are systematically undervalued by the market, intangible assets are in fact overvalued by the acquirer.
3 Table of contents Statement of Originality ... 2 Abstract ... 2 1. Introduction ... 4 2. Literature Review... 8 2.1 Acquisition premia ... 8 2.2 CEO entrenchment ... 9 2.3 Deal characteristics ... 10 2.4 Intangible assets ... 11 2.5 Reference points ... 12 2.6 Competition ... 12 3. Hypotheses ... 13
3.1 Intangible assets hypothesis ... 13
3.2 Blockholder hypothesis ... 14
3.3 Competition hypothesis ... 16
3.4 Reference points hypothesis ... 16
4. Data ... 16
4.1 Sample construction ... 16
4.2. Variable Definitions ... 17
4.3 Descriptive statistics ... 19
5. Methodology & Results ... 21
5.1 Event Studies ... 21
5.1.1 Methodology of the Event study ... 21
5.1.2 Event study results: Intangible Asset ... 21
5.1.3 Event study results: Blockholder ownership ... 22
5.1.4 Event study results: Stock volatility ... 23
5.1.5 Event study results: Competition ... 24
5.2. What caused the rise in acquisition premia? ... 25
5.3 Does the effect of intangible asset vary with type of intangible asset? ... 35
5.4 Do acquirers overvalue intangible assets or does the market undervalue them? ... 37
5.5. Is the negative effect of intangible assets corrected for in the long run? ... 39
6. Conclusion ... 41
Bibliography ... 46
4
1. Introduction
It is well documented that during mergers and acquisitions the acquiring firm typically pays in excess of the firm’s current market value (Haleblian, Devers, McNamara, Carpenter, & Davison, 2009). This difference between the fundamental value and the deal value is known as the acquisition premium (Slusky & Caves, 1991). To date there have been many
explanations for the payment of this premium. However, a few theories dominate prior literature. Initially these premia were attributed to the transferring of some of the synergistic gains of the acquisition to the target shareholders in order that they would agree to parting with their shares (Slusky & Caves, 1991). However, this has become increasingly disputed with some papers finding that the value of the synergistic gains to have an insignificant effect on the acquisition premia (Slusky & Caves, 1991). This led a wealth of the prior literature to attribute acquisition premia to be a result of managerial hubris, managerial incentives, and an absence of sufficient corporate governance to prevent CEO’s from acting upon these traits (Alexandridis, Mavrovitis, & Travlos, 2012; Haleblian et al. 2009; Hayward & Hambrick, 1997; Roll, 1986). Managerial hubris and entrenchment even became synonymous with acquisition premia to the extent that some papers even use acquisition premia as a proxy for hubris (Beckman & Haunschild, 2002).
Corporate governance was found to be a determining factor in the size of the premia paid because it reduces the agency problem and increased the shareholder’s power, meaning that CEOs are less able to act on their hubris. Alexandridis et al. (2012) made the link
between the increased focus on corporate governance and corporate governance regulation to explain the downward trend in acquisition premia between the Fifth merger wave (1993 to 1999) and Sixth merger wave (2003 to 2007). This trend in corporate governance has
continued since 2007 as seen in figures 1 and 2. Anti-takeover provisions have been found to increase CEO entrenchment (Gompers, Ishii, & Metrick, 2003). Figure 1 shows that the number of firms with anti-takeover provisions in the form of the poison pill and a staggered board has decreased between the year 2000 and 2016. Although the number of golden parachutes has increased during this period, the cost relative to deal value is relatively small and therefore, this should have a minor effect in terms of CEO entrenchment.
5
Figure 1: Anti-takeover provisions over time: Figure 1 plots the number of firms with each anti-takeover
provision in the sample of both acquired and acquiring firms per year.
Figure 2: Limitations to CEO rights over time: Figure 2 plots the number of firms with each limitation of a
6
Figure 2 shows that there has been an increase in the limitations to CEO’s abilities throughout this period. This means that CEOs are less able to manipulate the board in order to push through an acquisition irrespective of the cost. Despite the trend in corporate governance, acquisition premia throughout this period have risen as seen in figure 3. This results in the following question being explored in this paper: What has driven this rise in acquisition premia following the Sixth merger wave?
Figure 3: Acquisition premia over time: Figure 3 plots acquisition premia over time. Acquisition premia are
measured using the stock return for one month prior to the acquisition announcement. Acquisition premia are winsorised at the 1% level and averaged on an annual basis.
This paper tests four hypotheses of what may have caused the rise in acquisition premia: increased intangible assets, increased blockholder ownership, increased competition and increased stock volatility. Each of these hypotheses are initially tested using an event study, after which a multivariate analysis on three of the factors is used. In the case of
competition, a two stage least squares regression is implemented on the grounds that the level of competition is not directly observable.
The results of this paper indicate that competition and stock volatility have not been significant drivers in acquisition premia. Although, the percentage of a firm’s stock held by blockholders was found to be significant these results are not robust under fixed effects and the relatively small increase in percentage blockholder ownership would indicate that the rise
7
in blockholder ownership was not a key driver in acquisition premia. Overall intangible assets were found to have a robust and significant positive effect on acquisition premia. This in conjunction with trend of increased investment into intangibles indicates that intangible assets are the key driver of acquisition premia in this period.
The channel through which intangible assets affect acquisition premia was then further explored and resulted in the finding that intangible assets have a significant negative effect on acquisition premia. This would indicate that the acquirers typically overestimate the value of intangible assets. This was in contrast to the findings of Laamanen (2007), which found that acquirers accurately price intangible assets and the market undervalues intangible assets. Thus, to confirm that the negative short-term effect of intangible assets was not reversed in the long run due to the recognition of the intangibles true value, regressions were performed on acquirer’s long term Buy and Hold Abnormal Returns. The results of these regressions show that the acquirer typically overvalue intangible assets which results in an increased acquisition premium, which negatively impact the acquirers returns.
These findings are more consistent with prior literature. A long term systematic undervaluation of intangible assets in the market would eventually disappear due to investors taking advantage of the arbitrage opportunity (Fama, 1998). Therefore, it is more likely that the acquirer would overvalue intangible assets than the whole market would undervalue them. Additionally, acquirers overpaying is in line with prior literature on hubristic CEOs (Roll, 1986) and CEO’s incentives such as empire building (Hope & Thomas, 2008). As CEOs have been found to overestimate the true value of the synergies which can be extracted. This is then amplified by the fact that the greater the amount of intangible assets means the greater the uncertainty and speculation surrounding the targets true value. Empire building CEOs may use the uncertainty surrounding intangible assets true value as an opportunity to justify paying a higher price for the target and increasing the likelihood of a successful bid.
This paper contributes to the body of literature focusing on why an acquisition premium is paid and further develops the understanding of the mechanism by which
intangible assets influence acquisition premia. This is of particular interest as there has been high growth of intangible assets over recent years and this paper provides a better
understanding of the consequences of this trend. Where this paper is particularly distinct from prior literature is that it finds that intangibles are overvalued by the acquiring firm opposed to the market which was found by Laamanen (2007). Furthermore, the motivations for this
8
overvaluation are in line with prior literature regarding CEO motives during mergers and acquisitions (Roll, 1986). These results also indicate the future authors should be careful when using acquisition premia as a proxy for governance in the future as the premium is not solely driven by corporate governance.
This paper will start by reviewing the relevant prior literature in section 2 before outlining the hypotheses based on this literature in section 3. Following this, section 4 will explain how the data was gathered and subsequently how it was treated to get the variables of interest. Section 5 describes the methodology and results. Finally, section 6 will be the conclusion.
2. Literature Review
In this section, the underlying theories will be explained upon which the hypotheses are based. In section 2.1, acquisition premia will be defined. In the remaining sections, factors that influence these premia will be outlined. Specifically, CEO entrenchment, deal
characteristics, intangible assets, competition and behavioural arguments will be discussed respectively in sections 2.2, 2.3, 2.4, 2.5 and 2.6.
2.1 Acquisition premia
Acquisition premia are the difference between the amount paid by the acquirer for the firm and the target firm’s stock price prior to the announcement. There is a large body of academic financial literature which delves into why an acquirer would pay more than the current
market value for a firm (Haleblian et al., 2009). Some studies argue that the resulting premia were paid to give the target firm’s shareholders a share in the synergies created by combining the two firms (Haunschild, 1994). These synergies can be divided into three broad categories: financial synergies, operational synergies and collusive synergies (Chatterjee,1986;
Damodaran, 2005). Financial synergies are those synergies which result in a reduction of the cost of capital. Operational synergies encompass all benefits derived from an increase in administrative or productive efficiency and collusive synergies refer to synergies derived from increased market power (Chatterjee, 1986). However, many other papers attribute acquisition premia to other and additional factors. Furthermore, there are some papers such as Slutsky and Caves (1991) which found that synergies aren’t significant in determining
9
acquisition premia. The majority of this literature has focused on the role CEO entrenchment plays in determining acquisition premia both on the side of the target and the acquirer as well as certain deal characteristics (Haleblian et al., 2009).
2.2 CEO entrenchment
Prior literature to date has frequently attributed CEO hubris as one of the predominant causes of acquisition premia (Alexandridis et al., 2012; Hayward & Hambrick, 1997; Roll, 1986). Roll (1986) reasoned that due to hubris on the part of the CEO (i.e. the decision maker in bidding), there is a positive valuation error on the part of the acquirer. This results in the acquiring party paying a greater acquisition premium. Hayward and Hambrick (1997) further developed this theory by looking at factors which would increase the level of a CEO’s hubris such as recent performance, media coverage and a measure of CEO self-importance. Their results supported the hubris hypothesis and found that the effect of hubris is greater when there is reduced board oversight and the CEO is entrenched, because this enables the acquirer CEO greater freedom to act upon their hubristic tendencies. Furthermore, Alexandris et al. (2012) attributed the downward trend in acquisition premia between the early nineties and up to the end of the Sixth merger wave (in 2007) as a result of improved corporate governance. Throughout this time period, regulation related to corporate governance steadily increased. For instance, in July 2002, the Sarbanes-Oxley Act was implemented, which reduced CEO entrenchment and their ability to act upon their hubris.
The effect of CEO entrenchment on the acquirer does not necessarily result in a higher acquisition premium, as it has been found that during periods of high uncertainty, such as during a financial crisis, the risk aversion of entrenched CEOs has a moderating effect on acquisition premia (Fralich & Papadopoulos, 2017). This is due to there being greater
uncertainty surrounding the true value of the asset. Therefore, the risk aversion of entrenched CEOs causes them to reduce the acquisition premium in order to reduce their exposure to this risk. Furthermore, Fralich and Papadopoulos (2017) found that during the 2008 financial crisis, credit constraints weakened potential target’s negotiating power, thereby making targets more attractive. However, amongst the acquirers which maintained a line of credit, they engaged in pre-emptive bidding to deter other potential acquirers. Consequently, this resulted in a rise in acquisition premia. However, entrenched CEOs were found to have a
10
moderating effect on pre-emptive bidding. As a result of risk aversion and market uncertainty, CEOs made more conservative bids, resulting in lower acquisition premia. Academic literature has also found that the CEO entrenchment on the side of the target plays a role in determining acquisition premia (Song & Walking, 1993). Song and Walking (1993) found that targets with greater managerial ownership in combination with entrenchment have higher acquisition premia on the account of management requiring greater compensation to give up their power. However, when there is CEO entrenchment, target managers may be willing to trade acquisition premia for their own pay off (Qiu, Trapkov & Yakoub, 2014; Song & Walking, 1993). Moeller (2005) found that this issue is mitigated by increased shareholder power as it allows the shareholders to force management to hold out for a higher acquisition premium. One instance of this is the presence of a blockholder (a shareholder which holds more than 5% of shares) as blockholders on the target side have been found to increase the acquisition premia (Fich, Harford & Tran, 2015; Moeller, 2005). However, Fich et al. (2015) found that in addition to the positive effect on premia,
blockholders simultaneously decrease bid frequency and the deal completion rate. This results in a net effect of approximately zero on overall target shareholder wealth. Additionally, Qiu et al. (2017) found that when the future of entrenched target CEOs is uncertain post-merger they negotiate more vigorously and hold out for higher premia.
2.3 Deal characteristics
Deal characteristics also factor into acquisition premia. Huang and Walking (1987) found that the method of payment influences the acquisition premia as measured through abnormal returns. They concluded that a cash purchase results in a greater premium than that of an acquisition through stock. Schleifer and Vishny (2003) further evaluated the motivations for this and found that the market valuations are a factor in determining the method of payment. This worked on the basis that when the acquirers stock is undervalued then using stock is more cost effective. However, if the acquirer is undervalued then the cost of using stock is great than that of cash and therefore, cash is the preferred method of payment.
Alexandridis, Fuller, Terhaar and Travlos (2013) researched the paradoxical effect of deal size on acquisition premia. They found that larger deals result in a smaller premium, but simultaneously destroy more value for the acquirer. This counterintuitive effect was then
11
attributed to the complexity of larger deals, meaning that the economic benefits of the merger are more difficult to extract.
Anti-takeover provisions of the target firm have also been found to result in the acquirer paying a greater premium (Varaiya, 1987). This is due to the fact the anti-takeover provisions provide the target with greater bargaining power which subsequently allows them to drive up the price during negotiations.
Betton, Eckbo and Thorburn (2009) researched the utilisation of a toehold position within competitive bidding. Toehold bidding is when the acquiring firm already has shares in the target prior to the acquisition. Normally as there is a premium on control when trying to purchase a target, it would make sense to try and acquire as much of the firm as possible before inflicting said premium. Therefore, a toehold position would typically reduce the acquisition premium (Betton et al., 2009). However, toehold positions are increasingly rare because of its potential for rejection costs incurred through the purchasing of shares resulting in an unfavourable trade-off for many acquirers.
2.4 Intangible assets
Intangible assets have also previously been found to be a factor in determining acquisition premia (Laamanen, 2007). The way intangible assets have been reported has been a highly contested issue in the last 50 years (Lohrey, DiGabriele & Nicholson, 2017). Currently, U.S. firms only report intangible assets in the form of goodwill, thereby ignoring knowledge and organisational stock developed by the company which are by nature more complex to value. Subsequently this results in differing opinions of the fundamental value of intangible assets which leads to a dispute in the literature as to whether the market correctly values intangible assets. Chan, Lakonishok and Sougiannis (1999) found that R&D intensive stocks outperform stocks with little or no R&D indicating that R&D is not priced into the market’s stock price. Alternatively, Hall, Jaffe and Trajtenberg (2005) found that the market responds significantly to the announcement of patents and patent citations indicating that the market is responsive to successful R&D. Furthermore, Chen and Hwang (2005) found that investors place different weights on human capital, intellectual capital, physical capital and structural capital, meaning that intangible assets are factored into stock prices.
12
Laamanen (2007) argued that the root of intangible assets impact on acquisition premia is due to information asymmetry. Laamanen (2007) reasoned that the acquirer and target have greater knowledge of the fundamental value of R&D than the market due to their industry expertise. This means that when the acquirer bids on the target they must pay more than the current market price as both are aware of the actual value of the asset. Additionally, Laamanen (2007) found that intangible assets have a positive effect on the acquisition premia as well as on the acquirer’s abnormal returns. However, this is based on a sample exclusively made up of acquisitions in the tech industry.
2.5 Reference points
Prior literature has also investigated the impact of human psychology on acquisition premia. Baker, Pan and Wurgler (2011) provided evidence for non-standard preferences during the negotiation of an acquisition. Essentially, they argued that both the target and acquirer shareholders have an anchoring bias on the week stock price high. This is due to the 52-week high being a significant point of reference to the parties involved in acquisition. This is evidenced by the fact that asking prices and offer prices rise with the 52-week high and that there is an increased success rate when the offer exceeds the 52-week high (Baker et al., 2011). In relation to acquisition premia they further found that offer prices rise with the week high, meaning that the greater the difference between the current stock price and the 52-week high, the greater the acquisition premia.
2.6 Competition
Acquisition premia have also been attributed to being as a result of increased competition. Fishman (1985) first hypothesised that when there is increased competition in the mergers and acquisition market that the firms pre-emptively bid higher in order to deter a secondary bidder placing a bid on the same target. Fralich and Papadopoulos (2017) also found evidence for this during the 2008 financial crisis. They found that despite the crisis, a pool of healthy bidders actively sought suitable targets under financial distress. In order to avoid bidding wars acquirers offered higher premia to pre-empt a competitive bid from another potential acquirer.
13
3. Hypotheses
3.1 Intangible assets hypothesis
Of the hypotheses which may explain the rise in acquisition premia, the intangible asset hypothesis is the most credible. This is due to there being an observable increase in intangible asset investment and therefore the amount of intangible assets held by firms in the last decade as seen in figure 4.
Figure 4: Intangible assets over time: Figure 4 plots the annual average of the logarithm of intangible assets.
This is using the intangible assets variable as calculated through the perpetual inventory method for all firms in Compustat, as well as the subsection which were acquired between 2001 and 2016.
However, this paper will dispute the channel through which intangible assets drive up acquisition premia on the grounds that it is more likely that the CEO overvalues intangibles assets than there is a systematic undervaluation within the market. If the results of Chan et al. (1999) still hold and there remains an average systematic undervaluation of intangible assets of 6.12% per year it would beg the question why the market has not corrected for this by implementing a long position on high intangible asset firms and short on firms with little or no intangible assets with otherwise similar characteristics. Therefore, it is expected that the market does reflect intangible assets in a firm’s stock price and a different result to that of Laamanen (2007) can be expected. Furthermore, based on the theory that CEOs are empire
14
building (Hope & Thomas, 2008), it can be expected that intangible assets provide hubristic CEOs justification to the board to bid higher on a target as the potential synergies are more complex. Additionally, under the assumption that CEOs are also empire building it may be the case that intangibles allow them to increase their valuation of the target to increase the likelihood of the deal going through. Instead they may perceive that they can extract greater synergies from the intangible assets than is truly realistic. Based on this theory, the following hypothesis can be made:
Hypothesis 1: Intangible assets have a positive and significant effect on acquisition premia.
Additionally, the type of intangible assets may be a driving factor under this hypothesis as it some intangibles may be harder to value than others.
Hypothesis 2: The effect of intangible assets varies with the type of intangible assets.
Furthermore, as the overvaluation occurs on the side of the acquiring firm the acquirer would be transferring money to the target’s previous shareholders and outside of the
company. This loss should then result in the following two hypotheses:
Hypothesis 3: Intangible assets have a negative effect on the acquirers returns in the short run.
Hypothesis 4: The negative effect of intangible assets is not reversed in the long run.
3.2 Blockholder hypothesis
An alternative explanation for the rise in acquisition premia following the Sixth merger wave could be an increase in blockholder ownership. Stronger shareholders have been found to monitor their CEOs better and subsequently prevent them from exchanging acquisition premium for personal benefits (Moeller, 2005). Subsequently as blockholders have been found to exact this greater level of monitoring due to them having more to lose if the CEO makes a bad deal (Fich et al., 2015). It can then be hypothesised that a rise in the number of blockholders would increase acquisition premia as it would mean a greater presence of strong shareholders forcing the target CEO to negotiate on behalf of the shareholders and not
themselves. This leads to hypothesis 5.
Hypothesis 5: The presence of a blockholder in the target has a significant positive impact on the acquisition premia.
15
Also, it could be hypothesised that an increase in the percentage of shares held by blockholders may result in higher acquisition premia. This would increase the amount that a blockholder can lose as a result of a lower acquisition premium. Therefore, providing the blockholder a greater incentive to monitor the CEO’s actions.
Hypothesis 6: The percentage of shares held by a blockholder has a significant positive impact on the acquisition premia.
Figure 5 below shows the trend in blockholder ownership from 2000 to 2016 and presents a steady upwards trend in the percentage of a firm’s total shares held by blockholders. The number of targets with a blockholder presence increases initially but then declines post 2006 before increasing again. This indicates that an increase in the presence of a blockholder was not a driving factor in the increase in acquisition premia. However, the increase in the percentage of shares owned by blockholders may be a cause for these higher acquisition premia. This provides greater credibility to hypothesis 6 as a cause of the rise in acquisition premia.
Figure 5: Blockholder Ownership over time: Figure 5 plots the annual average percentage of shares owned by
16
3.3 Competition hypothesis
Acquisition premia may have also increased because of increased competition. Increased competition may drive acquirers to pre-emptively bid higher in order to deter other potential bidders starting a bidding war and driving the price up higher (Fishman, 1985). As the level of competition is not directly observable it must therefore be proxied for. Subsequently, this makes it difficult to measure whether it has increased throughout the period. The pre-emptive bidding theory leads to the following hypothesis:
Hypothesis 7: Competition has a significant and positive impact on acquisition premia.
3.4 Reference points hypothesis
As Baker et al. (2011) found that prices increase with the 52-week high it could be the case that an increase in targets stock price volatility has resulted in a greater difference between the 52-week high and current stock price. Therefore, the premium as measured by the stock price one month prior to acquisition and the offer price will be greater. Furthermore, it can be expected that this variance would only have a positive effect on acquisition premia. This is because in the event the current stock price exceeds the 52-week high, the anchoring bias would lead bidders to believe that the target firm is overvalued and therefore not acquire at this time. Although this is hypothesised in Baker et al. (2011), it was not tested in their paper. Hypothesis 8: Stock volatility for the year prior to the acquisition has a significant positive effect on acquisition premia.
4. Data
4.1 Sample construction
To analyse the potential driving forces behind the increase in acquisition premia following the Sixth merger wave, a combination of different databases will be used. Thomson One will be used to gather a sample of 1,518 acquisitions which occurred within the United States between 2000 and 2016. As this paper is analysing acquisition premia, only mergers and acquisitions between two publicly traded firms were used. Although a share value for private firms may be estimated through valuation methods this is reliant on assumptions and would therefore be subjective. Compustat, ISS and CRSP, were used in conjunction with Thomson
17
One to provide further details on the firm characteristics of both the target and the acquirer. Initially, Thomson One’s cusips were shortened from 8 digits to 6 digits in order that it could be merged with CRSP. From this point on the Permno for both the acquirer and target were used as company identifiers. In cases where the Permno was not readily available in a database, such as with the blockholder data taken from ISS – Incentive Lab, linking tables were used. If an acquirer or target were missing a Permno, the observations were dropped from the sample. This left a sample of 1,336 mergers and acquisitions of publicly traded firms which occurred between 2000 and 2016. Additionally, financial firms were dropped from the sample. Once this database was compiled additional variables were calculated such as Intangible assets and Likelihood of a secondary bidder.
4.2. Variable Definitions
Intangible assets
In order to measure intangible assets, the same methodology as Peters and Taylor (2017) was applied. Peters and Taylor (2017) reasoned that intangible assets are either purchased
externally or developed internally by the firm. However, as U.S. accounting rules dictate that intangible assets created internally by the firm are to only be expensed through the income statement, the value of an internally developed intangible asset does not appear on the balance sheet (Peters & Taylor, 2017). Based on this reasoning, intangible assets are a function of Knowledge Capital (e.g. patents, software ect.), Organisational Capital (e.g. the value of the brand and human capital) and Goodwill (purchased external capital).
In order to calculate the value of knowledge capital, a perpetual inventory system is applied to R&D spending using formula (1)
𝐺𝑖𝑡 = (1 − 𝛿𝑅&𝐷)𝐺𝑖,𝑡−1+ 𝑅&𝐷𝑖𝑡
Where Git represents the value of knowledge capital for the current year, 𝛿𝑅&𝐷 is the
depreciation rate of assets generated through R&D, Gi,t-1 is the knowledge capital of the prior year and R&Dit is the R&D investment for the current year (Peters & Taylor, 2017). This variable was then generated using Compustat-CRSP merged database using database dating back to 1961. In the event that a prior year was not available the prior year’s knowledge capital was assumed to be zero due to Peters and Taylor’s (2017) finding that this methodology provided robust results. The depreciation rate variable per industry was
18
gathered using Table 4 of Li (2012) and when there was an absence of data, Peter and Taylor’s (2017) assumption of a constant depreciation rate of 15% was applied.
Following this, Organisational capital was calculated by applying a similar perpetual inventory model. However, in this case SG&A was used instead of R&D and a depreciation rate of 30% was applied based on the model designed by Peters and Taylor (2017). It is also worth noting that due to Compustat frequently combining actual SG&A expense and R&D expense under the SG&A variable, R&D was deducted from SG&A to arrive at the true value of the SG&A expense.
Finally, Goodwill as taken from Compustat was added to represent the external purchase of intangible assets. These three variables were then summed together to reach the value of total intangible assets. Following this, total fixed assets were calculated by adding PPE to intangible assets. Having gathered these variables intangible assets were then scaled by total fixed assets to find the proportion of intangible assets for the target firms on an annual basis.
WRDS provides the same intangible assets database used by Peters and Taylor (2017). However, merging this directly with a linking table and then with CRSP resulted in fewer matches, than when using the CRSP/Compustat database to calculate the intangible assets manually. Hence the latter was used to measure intangibles.
Likelihood of a Secondary Bidder
This paper will also contribute through its analysis of what variables predict additional
bidders. I hypothesise that the same variables which predict the likelihood of a firm becoming a target will predict the likelihood of a second bidder. This is rationalised on the bases that the likelihood of a firm being a target is effectively the likelihood that there will be a bid to acquire said firm. Hence, the more attractive a target is the more likely that there will be a secondary bid. The target characteristic variables used will be those found in Kimball,
Dietrich and Sorensen (1984). Additionally, I hypothesise that the probability of a competitor becoming an acquirer will also predict the likelihood of an acquisition as no matter how attractive the target there must be viable competitive bidders within the market. Therefore, an industry average for each acquirer characteristic excluding the successful acquirer and target will be calculated. These characteristics will be taken from Trahan and Shawky (1992). Additionally, the number of acquisitions for that month will be used as the number of
19
mergers and acquisitions occurring during that time may also play a role in the likelihood of an acquisition. The greater the number of acquisitions may indicate that it is an opportune time in the market to acquire as well as decrease the number of alternative targets in the market. Using the aforementioned variables as independent variables and a dummy variable for whether there were one or more bidders involved in the acquisition process a logit model will be carried out in order to find an instrument for the likelihood of a secondary bidder. Once this logit model has been carried out it will be used to predict the likelihood of a secondary bidder as this is a latent variable and otherwise could not be observed.
Synergies
As synergies are not directly observable based on the data which is publicly available and due to it being impossible to observe what would have happened if the merger or acquisition did not occur, a proxy for synergies is required. This paper will use the proxy used by
Alexandridis et al. (2012) which used the market estimate of synergies created from the acquisition by calculating the weighted average cumulative abnormal return of both the target and the acquirer. This was carried out through conducting an event study with an event window from 10 days prior and 10 days post the announcement of the merger. A market adjusted model was applied with the estimation period consisting of 250 days prior to the acquisition to 15 days prior (-265, -15). From this the CAR for the end of the 21-day event window and the Market value of both the target and acquirer were used to reach a weighted average.
𝑆𝑦𝑛𝑒𝑟𝑔𝑦 =𝐶𝐴𝑅10𝐴∗ 𝑀𝑉𝐴 + 𝐶𝐴𝑅10𝑇∗ 𝑀𝑉𝑇 𝑀𝑉𝐴+ 𝑀𝑉𝑇
4.3 Descriptive statistics
Table 1 shows the descriptive statistics for the sample. It is notable that throughout this period, acquirers typically have a positive return of 7% for the month prior to the acquisitions announcement but a negative long term BHAR of -5% indicating that some of the return is lost in the long run. Additionally, the acquisition premia on average are 20% for the sample. This would indicate that on average both acquirers and the target benefit from the acquisition. It is also observed that cash is more frequently used than stock throughout this period.
20
Table 1: Descriptive statistics: Table 1 shows the descriptive statistics for 1,336 deals completed between
2000 and 2016. The table provides statistics on deal characteristics as well as governance characteristics of both the acquirer and the target.
Mean Median Standard Deviation
Deal characteristics
Acquisition premia 0.20 0.16 0.68
Acquirer 1-month return 0.07 0.01 0.45
BHAR -0.05 -0.08 0.75
Synergy 0.17 0.12 0.34
Purchased with cash 0.40 0.00 0.49
Purchased with stock 0.27 0.00 0.45
Probability of secondary bid 0.05 0.04 0.02
Target intangible assets 0.42 0.45 0.21
Blockholder ownership as a % 0.05 0.00 0.16
Target stock volatility 0.00 0.00 0.00
Governance characteristics - Target
Classified Board 0.08 0.00 0.27
Golden Parachutes 0.06 0.00 0.24
Limit Ability to Act by Written Consent
0.08 0.00 0.28
Limit Ability to Call Special Meeting
0.07 0.00 0.26
Limit Ability to Amend Charter 0.07 0.00 0.25
Governance characteristics - Acquirer
Classified Board 0.19 0.00 0.40
Golden Parachutes 0.20 0.00 0.40
Limit Ability to Act by Written Consent
0.21 0.00 0.41
Limit Ability to Call Special Meeting
0.25 0.00 0.43
Limit Ability to Amend Charter 0.26 0.00 0.44
Observations 1336
Interestingly there is a significant difference in governance between the target and the acquirer. Typically, acquirers have more takeover defences but put greater limitations on the abilities of the CEO. Therefore, it is difficult to discern which party has better corporate governance.
21
5. Methodology & Results
Section 5 will be divided into 5 subsections which will contain both the methodology for each test as well as the analysis of the results.
5.1 Event Studies
5.1.1 Methodology of the Event study
A series of event studies will be carried out in order to determine if the hypothesised channels impact acquisition premia as measured by cumulative average abnormal returns. A significant difference between cumulative average abnormal returns of quintiles of the hypothesised variable will indicate that the variable has a significant impact on acquisition premia. The event date (day 0) will be the announcement date and the abnormal returns will be calculated using the adjusted market model. Following this, the abnormal returns will be combined to find the cumulative abnormal returns (CAR) per target for each day of the event window. Then, the CAR of each firm will be averaged for each event day to arrive at the cumulative average abnormal returns (CAAR) for the sample. The CAAR will then be plotted in order that the difference in acquisition premia can be directly observed. The window which will be observed will be for the 30-day period before and after the announcement date. Expected returns will be calculated based on a 120-day estimation window from day -60 to -180 where the announcement date is day zero. Each sample will be divided into quintiles based on the variable of interest. Then two event studies will be done for the highest and lowest quintile in order that the difference between the two can be observed. Following this t-tests were
performed on the difference in CAAR between the top and bottom quintiles for different event windows which can be found in appendix I.
5.1.2 Event study results: Intangible Asset
The above-mentioned method of dividing the event study quintile was applied to intangible assets. These initial results indicate that intangible assets do play a significant role in determining acquisition premia as at day 0 the 5th quintile increased from previously being 3.01% below that of the first quintile to being 8.87% greater. This margin fluctuates between the two quintiles but maintains a difference of 8.02% at the end of the observation window. This result strongly supports the hypothesis that the increase in intangible assets was a driving force in rising acquisition premia. Furthermore, prior to the announcement date, the bottom quintile is greater indicating that the market is more sceptical of a rumoured deal
22
going through if the firm has high intangible assets proportional to total assets. Appendix I shows these results to be significant to a 1% level for a 3-day event window, 5% level for in the 21-day event windows and 10% in the 61-day event window. Overall this supports the hypothesis that acquisition premia are significantly affected by intangible assets.
Figure 6: Cumulative Average Abnormal Returns (CAAR) by intangible asset quintile: Figure 6 plots
CAAR per event day. Quintiles are based on intangible assets as calculated by the perpetual inventory method (Peters and Taylor, 2017). CAAR is then calculated based on the adjusted market model and a 120-day estimation window (-180, -60) for the 61-day period surrounding the announcement of an acquisition.
5.1.3 Event study results: Blockholder ownership
The same methodology was then applied to quintiles depending on percentage of outstanding shares owned by a blockholder. In figure 7 it is observed that there is a little difference
between the CAAR in the run up to the announcement of the acquisition. However, following the acquisitions announcement there is a significant difference of 8.13% before continuing to increase to 13.46% at the end of the event window. This would support the hypothesis that the percentage ownership shares by blockholders increases blockholder monitoring and therefore drives up the acquisition premia. Appendix I shows that blockholder ownership significantly impacts acquisition premia in each of the three event windows. This positive effect is significant to a 1% level for (-1, +1) and (-10, +10), but then diminishes in significance to a 5% level for the (-30, +30) window. The robust significance of these findings supports the hypothesis that blockholder ownership significantly affects acquisition premia.
23
Although these results for blockholder ownership are the greatest in size and significance in the event study results, it is worth noting that they are endogenous on the account of there being an absence of controls for other factors. As the decision for an investor to become a blockholder is based on the firm’s characteristics and a univariate event study cannot control for firm characteristics. The blockholder ownership variables will effectively proxy for these characteristics resulting in an overestimation of the effect of blockholder ownership.
Figure 7: Cumulative Average Abnormal Returns (CAAR) by percentage blockholder ownership quintile: Figure 7 plots CAAR per event day. Quintiles are based on the percentage of shares owned by
blockholders as retrieved from ISS Incentive Lab. CAAR is then calculated based on the adjusted market model and a 120-day estimation window (-180, -60) for the 61-day period surrounding the announcement of an acquisition.
5.1.4 Event study results: Stock volatility
Figure 8 shows the results of the stock volatility by quintile event study. These results are in direct opposition to what was hypothesised by the behavioural channel as the cumulative abnormal returns of the first quintile exceed those of the fifth throughout most of the event window. These results would indicate that the behavioural channel is not significant in determining acquisition premia as can be seen also in appendix I.
These results may indicate that stock volatility is factored into acquisition premia but through another channel. It could be the case that stock volatility proxies for the uncertainty
24
surrounding the target firms true value. Therefore, when negotiating the acquisition the buying party may wish to be compensated for the additional risk underlying the target’s fundamental value and subsequently pay a lower premium. This would then explain why there is a greater premium paid for targets with a lower stock volatility. However, this analysis is purely speculative and would require further research. Additionally, as these results are not significantly different from zero it could just be as a result of statistical noise.
Although, this does provide counterevidence to the behavioural channel hypothesis this methodology is fairly rudimentory as it is subject to endogeneity through issues such as omitted variable bias. Therefore, to fully evaluate the behaviour hypothesis a multivariate analysis is required.
Figure 8: Cumulative Average Abnormal Returns (CAAR) by Target stock volatility quintile: Figure 8
plots CAAR per event day. Quintiles are based on target stock volatility for one year prior to announcement as calculated using date retrieved from CRSP. CAAR is then calculated based on the adjusted market model and a 120-day estimation window (-180, -60) for the 61-day period surrounding the announcement of an acquisition.
5.1.5 Event study results:Competition
The event study was then repeated using quintiles as determined by the competition variable which was constructed through a logit model. The resulting graph shows an initial deviation between the quintiles with the higher competition quintile having greater CAR. This is likely due to the greater the likelihood of a secondary bidder also resulting in a greater likelihood of a deal being announced. Following the announcement date, the fifth quintile is still greater
25
however the size of the difference between quintiles has diminished post announcement (from 8.22% to 2.77%). This would indicate that this initial deviation is as a result of the market placing greater probability on a deal going through for more attractive firms. Post announcement date the difference between quintiles is small but consistent up until the end of the event window. However, the difference between quintiles is found to be insignificant and negative in for the 3-day and 21-day windows in appendix I. This indicates that the small difference in acquisition premia, although significant to a 10% level across a 61-day event window, does not have a robust and significant positive effect on acquisition premia.
Figure 9: Cumulative Average Abnormal Returns (CAAR) by Competition quintile: Figure 9 plots CAAR
per event day. Quintiles are based on the probability of a secondary bidder as estimated using a logit model and variables relating to target characteristics, industry characteristics and annual number of acquisitions. CAAR is then calculated based on the adjusted market model and a 120-day estimation window (-180, -60) for the 61-day period surrounding the announcement of an acquisition.
5.2. What caused the rise in acquisition premia?
Having gained a preliminary insight into the effect of these hypotheses, a multivariate
analysis is carried out. There are two reasons for this, the first being that a multivariate allows for controls meaning that some of the statistical noise within the dataset can be controlled for, thus providing more reliable results. Secondly it enables the testing of the robustness of this paper’s initial findings.
26
Multiple multivariate OLS regressions will be performed in which the acquisition premium will be the dependent variable. Acquisition premia are defined as the premium from the stock price one month prior to announcement to the stock price on the day of
announcement as has been used in prior literature as a measure of acquisition premia (Alexandridis et al., 2012; Laamanen, 2007)
CEO entrenchment cannot be directly proxied hence multiple control variables have been applied. These control variables are divided into two categories: those which entrench the CEO through takeover defences such as a golden parachute clause and a staggered board and those which actively limit the actions available to a CEO. GIM index is often used as a proxy for corporate governance and consequently CEO entrenchment. However, as this index is based on anti-takeover provisions (Gompers et al., 2013) its effects are twofold. As
previously mentioned, takeover provisions provide the target with a stronger negotiating position which allows to drive up the premium paid on the acquisition (Varaiya, 1987). Additionally, Gompers et al. (2003) found that anti-takeover provisions provide management with extensive power, making them more entrenched. This entrenchment would allow
management to act upon its hubristic tendencies and therefore be willing to pay a higher premium. Therefore, to try and separate these two effects on acquisition premia limits to CEO powers will also be used to isolate the effect of governance. The governance measures are used for both the acquirer and the target because prior literature has found corporate governance to influence both the target and acquirer.
While carrying out the multivariate analysis, the same measure of synergies as Alexandridis et al. (2012) will be used. This measures synergies by calculating the weighted average CAR for the end of a 21 days surround the announcement date of an acquisition, in order to capture the market’s expectation of synergies for the combined entity. This variable should also absorb the effect of firm size as Meoller, Schlingemann and Stulz (2004) found a correlation between deal size in the respect that large firms will enter acquisitions with negative synergy gains whilst smaller firms do not. Therefore, as synergies and deal size are correlated, synergies should also proxy for the effects of deal size in this model.
(1) 𝑃𝑟𝑒𝑚𝑖𝑎𝑖𝑡 = 𝛽0𝐼𝑛𝑡𝑎𝑛𝑔𝑖𝑏𝑙𝑒 𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡+ 𝛽1𝐴𝑇𝑃𝑖𝑡𝑝+ 𝛽2𝑙𝑖𝑚𝑖𝑡𝑠𝑖𝑡𝑝+ 𝛽3𝑆𝑦𝑛𝑒𝑟𝑔𝑦𝑖𝑡 + 𝛽4𝐶𝑎𝑠ℎ𝑖𝑡 + 𝛽5𝑆𝑡𝑜𝑐𝑘𝑖𝑡 + 𝛾𝑡+ 𝛾𝐼𝑛𝑑+ 𝛾𝑎+ 𝜀𝑖𝑡
Premiait is the acquisition premia as calculated by the percentage increase in the targets stock prices for the month prior to announcements. The Intangible assetsit variable are
27
the targets intangible assets scaled by total fixed assets. This variable will be replaced in order to test for other hypotheses. Due to the distinction between the two, blockholder ownership both as a dummy variable where it takes the value of one if a blockholder is present and the percentage of shares held by blockholders will also be used in the place of intangible assets. Additionally, Stock volatility of the target for the year prior to
announcement will be used to test for the behavioural channel will also take this place. ATPitp is a control for the anti-takeover provisions of both the target and the acquirer such as the golden parachute and staggered board. The anti-takeover provisions used are the staggered board and golden parachute. The limitsitp variable is a control for whether there are
limitations in place on both the acquirer’s and target’s CEO. The limitations used are: limited ability to act by written consent, limited ability to call special meetings and limited ability to amend charter. Dummy variables are used with a value of 1 if the limitation or anti-takeover provisions is in place and 0 otherwise. Synergy is the control for synergistic gains from the merger as calculated using the methodology put forward by Alexandridis et al. (2012). Cash is a dummy variable for whether only cash was used to purchase the target where it takes the value of 1 in the event that cash was used or 0 otherwise. Stock is a dummy variable for whether only stock was used as the method of payment in the acquisition, when stock was used the variable is equal to 1 and otherwise it is 0. 𝛾𝑡 represents year fixed effects. 𝛾𝐼𝑛𝑑 represents industry fixed effects. 𝛾𝑎 is used to represent acquirer fixed effects this provides a proxy for CEO hubris within the acquiring firm as well as acquiring firm characteristics.
Although using acquirer fixed effects reduces the sample size it does provide a more robust control for the aforementioned factors. Despite the reduction in sample size, I am still left with a statistically significant sample as in only 46.15% cases did the acquirer only acquire one firm. This leaves a sample of the 53.85% (692 observations) of the acquisitions in which the acquirer bought multiple firms as can be seen in appendix II. The use of acquirer fixed effects also raises an endogeneity concern in the respect that it may be the case that the frequency of acquisitions by the acquirer has an effect on acquisition premia. This could be due to this sample being better at spotting profitable acquisitions and therefore acquire more firms at a lower premium. On the other hand, it may also be the case that this sample is made up of entrenched CEOs who are all actively building an empire. In order to test this
endogeneity concern regressions using the number of acquisitions per acquirer as the
independent variable for each of the models used throughout this paper in appendix III were performed. Appendix III shows that the number of acquisitions by the acquirer during the
28
period has no significant effect on any of the dependent variables used so the endogeneity concern with the use of acquirer fixed effects is low.
Increased competition’s effect on acquisition premia can result from two channels: i) the first being whether there is an active second bidder during the acquisition process. ii) the second is through pre-emptive bidding by the acquirer in order to deter additional bidders from making a bid. The first channel in which multiple other bidders drive up the acquisition price, and subsequently the abnormal return, has been well documented in other papers (Haleblian et al., 2009). This first channel is proxied for using a variable for the number of additional bids. However, the pre-emptive bidding channel is less well documented due to the probability of a secondary bidder being a latent variable. Furthermore, I hope to mitigate this issue through the use of a two stage least squares regression. For this purpose, the likelihood of a second bidder will be found using a logit regression, before being used as an instrument in the second stage regression. The second stage will be the same as model 1 but with competition replacing intangible assets.
Intangible assets
In Table 2, univariate and multivariate regressions are carried out both with and without fixed effects to test if intangible assets have an effect on the acquisition premium. Intangible assets are found to be significant to a 10% level in columns 3 and 4 when fixed effects are applied. Although this significance isn’t found in columns 1 and 2, the fact that it is found when stricter controls are applied in the form of fixed effects, adds greater validity to this result. Additionally, the magnitude of the coefficients is also economically significant as it indicates that a one percent increase in intangible assets would result in an increase in the acquisition premia between 2.5996% and 2.3819%. This means that the trend in intangible assets over this period is a credible cause of the increased acquisition premia. Furthermore, this supports the case that intangible assets have a robust significant effect on acquisition premia, as these results are in line with the results of the event study and are replicated when using the logarithm of intangible assets as seen in appendix IV.
29
Table 2: The effect of Intangible assets, Stock volatility and Competition on acquisition premia. The OLS regressions 1 to 8 and IV regressions 9 to 12 test whether Intangible Assets
and Stock volatility impact acquisition premia. Acquisition premia is measured as the stock return one month prior to the announcement date of the acquisition. The intangible assets variable is derived the perpetual inventory method as outlined by Peters and Taylor (2017) and scaled by total assets adjusted for the perpetual inventory method, Target stock volatility is calculated using CRSP daily stock returns for the year prior to acquisition announcement and Competition is the probability of a secondary bidder as predicted by a logit model. Additionally, in the regressions in which fixed effects are not included a constant is added to the model but omitted from the results. Standard errors are in parentheses and significance is indicated with the following *** p<0.01, ** p<0.05, * p<0.1.
Intangible Assets Stock volatility Competition
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Dependent
Variable: Premium Premium Premium Premium Premium Premium Premium Premium Premium Premium Premium Premium
Intangible Assets 0.3403 0.4519 2.5996* 2.3819* (0.4414) (0.5471) (1.3983) (1.3807) Target Stock volatility -80.6105 62.2424 440.0337 653.6532 (135.9842) (217.3653) (580.2483) (527.8530) Competition 4.1169 8.8583 12.2759 11.6717 (5.1755) (6.4041) (14.5205) (13.9987) Staggered Board Target 0.2620 0.4492 0.7602 -12.7371** 0.3742 0.4035 (0.5323) (1.0465) (1.7929) (4.9825) (0.4975) (0.9604) Golden parachute Target 0.8598 0.9351 3.3391 6.5365 0.9231 0.8007 (0.8667) (1.7488) (2.1664) (5.6008) (0.8470) (1.7275) Limit Ability to Act by Written Consent Target -0.5774 -0.5518 -0.1531 16.5102*** -0.6131 -0.4572 (0.6171) (1.1963) (2.3730) (5.6237) (0.6008) (1.1415) Limit Ability to Call Special Meeting Target 0.6190 0.3363 0.2722 0.7189 0.5595 (0.6277) (1.0578) (2.2162) (0.6069) (1.0261)
30 Limit Ability to Amend Charter Target -0.3256 -0.0943 1.3775 -0.3573 -0.1939 (0.8518) (1.6534) (2.0991) (0.8324) (1.6368) Staggered Board Acquirer -0.0980 -1.6835** -0.2683 -8.1901** -0.1628 -1.8920*** (0.3248) (0.7546) (0.9096) (3.0059) (0.3013) (0.7013) Golden parachute Acquirer 0.0677 1.2329 0.3941 7.3417*** 0.2305 1.4299* (0.3971) (0.8104) (0.9849) (2.3686) (0.3719) (0.7270) Limit Ability to Act by Written Consent Acquirer 0.2078 1.0923* 0.4886 2.1282 0.3038 1.2315** (0.3298) (0.6303) (0.9174) (4.2623) (0.3026) (0.5926) Limit Ability to Call Special Meeting Acquirer -0.1040 0.3066 0.3131 -6.0556 -0.1983 -0.1504 (0.3207) (0.6302) (0.8833) (3.7726) (0.2962) (0.5937) Limit Ability to Amend Charter Acquirer 0.1624 0.2347 -0.1550 5.8326 0.0859 0.1136 (0.3898) (1.0562) (0.9673) (4.9875) (0.3664) (1.0194) Synergy 0.2645 0.5687 -0.0950 -0.1617 0.3567 0.4213 (0.3211) (0.4654) (0.6152) (0.9602) (0.3114) (0.4561) Cash payment -0.2628 -0.1642 -0.3977 -2.9307 -0.1797 0.0166 (0.2660) (0.4820) (0.7257) (2.0803) (0.2504) (0.4627) Stock payment -0.6522** -0.5746 -1.4640* -6.0398 -0.5270* -0.4649 (0.3117) (0.6697) (0.8613) (3.8298) (0.2992) (0.6551) Observations 975 728 448 330 429 264 103 72 976 785 481 364 R-squared 0.0006 0.0192 0.4257 0.4570 0.0008 0.0747 0.7030 0.8851 0.0006 0.0240 0.4136 0.4241
Year FE No No Yes Yes No No Yes Yes No No Yes Yes
Acquiror FE No No Yes Yes No No Yes Yes No No Yes Yes
31
Although these results support the hypothesis that the amount of intangible assets held affect the acquisition premia, they do not explain why intangible assets impact acquisition premia. In order to further develop this channel, it is important to work out which component of the three components of intangible assets are driving the acquisition premia. Based on the findings of Laamanen (2007) that intangible assets impact acquisition premia through the asymmetric information between the acquirer, target and market, acquisition premia may depend on the type of intangible asset. It could be expected that the more difficult to value aspects of intangible assets such as knowledge stock and organisational stock should have a greater positive effect on acquisition premia. Alternatively, the goodwill components previous market value is already known by both parties and it is also known whether the purchase of the goodwill asset was overvalued or undervalued. Under this assumption goodwill intangible assets should have no effect on the acquisition premia.
Stock volatility
The results in table 2 do not support the reference point hypothesis. Stock volatility is insignificant in each of the regressions and in addition to this the magnitude of the
coefficients fluctuates between regressions indicating that there is no consistent economic effect. Overall this indicates that stock volatility is not driving the trend in acquisition premia.
However, due to data limitations regarding CRSP’s daily return coverage of smaller firms, the stock volatility data is limited and subsequently the number of observations is low in these results. Additionally, this small sample resulted in instance of multicollinearity within limitations to the abilities of the target CEO. This resulted in these variables being dropped from the regression in column 8. Therefore, the accuracy of this finding is limited due to data constraints, so may require further research with a larger dataset in the future.
Competition
Although in table 2 the magnitudes of the competition coefficients are consistent in columns 3 and 4 these results are insignificant which indicates an absence of pre-emptive bidding during this period. Furthermore, these results indicate that it is unlikely that acquisition premia have increased as a result of increased competition and acquirers employing a tactic of pre-emptive bidding.
32 Blockholder Ownership
Columns 1 to 4 in table 3 show the effect of the percentage of total shares owned by a blockholder effects acquisition premium. Columns 1 and 2 support the hypothesis that when there is a higher percentage blockholder ownership there is greater monitoring on the target CEO therefore the CEO cannot as easily exchange some of the acquisition premium for personal gain whilst negotiating the acquisition. The univariate regression without fixed effects finds the percentage owned by blockholders significant to a 0.1% level, whereas the multivariate regression without fixed effects finds percentage owned by blockholders
significant to a 5% level. Additionally, both of these results are consistent in magnitude with magnitudes of 1.4371 and 1.2199 respectively. These results indicate that a 1% increase in the percentage of shares owned by a blockholder results in an increase between 1.2199%. and 1.4371%
However, in columns 3 and 4 when fixed effects are added to the model both results are insignificant, and the coefficients become close to zero and negative. Subsequently, this brings into question the robustness of the results in columns 1 and 2 as the use of fixed effects normally provide the more accurate results. These results still show some promise of a
relationship between acquisition premia and blockholder ownership. However, the results are not as compelling as those of intangible assets hence further research into the effect of the percentage of stocks held by the blockholder are continued in the appendix V and not in the main body of this paper.
Blockholder presence
Table 3 finds in the univariate regression in column 5 that the presence of a blockholder has a positive effect with a 10% significance level. However, upon the addition of controls and fixed effects to the model blockholder presence becomes insignificant and very small in magnitude. This would indicate that a rise in the presence of blockholders does not result in an increase in acquisition premia.
33
Table 3: The effect of blockholder ownership on acquisition premia. The OLS regressions 1 to 4 test whether the percentage of shares held by a blockholder impact acquisition premium
and 5 to 8 tests whether the presence of a blockholder impacts acquisition premia. Acquisition premia is measured as the stock return one month prior to the announcement date of the acquisition. The percentage of shares owned by a blockholder prior to acquisition as taken from the ISS database. The Blockholder presence dummy variable is one when a target has a blockholder and zero otherwise. Additionally, in the regressions in which fixed effects are not included a constant is added to the model but omitted from the results Standard errors are in parentheses and significance is indicated with the following *** p<0.01, ** p<0.05, * p<0.1.
Percentage owned by Blockholders Blockholder presence Dummy variable
(1) (2) (3) (4) (5) (6) (7) (8)
Dependent Variable Premium Premium Premium Premium Premium Premium Premium Premium
Percentage owned by Blockholders 1.4371*** 1.2199** -0.5888 -0.0157
(0.4651) (0.5582) (1.3156) (1.5039)
Blockholder presence Dummy variable 0.4004* 0.1310 -0.4581 -0.2895
(0.2133) (0.2898) (0.4315) (0.5063)
Staggered Board Target 0.3251 0.4539 0.3599 0.5135
(0.4844) (0.9712) (0.4865) (0.9721)
Golden parachute Target 1.0010 0.9492 0.9396 0.8050
(0.8332) (1.8035) (0.8368) (1.8045)
Limit Ability to Act by Written
Consent Target -0.5866 -0.6548 -0.5951 -0.6564
(0.5891) (1.1689) (0.5917) (1.1681)
Limit Ability to Call Special Meeting
Target 0.6907 0.7315 0.7017 0.7686
(0.5966) (1.0648) (0.5985) (1.0659)
Limit Ability to Amend Charter Target -0.5138 -0.2374 -0.4043 -0.0648
(0.8215) (1.7292) (0.8291) (1.7328)
Staggered Board Acquirer -0.1592 -1.7610** -0.1765 -1.7308**
(0.2938) (0.7012) (0.2946) (0.7009)
Golden parachute Acquirer 0.1114 1.2327 0.1197 1.2266
(0.3622) (0.7483) (0.3632) (0.7478)
Limit Ability to Act by Written
Consent Acquirer 0.3236 1.3095** 0.3407 1.3190**
34
Limit Ability to Call Special Meeting
Acquirer -0.2209 -0.3366 -0.2149 -0.3425
(0.2897) (0.5940) (0.2909) (0.5931)
Limit Ability to Amend Charter
Acquirer 0.1603 0.5698 0.1744 0.5620 (0.3573) (1.0023) (0.3589) (1.0012) Synergy 0.3373 0.3985 0.3300 0.3690 (0.3016) (0.4749) (0.3026) (0.4740) Cash payment -0.1032 -0.2996 -0.1391 -0.3379 (0.2430) (0.4657) (0.2438) (0.4657) Stock payment -0.5131* -0.5719 -0.5304* -0.5543 (0.2866) (0.6449) (0.2874) (0.6450) Observations 1,314 816 555 383 1,314 816 555 383 R-squared 0.0072 0.0282 0.3976 0.4191 0.0027 0.0227 0.3992 0.4199
Year FE No No Yes Yes No No Yes Yes
Acquiror FE No No Yes Yes No No Yes Yes
35
5.3 Does the effect of intangible asset vary with type of intangible asset?
Peters and Taylor (2017) noted that there are three types of intangible assets: goodwill, knowledge stock and organisational stock. Goodwill refers to those intangible assets bought through acquisitions and developed within the firm. Knowledge stock are intangible assets such as patents which are gained through research and development. Organisational stock refers to intangible assets such as branding and human capital which is developed through investing into SG&A. Although each of the above are components of the intangible asset value constructed through the perpetual inventory method, they may have different effects on the acquisition premia. For instance, one type of intangible asset may be more frequently overvalued than the other. Alternatively, it may be the case that each of these components are insignificant meaning that their proportion relative to total intangible assets plays no role in determining acquisition premia.
To test these hypotheses three additional variables were added to the initial model: goodwill, knowledge stock and organisational stock. These variables were each of the three intangible asset types divided by total intangible assets to get a portion of each subcategory. This was then regressed against the acquisition premia in the place of intangible assets in model 1. A significant negative coefficient for these variables would indicate that an increase in this form of intangible assets could diminish the effect of intangible assets. A significant positive coefficient would indicate that an increase in this asset type would increase the acquisition premia. In the event that there are no significant results this would indicate that the weighting of the three components of intangible assets has no effect on acquisition premia. Therefore, each as asset is equally complex to value.
Table 4 finds that none of the three components are significant. The Goodwill coefficients are positive in each of the first three regressions, however, in the fourth the coefficient becomes negative and increases in magnitude. Knowledge stock also follows this pattern of being small, positive and insignificant in columns one through three before
increasing in magnitude and becoming negative. Organisational stock is insignificant in all four regressions but has positive coefficients with fixed effects and negative coefficients with fixed effects. This also follows the pattern of becoming larger and negative in the
multivariate fixed effect regressions. Overall these results indicate that the weighting of the components of intangible assets plays no significant role in determining the acquisition
