Master Thesis
Long-Term Bidder and Target Valuation Effects of M&As in the Energy Industry
Name: Jing Wan Student number: s3135780
Program: Msc Finance Supervisor: Wolfgang Bessler
Abstract
This thesis conducts an event study on the mergers and acquisitions’ valuation effects (stock price reactions) in both the short term and long term for bidders and targets in the energy industry between 1992 and 2015. By analyzing the cumulative abnormal return and buy-and-hold abnormal return, I conclude that mergers and acquisitions in the energy sector have a negative effect on the bidder’s short-term stock returns while a positive effect on the target’s short-short-term stock performance. Also, in the long-run, mergers and acquisitions in the energy industry have significant positive valuation effects of bidders and targets. Moreover, the regression results show that the acquisition share payment method lowers the bidder’s and target’s stock performance of long-run and short-run comparing to cash payment method.
1. Introduction
Merger and acquisition are financial and operating activities that combine two companies into a single entity. In this case, an acquiring corporation (bidder) acquires the control rights and ownership of an acquired corporation (target). Moreover, mergers and acquisitions are an alternative growth or expansion strategy when a company decides to enter into a new industry or market. Rainer and Katrin (2006) define mergers and acquisitions as the attempts of pursuing economies of scales, promoting a reorganization within the company, and decreasing the competition within in and between various industries and firms as well as to generate inorganic corporate growth.
Consensus holds that environmental issues such as energy shortage and environmental pollution are increasingly unresolved and imminent problems in the energy industry with irreparable consequences that must be solved. One of the most effective and efficient approaches is to reduce the dependence on traditional energy sources such as coal and nuclear power and rely more heavily on renewable energy resources instead. Although the energy transition and technology innovation have made substantial progress in recent years, the early stage of research and development, which requires large amounts of capital and investment, limits firms’ participation in developing new renewable energy resources. The renewable energy represents a potential pathway to help the energy sector to make a difference and get out of this current dilemma In 2015, the Paris Agreement, which is signed by all parties from the United Nations Framework Convention on Climate Change (UNFCCC), provides real steps to achieve a sustainable and low carbon future.
Verde (2008) introduces the idea of the implication of the wave of mergers and acquisitions to energy markets. With the existence of the wave of mergers, the energy market has experienced a structural transformation. During the reshaping, merger policy has played a significant part in enhancing the liberalization and consolidation of energy industries, especially in Europe. The liberalization in the European and American energy markets allow corporations to adapt their business plan to the external trends. Hence, with this driving force, in recent years there is a surge in the number of mergers and acquisitions across the energy industry, not only Europe and the United States but globally as well. According to KPMG (2018), the number of merger and acquisition in the renewable energy sector has seen a steady increase since 2010, and in the first half of 2017, the merger value amounts to EUR 22.5 billion, with nearly 200 transactions worldwide.
of merger and acquisition in the traditional energy sector could be achieving economies of scales by obtaining existing customer lists. Reddy and Xie (2017) argue in their research that Chinese oil companies made acquisitions for reducing production costs, stepping into worldwide markets, global market expansion, and decreasing the learning curve of acquiring new technologies and knowledge. This argument demonstrates that the present research topic examining why traditional energy companies, like oil firms, are willing to make acquisitions is timely and critical to understanding new trends in the energy industry. Mergers and acquisitions in the energy industries are conventional means of corporate restructuring and expansion. However, even though mergers and acquisitions in the traditional energy sector such as gas and oil have come under scrutiny, only a few articles focus on the interaction between traditional and renewable energy industries. Moreover, the extant literature often only pays more attention to bidder companies’ short- term post-acquisition stock price performance. Thus, the gap for the long-term post-acquisition valuation effects for bidder and targets in the energy sector needs to be filled in. The evidence in Porter (1987) suggests that the short-term post-acquisition performance cannot be representative for the key measure for the success of mergers or acquisitions as it only suggests that the financial markets fail to make immediate and full price adjustments concerning the future performance of the combined firms. Therefore, the long-term post-merger analysis is important in evaluating the true performance of merger and acquisition. In addition, there is limited research which addresses the effect of merger and acquisition on the target companies’ post-acquisition performance. Overall, there is rarely research that includes both long-run and short-run post-acquisition performances for both bidders and targets. Therefore, my contribution to the current research provides a comprehensive perspective on the effects of mergers and acquisitions on both the bidder and target firm in both the short- and long-run.
Besides the motivation mentioned above, there are abundant studies on how mergers and acquisitions affect bidders’ or targets’ short-term or long-turn stock returns. Most research has explored the abnormal returns in respect of post-acquisition performance for targets and bidders, while the empirical evidence and results are contradictory and diverse. Most of the arguments about short-term performance for bidder’s short-term abnormal returns are negative or zero while for target’s short-term abnormal performance is positive as it is most likely that the target firm captures the created value from mergers or acquisitions (see e.g., Graham et.al, 2002; Mulherin and Boone, 2000; Chatterjee, 1986). Nevertheless, the views on the long-term stock performance are much more complicated. As the methodologies for long-term abnormal returns are controversial, and all the methods have their own benefits and weaknesses. As such, this will be further investigated.
hypothesis. Chapter 3 contains the data description and the data selection. Chapter 4 describes the adopted methodology and Chapter 5 displays the empirical results. The last chapter is about the discussion and conclusion of this thesis.
As a result, the two main research questions of this paper are as follows:
1. What is the impact of merger and acquisition on the short-term stock abnormal stock price reactions (valuation effects) of both targets and bidders in the energy sector?
2. What is the impact of merger and acquisition on the long-term stock abnormal stock price reactions (valuation effects) of both target and bidder companies in the energy sector?
2. Review of the literature
2.1 Definition of the merger and acquisition in this paper
Although the definition of merger and acquisition is known and transparent, it is worth mentioning the narrow definition of merger and acquisition which is applied in this paper. Normally a merger happens when two companies, one is the bidder, the other is the target, legally integrate into one combined firm (Weston and Halpern 1983), usually these two firms are of similar size while acquisition generally equals to takeover, which represents the ownership transferring from one smaller-size firm to a bigger-size firm (Hacker 1964). In short, mergers and acquisitions are not categorized as two different events in this thesis, they are treated as the same. As concluded by Carpron et al. (1998), the acquisitions taking place between two firms in the same industry, they can be classified as horizontal acquisitions. Thus, our focus on the mergers and acquisitions in the energy sector is also horizontal mergers and acquisitions.
According to the classification of merger and acquisition in Thomson One Database, merger and acquisition can be classified as mergers, acquisitions, minority stake purchases, leveraged buyouts, spin-offs, repurchases, and exchange offers, etc. However only mergers and acquisitions will be included in this paper.
2.2 The motivation for the merger and acquisition
Apparently, more and more firms choose merger and acquisition as a corporate strategy for obtaining the operating benefit.
From the managerial perspective
Megginson et al. (2004) investigate in questionnaires to seven CFOs in Houston about why they choose mergers or acquisitions as the expansion strategy for their companies. The survey results show that their actions are all motivated by the operating synergies from the merger and acquisition. Thus, mergers and acquisitions here support management’s preference and personal choice as they treat mergers and acquisitions as a strategy to manage the firm-risk while expanding the business.
From the conflict of interest perspective
Wright et al. (2002) articulate that due to insufficient incentive mechanism and inappropriate performance monitoring within a corporation, whenever there are two parties with an engagement, there is the possibility that each side of the party is prone to pursuing their own interests to the disadvantage of the other party. This justification gives rise to the explanation of the agency theory. Sometimes, even worse, the terms of the contract are partial to protecting the management instead of the principal (shareholders). In addition, Trautwein (2013) proposes the empire-building theory that top management will incline to engage in merger and acquisition activities for the firm if those activities would increase their own benefits to acquire more ownership rather than maximize shareholder wealth.
From management hubris perspective
Roll (1986) proposes that managerial hubris phenomenon exists for both bidder and target firm when facing the decision of mergers and acquisitions. The bidders always pay too much for acquiring the target, as they tend to make too-high synergy estimation and even make mistakes in determining the valuation effects rather than acting as a rational investor. The reason behind this phenomenon might be bidders are too overoptimistic and putting too much expectation on the combined firm’s future performance. They may also have poor bargaining skills.
From synergy-creation perspective
Most of the time, the acquired firms should have at least some competitive advantages which can compensate for deficits in the bidding firms.
2.3 Efficient market theory
It is important to assume that the market is efficient when conducting event studies. The presence of efficient capital markets is the hypothetical and theoretical background of event studies as suggested by Malkiel and Fama (1970). Hereby, it is necessary to make the efficient market hypothesis that all the information available in the markets are fully reflected in the stock price. If the efficient market hypothesis is not rejected, then the merger and acquisition announcement information can be regarded as being incorporated into the stock price.
There are three main patterns of manifestation of an efficient market. One is weak form efficiency, the lowest extent of efficiency that all the historical information is included in the current stock prices. The second type is the semi-strong form efficiency, which indicates that all the public information has been reflected in today’s market price. The last kind is the strong form of market efficiency, which incorporates the first two forms and requires that all the public and private information are integrated into the current stock prices.
2.4 Literature review on merger and acquisition’s valuation effects on the bidder and target short-term performance
Most literature analyzes the merger and acquisition’s announcement effect for the bidder and target companies’ stock price. Rau and Vermaelen (1998) summarize that mergers and acquisitions have no or a negative effect for the bidders’ short-term stock returns, while the target captures most of the acquisition synergies, consequently having positive abnormal return around the acquisition date. Furthermore, Bartunek et al. (1993) acquisitions among public utilities firms have an unfavorable effect for bidders while a positive effect on the target’s wealth.
even cause detrimental effects on the merging firms’ financial performances. Furthermore, Wright et al. (2002) and Campa and Hernando (2004) both shed light upon mergers and acquisitions have a significant influence on the target firm’s acquisition performance in the short term.
2.5 Past literature on the merger and acquisition’s impact on the target and bidder companies’ long-term performance
While most articles are about the firms’ short-term performance in the post-acquisition period, Rao-Nicholson et al. (2016) shift their research focus to the long-term post-merger company performance. They demonstrate that merger and acquisition activities have negative effects on the long-term performance of the combined firms. However, we need to take into considerations that in their research, they use the ratio of return on assets as the benchmark to evaluate the combined firms’ post-acquisition long-term operating performance instead of stock prices. This ratio is a different index with the one I will adopt in my research. Therefore, the comparability of this paper will also be considered in concluding my research.
In a more detailed analysis of the long-term abnormal performance, Agrawal et al. (1992) uncover the fact that merger events for bidder firms cause a lower five-year post-merger stock price by 10%, indicating that mergers harm the bidder firm’s shareholder value. Similarly, studies such as Fanks et al. (1991), Higson and Elliott (1993) also conduct the long-term abnormal return analysis and both draw the conclusion that acquisitions have a negative effect on the bidder’s long-term stock returns.
Despite the fact that many research papers have found that merger and acquisition activities have an impact on the firm’s stock returns, some researchers have concluded that mergers and acquisitions have no significant effect on the long-run abnormal performance. For example, Mitchell and Stafford (2000) provide evidence that both buy-and-hold abnormal return and calendar-time abnormal return have hardly any impact on long-term abnormal performance. Meanwhile, Becker-Blease et al. (2007) failed to find that mergers and acquisitions create long-term value for shareholders in the electric power industry and that the stock prices and the operating performance of bidders even are seldom better and often underperformed compared to the firms, which are not involved in the merger activities. More recently, Dutta and Jog (2009) also share similar results that Canadian merger and acquisitions among 1993-2002 have no significant negative impact on abnormal returns under both event-time and calendar-time method.
the 2008 financial crisis, there is increasing attention paid to the firm’s strategic risks, with a higher focus on the higher level of management’s professional expertise and diversity, the introduction of stakeholder democracy into the business strategy, and increased emphasis on structural incentive mechanism within the corporate context. Furthermore, Healy et al. (1992) conduct an event study for both target and bidding firms and found that the combined companies operating cash flow performance and abnormal stock return performance both are improved in the following five years after the mergers or acquisitions occur. There is some other evidence for a positive post-merger performance when using Tobin’s Q and excess value for the performance measurement. The Tobin’s Q and excess value are all gradually enhancing in the consecutive five years after the mergers. Although Tobin’s Q and excess value are different measures for the long-term post-acquisition performance, they can also provide insights for the mergers and acquisitions’ valuation effects.
2.6 Hypotheses
Based on the literature review displayed above, the following first four alternative hypotheses are proposed.
First hypothesis:
𝐻1: Mergers and acquisitions have a negative effect on the short-term abnormal stock price
performance of bidder companies in the energy sector. Second hypothesis:
𝐻2: Mergers and acquisitions have a positive effect on the abnormal short-term stock price performance of target companies in the energy sector.
Third hypothesis:
𝐻3: Mergers and acquisitions have a positive effect on the long-term abnormal stock price performance of bidder companies in the energy sector.
Fourth hypothesis:
𝐻4: Mergers and acquisitions have a positive effect on the abnormal long-term stock price performance of target companies in the energy sector.
acquisitions. Thus, based on these two views, I develop the fifth, sixth, seventh, and eighth alternative hypotheses.
Fifth hypothesis:
𝐻5: The four different energy acquisition types will affect the short-term stock price of bidder companies differently.
Sixth hypothesis:
𝐻6: The four different energy acquisition types will affect the short-term stock price of target companies differently.
Seventh hypothesis:
𝐻7: The four different energy acquisition types will affect the long-term stock price of bidder companies differently.
Eighth hypothesis:
𝐻8: The four different energy acquisition types will affect the long-term stock price of target
companies differently.
According to the statement made by Loughran and Vijh (1997), when bidders choose to pay targets by cash, this would be an indicator for the efficient capital market that this acquisition or merger would be a value-creation event. Thus, it is reasonable to make the last two alternative hypotheses that the stock returns with the cash payment for mergers or acquisitions would outperform the short-term and long-term abnormal stock returns with the share payment method.
Ninth hypothesis:
𝐻9: The bidders in the energy sector which chose cash as the method of payment for the merger and acquisition in the short-term outperform the bidders, which choose share payment.
Tenth hypothesis:
𝐻10: The bidders in the energy sector which chose cash as the method of payment for the merger and acquisition in the long-term outperform the bidders, which choose share payment.
3. Data
3.1 Data availability and selection
Additionally, the variables (e.g., transaction value, acquisition attitude, etc.) for the linear regression are accessed via the Thomson One Database. Both the target and bidder companies’ daily and monthly stock price are derived from the DataStream databases. All the stock prices are in US Dollar.
To investigate how mergers and acquisitions affect target and bidding firms’ short-term and long-term stock abnormal prices, I use the merger and acquisition transactions with announcement dates are between 01/01/1992 and 11/24/2015. I categorize firms in the energy sector into two types:
1) The firms’ main business is in the traditional energy sector (based on the description of company’s Standard Industrial Classification (SIC) code in Thomson One Database, which is in relation to natural gas, gas, and oil)1.
2) The other’s firm’s primary business is associated with renewable energy (based on the description of company SIC code in Thomson One Database which is regarding cogeneration and alternative energy sources).
To further analyze the effect of merger and acquisition in different types of energy corporations, I divide the kind of energy merger and acquisition into four categories:
• the traditional bidder with the traditional target (acquisition type – TT), • the traditional bidder with the renewable target (acquisition type – TR), • the renewable bidder with the traditional target (acquisition type – RT), and • the renewable bidder with the renewable target (acquisition type – RR).
After defining the industry of the firms, the deal types of the merger and acquisition are chosen from Thomson One Database as disclosed value M&A and undisclosed value M&A. In this sense, I only consider mergers or acquisitions in my study. To note, as the focus of my research is the post-acquisition long-term performance, nearly completed deals are also considered in my sample data.
When reviewing the sample data for linear regression, there are many variables for the regression missing under many merger and acquisition events in the database. In order to ensure a large sample, I choose to separate the sample for calculating abnormal returns and for calculating the regression. More specifically, the sample data for regression is a subset of the
1 For traditional energy company, specially I choose the all SIC industry description which are natural gas
sample data for the abnormal since a lot of variables’ missing. I delete the merger and acquisition events at the stage of regression if regression variables are missing or not applicable to the event. Also, to calculate the short-term abnormal return and long-term abnormal return by using the daily stock price and monthly stock price, I did not evaluate events which have no stock prices in the estimation window and event window as well as the events whose daily or monthly stock prices are all zero or have no change during this period. Overall, the acquisition events for both bidders and targets are the same, for each bidder it has its corresponding target. However, the bidder’s missing samples are slightly different from the target’s, so the research is conducted separately on both the target and the bidder. Accordingly, for short-term post-acquisition performance, the daily stock price is used, while for long-term post-post-acquisition performance the monthly stock price is applied, as such samples for short-term and long-term performance are analyzed separately.
3.2 Descriptive statistics for sample data
Figure 1 and Figure 2 display how the different types of merger and acquisition are distributed over the sample time. Before the sample firms with missing variables for regression are removed, there are 911 bidders and 917 targets. From Figure 1 and Figure 2, for both bidder and target firm, it is easy to tell the TT acquisition type accounts for most of the merger and acquisition sample. The first notable point to mention is the massive increase in the number of merger and acquisition in 1999, which correspond to Martin and Sayrak’s statement in 2003 that in 1999, the global merger and acquisition transaction volume climbs to $2300 billion, which is over a 20% increase from 1985. In 2007, the number of merger and acquisition reaches the highest point which is line with UNCTAD (2008), showing a sudden worldwide boost in cross-border mergers and acquisitions, with the value of merger and acquisition transactions 21% greater than the previous highest number in 2000. Interestingly, from 2005 to 2015, the number of mergers and acquisitions in the energy sector does not change dramatically.
Table 1 exhibits the sample data for the short-term regression analysis grouped into different years and countries. After removing the missing variables, in Panel A, the number of merger and acquisition for the bidder firm for the short-term performance drops to 670, and for the target firm, the number decreases to 741. From Panel B, the number of merger and acquisition in Canada and the United States outnumber other countries.
Table 2 presents sample data for long-term regression analysis. The trend in long-term performance data remains roughly constant with the data for short-term regression estimation. Figure 1 and Figure 2: merger and acquisition distribution by years before regression
Figure 1. Bidders’ M&A distribution by years before regression
Figure 2. Targets’ M&A distribution by years before regression
Table 1 Merger and acquisition distribution for short-term performance after regression
Panel A and Panel B separately present the bidder’s and target’s merger and acquisition number distribution by announcement years after running the OLS regression for the short-term valuation effect. Panel A. Bidder and target M&A for short-term performance distribution by years after regression
Year Bidder Freq. Target Freq. Total Percentage
1992 3 3 6 0.43% 1993 2 3 5 0.35% 1994 13 10 23 1.63% 1995 9 19 28 1.98% 1996 10 20 30 2.13% 1997 18 21 39 2.76% 1998 26 55 81 5.74% 1999 27 44 71 5.03% 2000 23 20 43 3.05% 2001 30 25 55 3.90% 0 10 20 30 40 50 60 Number of acquisitions Year Bidder's M&A dsitribution by years before regression
TT Freq. TR Freq. RT Freq. RR Freq.
0 10 20 30 40 50 60 Number of acquision Year Target's M&A dsitribution by years before regression
2002 22 29 51 3.61% 2003 16 19 35 2.48% 2004 31 34 65 4.61% 2005 44 43 87 6.17% 2006 49 49 98 6.95% 2007 49 48 97 6.87% 2008 46 51 97 6.875 2009 41 37 78 5.53% 2010 33 36 69 4.89% 2011 44 42 86 6.09% 2012 38 36 74 5.24% 2013 29 27 56 3.97% 2014 32 37 69 4.89% 2015 35 33 68 4.82% Total 670 741 1,411 100.00%
Panel B. Bidder and target M&A for short-term performance distribution by countries after regression
Nation Bidder Freq. Target Freq. Total Percentage
Norway 9 15 24 1.70% Peru 0 1 1 0.07% Philippines 0 1 1 0.07% Romania 0 1 1 0.07% Russian Federation 7 6 13 0.92% Singapore 0 1 1 0.07% Slovak Republic 0 1 1 0.07% South Africa 0 1 1 0.07% South Korea 1 1 2 0.14% Spain 8 31 39 2.76% Sweden 1 2 3 0.21% Switzerland 1 1 2 0.14%
Trinidad & Tobago 0 1 1 0.07%
United Kingdom 25 36 61 4.32%
United States 281 265 546 38.70%
Utd Arab Emirates 1 0 1 0.07%
Total 670 741 1,411 100.00%
Table 2 Merger and acquisition distribution for long-term performance after regression
Panel A and Panel B separately present the bidder’s and target’s merger and acquisition number distribution by announcement years after running the OLS regression for the long-term valuation effect. Panel A. Bidder and target M&A for long-term performance distribution by years after regression
Year Acquire Freq. Target Freq. Total Percent
2012 36 32 68 5.23%
2013 28 27 55 4.23%
2014 32 35 67 5.15%
2015 35 34 69 5.30%
Total 609 692 1301 100.00%
Panel B. Bidder and target M&A for long-term performance distribution by countries after regression Nation Acquire Freq. Target Freq. total percent
Sweden 1 2 3 0.23%
Switzerland 1 1 2 0.15%
Trinidad & Tobago 0 1 1 0.08%
United Kingdom 24 30 54 4.15%
United States 266 249 515 39.58%
Utd Arab Emirates 1 0 1 0.08%
Total 609 692 1301 100.00%
4. Methodology
4.1. Methodology for short-term post-acquisition performance
To explore the relationship between acquisition announcements and abnormal returns, the event study methodology described by MacKinley (1997) is used.
To calculate the daily return from the individual stock price accurately, the logarithmic returns is applied to reduce the skewness and kurtosis of the return distribution.
𝑅𝑖,𝑡 = 𝐿𝑛 (𝑃𝑖,𝑡⁄𝑃𝑖,𝑡−1) (1)
Where 𝑃𝑖,𝑡 is the daily stock price for firm 𝑖 at event date t, 𝑃𝑖,𝑡−1 is firm 𝑖’s daily stock price at
one day before the event date t.
With the daily return, the abnormal return can be achieved by using the market model, which assumes the markets are always efficient to reflect all the public information into the daily stock price.
𝑅𝑖,𝑡 = 𝛼𝑖 + 𝛽𝑖𝑅𝑚𝑡+ 𝜀𝑖,𝑡 (2) Where 𝑅𝑖,𝑡 is the firm 𝑖’s daily actual return at event date t, 𝑅𝑚𝑡 is the corresponding market
portfolio’s daily return at same event date t, as the nation of the merger and acquisition events are worldwide, so the World Energy Daily Index is chosen as the benchmark portfolio for the calculation of the market return.
The abnormal returns for firm 𝑖 at event date t can be then developed by using the formula below.
𝐴𝑅𝑖,𝑡 = 𝑅𝑖,𝑡− 𝑅̅̅̅̅ (3) 𝑖,𝑡 𝐴𝑅𝑖,𝑡 = 𝑅𝑖,𝑡− (𝛼̂ + 𝛽𝑖 ̂ 𝑅𝑖 𝑚𝑡) (4) 𝑅𝑖,𝑡
̅̅̅̅ denotes the expected return by using the market model, which are normal returns without considering the effect of merger and acquisition event. 𝛼̂ and 𝛽𝑖 ̂ are the estimated 𝑖 coefficients of the market model. The two parameters of the market model can be calculated by using the estimation window of (-90, -11). 𝐴𝑅𝑖,𝑡 represents the abnormal return for
To test if merger and acquisition have an effect on the target or bidder company’s stock price, both the mean of abnormal return (AAR) and cumulative abnormal return (CAAR) are tested for the null hypothesis. The event window for the test is (-10, 10).
Given N firms, the average abnormal return (AAR) for all the events on event date i is AAR𝑡 = 𝐴𝑅̅̅̅̅̅ = 𝑖 1
𝑁∑ 𝐴𝑅𝑖,𝑡 𝑁
𝑖=1 (5)
The null hypothesis to test the average abnormal return is
𝐻0: 𝐴𝑅̅̅̅̅̅ = 0 (6) 𝑖 Which means θ = 𝐴𝑅̅̅̅̅̅𝑖 𝑉𝑎𝑟(𝐴𝑅̅̅̅̅̅)𝑖 1 2 ⁄ ~ N(0,1) (7) To get the average cumulative abnormal return for the period from 𝑡1 to 𝑡2,
𝐶𝐴𝑅𝑡1,𝑡2 = ∑ 𝐴𝑅𝑡2𝑡1 𝑖,𝑡 (8) 𝐶𝐴𝑅𝑡1,𝑡2
̅̅̅̅̅̅̅̅̅̅ = 𝐶𝐴𝐴𝑅𝑡1,𝑡2 = ∑ AAR𝑡
𝑡2
𝑡1 (9) The null hypothesis for the average cumulative abnormal return is
𝐻0: 𝐶𝐴𝑅̅̅̅̅̅̅̅̅̅̅ = 0 (10) 𝑡1,𝑡2 Which means 𝜃 = 𝐶𝐴𝑅̅̅̅̅̅̅̅̅̅̅̅̅𝑡1,𝑡2
𝑉𝑎𝑟( 𝐶𝐴𝑅̅̅̅̅̅̅̅̅̅̅̅̅)𝑡1,𝑡2 1⁄2 ~ 𝑁(0,1) (11) To further test what factors affect the short-term stock return performance, the ordinary least square (OLS) method is used for the regression analysis to assess how the cumulative abnormal return is being influenced for both target and bidder company. The OLS regression model is also tested by Stata 2015.
CAR( 𝑡1, 𝑡2) = α0 + β1(𝐼𝑅) + β2(𝑇𝑉) + β3(𝑃𝑀) + β4(𝐶𝐵) + β5(𝑅𝑆) + β6(𝑀𝑉) + β7(𝑀𝐷𝑇) + β8( 𝐴𝑇) + β9(𝐴𝐴) (12)
cross-border event, which is a dummy variable. RS is the acronym of the relative size of the acquisition, which is the ratio of the transaction value to the market value2 of bidder or target firm. MV stands for the logarithmic market value3 of the bidder or target company. MDT means
the master deal type, it is a dummy variable of type MA (value disclosed M&A) and IMA (value undisclosed M&A). AT is short for acquisition type, there are four acquisition types (RR, TR, RT, RR), which is also a dummy variable. The last dummy variable AA is the shorthand for acquisition attitude. All the acquisition attitudes except for friendly are categorized as other attitudes. α0 and from β1 to β9 are all parameters for the regression.
4.2. Methodology for long-term post-acquisition performance
Since Loughran and Vijh (1997) prove that cumulative abnormal return cannot explain the long-term post-acquisition performance well, for long-long-term abnormal return, the methodology is different from the short-term methodology. This is because of the long-term abnormal return’s sensitivity to the methods adopted to compute the returns. There are two main classical approaches to conduct long-term event studies. One is the buy-and-hold abnormal returns, and the other is calendar-time returns proposed by Fama and French (1993). Owing to time constraints, only buy-and-hold abnormal return approach is taken into consideration in this thesis.
In 1991, Ritter (1991) first proposed the method to compute the long-term abnormal return as mean buy-and-hold-abnormal-return, which becomes the most applicable pervasive approach to calculate the long-run abnormal performance. Several years later, Barber and Lyon (1997) compare the difference between one-year cumulative abnormal returns and buy-and-hold-abnormal returns and found the one-year buy-and-hold buy-and-hold-abnormal return rises a lot than cumulative abnormal returns. They conclude the reason behind this result is buy-and-hold abnormal returns consider the returns’ compounding impact.
Thus, they propose buy-and-hold abnormal returns (BHAR) can be calculated as following: 𝐵𝐻𝐴𝑅𝑖,𝑡2 = ∏ [1 + 𝑅𝑖,𝑡 ] − ∏𝑡2 [1 + 𝐸(𝑅𝑖,𝑡)]
𝑡=1 𝑡2
𝑡=1 (13)
Where 𝐸(𝑅𝑖,𝑡) is the market portfolio index, which uses the World Energy monthly index as the reference portfolio. 𝑅𝑖,𝑡 is the monthly return for the individual stock price, and it can be calculated as follows:
𝑅𝑖,𝑡 = (𝑃𝑖,𝑡⁄𝑃𝑖,𝑡−1) − 1 (14)
2 The market value of the bidder and target firm for regression use the market value four weeks prior to the
merger and acquisition announcement date.
As explained before, 𝑃𝑖,𝑡 is the daily stock price for firm 𝑖 at event date t, 𝑃𝑖,𝑡−1 is firm 𝑖’s daily stock price at one day before the event date t. The event window is from 12 months before the acquisition and 35 months after the announcement event date (As for some companies the 36th
month’s stock price information is missing so I only use the 35th’s stock price). All the returns
and buy-and-hold abnormal returns are calculated by Stata 2015 as well. The null hypothesis for the buy-and-hold abnormal return is:
𝐻0: 𝐵𝐻𝐴𝑅̅̅̅̅̅̅̅̅̅̅̅̅ = 0 (15) 𝑖,𝑡2 Which means 𝜃 = 𝐵𝐻𝐴𝑅̅̅̅̅̅̅̅̅̅̅̅̅̅𝑖,𝑡2 𝑉𝑎𝑟( 𝐵𝐻𝐴𝑅̅̅̅̅̅̅̅̅̅̅̅̅̅)𝑖,𝑡2 1 2 ⁄ ~ 𝑁(0,1) (16)
The OLS regression is also applied to investigate what factors impact the long-run stock return performance.
BHAR (𝑡2) = α0+ β1(𝐼𝑅) + β2(𝑇𝑉) + β3(𝑃𝑀) + β4(𝐶𝐵) + β5(𝑅𝑆) + β6(𝑀𝑉) +
β7(𝑀𝐷𝑇) + β8( 𝐴𝑇) + β9(𝐴𝐴) (17)4
5. Empirical results
5.1 Empirical results for short-term performance analysis
The empirical results regarding abnormal return and cumulative abnormal return for the short-term performance analysis are presented in Table 3, the results of cumulative abnormal return regression determinants are presented in Table 4. The t-test results are the one-sample t-test obtained by SPSS 25.
5.1.1 One-sample T-test result for short-term abnormal returns
From Panel A in Table 3, it is evident that for the bidder, only at the acquisition announcement date, the abnormal return is significantly negative at any significant level. Overall, for bidders, the first null hypothesis that mergers and acquisitions have a positive effect on the short-term stock price of bidders in the energy sector has to be rejected. As in most of the event windows, from 10 days before the announcement date to 10 days after, the mean abnormal return (AAR) does not show statistical significance except on the acquisition date on which the average abnormal return (-0.46%) is negatively significant. While for the target in Panel B, the t-test results are much more complicated. I can conclude, for most event windows, the mean abnormal returns are statistically significant, however, from Figure 4, especially around the announcement date such as -5 days (0.52%), -4 days (0.47%), -1 day (1.1%), announcement
4 The abbreviated variable in the BHAR regression share the same definition of the variables in the regression
date (8.34%), +1 days (1.92%), the mean abnormal return is positively significant while several days after the announcement date, it begins to show the average abnormal return is negatively significant. Thus, we can at least draw the conclusion that the merger and acquisition announcement events affect the short-term stock returns for the target company. However, how the events affect both bidders and targets’ short-term stock performance should be further investigated.
Therefore, I conduct the t-test for cumulative abnormal return (CAR). The CAR is divided into two categories. One is the cumulative abnormal return in consecutive days (Panel C and D in Table 3), the other is selected cumulative abnormal returns (Panel E and F in Table 3), whose selection is based on the event window taken by Shusterman et al. (2000) which are significant for the short-term abnormal returns. For bidder companies, none of the event windows show statistical significance for mean cumulative abnormal return in consecutive days at any significant levels, which means the first null hypothesis that mergers and acquisitions have a positive effect on the short-term stock price of bidders in the energy sector can be rejected. On the other side, for the target firm, most of the CARs in continuous event windows are positively significant. Moreover, for the selected cumulative abnormal return window in Panel E and F in Table 3, the bidder firm’s cumulative abnormal return on the announcement date, CAR (-1, 0) and CAR (6, 10) all have significant adverse effects on short-run abnormal performance, indicating that the first null hypothesis that mergers and acquisitions do not have a negative effect on the short-term stock price of bidders in the energy sector can be further rejected. Furthermore, for target firms, in each event window listed in Panel F in Table 3, the average cumulative abnormal return all shows positively statistically significant. In this sense, the second null hypothesis that mergers and acquisitions do not have a positive effect (no effect or a negative effect) on the short-term stock price of target companies in the energy sector can be rejected. In addition, many CAR event windows have been included for robustness check, and the results are all robust for these event windows.
Table 3 T-test results for abnormal return and cumulative abnormal return
Panel A and Panel B in Table 3 shows the t-test results for short-term abnormal returns in the event window AR (-10, 10). Figure 3 and Figure 4 are the line trend graph of bidder’s and target’s short-term abnormal returns. Panel C and Panel D display the cumulative abnormal returns in the consecutive days for bidders and target. Figure 5 and Figure 6 are the corresponding line trend pictures. Panel E and Panel F are the chosen event windows of cumulative abnormal returns for bidder and target. The line trend graphs of selected CAR for bidder and target are Figure 7 and Figure 8.
Panel A. T-test results for bidder’s abnormal return (AR)
One-Sample T Test
Test Value AAR (Bidder) = 0
Event window Average abnormal return t-value (Bidder) Number of
observation Sig. p-value (2-tailed)
-9 0.03% 0.253 900 0.800 -8 0.20% 0.901 900 0.368 -7 0.00% -0.002 900 0.998 -6 0.11% 0.804 900 0.421 -5 -0.09% -0.673 900 0.501 -4 0.15% 1.269 900 0.205 -3 0.05% 0.414 900 0.679 -2 0.15% 1.105 900 0.270 -1 -0.05% -0.499 900 0.618 0 -0.46% -2.117** 900 0.035 1 0.13% 0.767 900 0.443 2 -0.08% -0.575 900 0.566 3 0.06% 0.432 900 0.666 4 0.04% 0.288 900 0.773 5 0.06% 0.231 900 0.817 6 -0.20% -1.420 900 0.156 7 -0.13% -0.945 900 0.345 8 -0.09% -0.803 900 0.422 9 -0.05% -0.359 900 0.719 10 -0.11% -0.797 900 0.426
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 3 Line graph of bidder’s AAR
Panel B. T-test results for target’s abnormal return (AR)
One-Sample T Test
Test Value AAR (Target) = 0
Event window Average abnormal return t-value (Target) Number of
observation Sig. p-value (2-tailed)
-6 0.20% 1.151 905 0.250 -5 0.52% 3.809*** 905 0.000 -4 0.47% 2.963*** 905 0.003 -3 0.42% 2.414*** 905 0.016 -2 -0.08% -0.446 905 0.656 -1 1.10% 6.153*** 905 0.000 0 8.34% 14.442*** 905 0.000 1 1.92% 4.804*** 905 0.000 2 -0.13% -0.611 905 0.542 3 -0.05% -0.308 905 0.758 4 0.13% 0.963 905 0.336 5 -0.31% -2.796*** 905 0.005 6 0.44% 2.669*** 905 0.008 7 -0.24% -1.839 905 0.066 8 -0.18% -0.822 905 0.411 9 0.14% 0.794 905 0.428 10 -0.14% -0.918 905 0.359
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 4 Line graph of target’s AAR
Panel C. T-test results for bidder’s cumulative abnormal return in consecutive days (CAR)
One-Sample T Test
Test Value CAAR (Bidder) = 0
Event window Cumulative average abnormal return t-value (Bidder) Number of
observation Sig. p-value (2-tailed)
CAR (-10,+2) 0.14% 0.289 900 0.773 CAR (-10,+3) 0.20% 0.393 900 0.694 CAR (-10,+4) 0.23% 0.452 900 0.652 CAR (-10,+5) 0.30% 0.513 900 0.608 CAR (-10,+6) 0.10% 0.173 900 0.862 CAR (-10,+7) -0.03% -0.052 900 0.959 CAR (-10,+8) -0.12% -0.206 900 0.837 CAR (-10,+9) -0.17% -0.283 900 0.777 CAR (-10,+10) -0.28% -0.438 900 0.662
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 5 Line graph of target’s CAAR in consecutive days
Panel D. T-test results for target’s cumulative abnormal return in consecutive days (CAR)
One-Sample T Test
Test Value CAAR (Target) = 0
Event window Cumulative average abnormal return t-value (Bidder) Number of
observation Sig. p-value (2-tailed)
CAR (-10, 7) 13.43% 15.948*** 905 0.000
CAR (-10, 8) 13.25% 15.054*** 905 0.000
CAR (-10, 9) 13.39% 15.016*** 905 0.000
CAR (-10, 10) 13.25% 14.689*** 905 0.000
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 6 Line graph of target’s CAAR in consecutive days
Panel E. T-test results for bidder’s cumulative abnormal return (CAR)
One-Sample T Test
Test Value CAAR (Bidder) = 0
Event window Cumulative average abnormal return t-value (Bidder) Number of
observation Sig. p-value (2-tailed)
CAR (-10, 10) -0.28% -0.438 900 0.662 CAR (-5, 5) -0.04% -0.088 900 0.930 CAR (-2, 2) -0.31% -0.895 900 0.371 CAR (-1, 1) -0.38% -1.350 900 0.177 CAR 0 -0.46% -2.117** 900 0.035 CAR (-10,-6) 0.34% 1.148 900 0.251 CAR (-5,-1) 0.21% 0.789 900 0.430 CAR (-3,-1) 0.15% 0.723 900 0.470 CAR (-1,0) -0.51% -2.207** 900 0.028 CAR(0, 10) -0.83% -1.511 900 0.131 CAR (-1, 10) -0.88% -1.608 900 0.108 CAR (1, 5) 0.21% 0.502 900 0.616 CAR (6, 10) -0.58% -2.120** 900 0.034 CAR (-5, 1) -0.12% -0.334 900 0.739 CAR (-1, 5) -0.30% -0.659 900 0.510
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 7 Line graph of bidder’s CAAR
Panel F. T-test results for target’s cumulative abnormal return (CAR)
One-Sample T Test
Test Value CAAR (Target) = 0
Event window Cumulative average abnormal return t-value (Bidder) Number of observation Sig. p-value (2-tailed) CAR (-10, 10) 13.25% 14.689*** 905 0.000 CAR (-5, 5) 12.35% 16.301*** 905 0.000 CAR (-2, 2) 11.15% 16.109*** 905 0.000 CAR (-1, 1) 11.36% 17.148*** 905 0.000 CAR 0 8.34% 14.442*** 905 0.000 CAR (-10,-6) 0.89% 2.346*** 905 0.019 CAR (-5,-1) 2.44% 7.345*** 905 0.000 CAR (-3,-1) 1.44% 5.176*** 905 0.000 CAR (-1,0) 9.43% 15.776*** 905 0.000 CAR(0, 10) 9.92% 12.822*** 905 0.000 CAR (-1, 10) 11.02% 13.929*** 905 0.000 CAR (1, 5) 1.57% 3.378*** 905 0.001 CAR (6, 10) 0.01% 0.040 905 0.968 CAR (-5, 1) 12.70% 17.975*** 905 0.000 CAR (-1, 5) 11.00% 15.501*** 905 0.000
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 8 Line graph of target’s CAAR
-1.00% -0.80% -0.60% -0.40% -0.20% 0.00% 0.20% 0.40% 0.60% CAAR Event window Cumulatiev average abnormal return - Bidder
5.1.2 Determinants of short-term performance
In this section, the results of the regression are discussed. The variables for regression are divided into two categories. One is the continuous variables, like transaction value and relative size, the other is the categorical variables, such as payment method and acquisition attitude. From Panel A in Table 4, the continuous variables’ maximum, minimum, median and standard deviation are not significantly different in the digit, which means there are no obvious outliers. From Panel B in Table 4, the dummy variables into the regression are the variables which have the same first two SIC code as the corresponding bidder/target company, and whose payment method is share payment, which is cross-border merger and acquisition, whose master deal type is undisclosed value M&A, and whose acquisition attitude is friendly.
For bidders
From Panel A in Table 5, when including all the variables in Table 4 in the OLS regression analysis, first take bidder firm’s CAR (-10, 10) as an example. This regression model’s F-value is 2.047, the p-value is 0.019 < 0.05 if the significance level is at 5%, which means the general F-test’s null hypothesis that the regression model with no independent variables fits better than the current model can be rejected. We can conclude that the model with multiple variables fits better than the model without any independent variables, which means at least one included variable can impact the cumulative abnormal return. Besides that, the adjusted 𝑅2 is 1.8%, which is much smaller than 1, indicating that the included independent variables cannot explain the dependent variable well and suggesting that there might be some other independent variables not being included into the regression model. It is surprising that for all the variables in the regression model, only transaction value has a statistically significant negative relation with cumulative abnormal returns. Furthermore, the coefficient of transaction value is -0.015, indicating that every time there is one unit increase in the logarithmic transaction value, the CAR (-10, 10) will decrease 0.015 units. For robustness check, the similar event window CAR (-1, 1), CAR (-5, 5), CAR (-1, 10), CAR (-1, 5) and CAR (-5, 1) have been checked as well. The evidence upon all sides leads to a compelling conclusion that only transaction value has a statistically negative significant effect on the cumulative abnormal returns in the event windows mentioned above. Besides, the cross-border merger and acquisition events have a significant positive effect on CAR (-1, 1) and CAR (-1, 5) and friendly takeover attitude also have a significant adverse influence on CAR (-5, 5) and CAR (-1, 5). Moreover, comparing to MAs, IMAs have a significant negative effect on CAR (-5, 5), CAR (-1, 1) and CAR (-1, 5). In an overall review, the regression results are robust among different event windows.
model. From this perspective, the stepwise regression model is executed. For the stepwise regression in this thesis, the independent variable removing standards is only when the p-value of the independent variables is smaller than 0.05, it will remain in the regression. From Panel C in Table 5, the adjusted R2 for CAR (-10, 10) has improved from 1.8% to 9.7% at the
significant level 5%, which means the data fits in the stepwise regression model better. In addition to the transaction value, the acquisition attitude, the master deal type, cross-border activity and the acquisition attitude which have been detected to have the effect on the bidders’ short-run stock returns, the acquisition type – TT have a negative effect on CAR(-1, 1) when contrasting to acquisition type – RR. However, only at the event window (-8,8), (-5, 5), the friendly acquisition attitude has a statistically adverse effect on the cumulative abnormal return. To summarize, for bidder’s short-term abnormal performance, two regression approaches share a logical conclusion. The transaction value has a negative effect on the cumulative abnormal return. Also, the friendly acquisition attitude, IMA and share acquisition payment method can both negatively influence the cumulative abnormal return when compared to other acquisition attitudes, MA and cash payment. In this case, the ninth null hypothesis that the bidders in the energy sector which chose cash as the method of payment for the merger and acquisition in the short-term underperform the bidders which choose share payment can be rejected. Conversely, the overseas mergers and acquisitions will increase the bidder’s short-term stock returns. Furthermore, the acquisition type – TT comparing to acquisition type – RR can impact the cumulative abnormal return negatively, resulting in that the fifth null hypothesis that the four different energy acquisition types will not affect the short-term stock price of bidder companies differently can be rejected.
For Target
The regression results for target firms are much better than for the bidder. From Panel B in Table 5, the F-value in CAR (-10,10) is 7.481, p-value = 0.000 < 0.01, which means at least one included independent variable’s coefficient is not equal to zero, while the adjusted 𝑅2 is
9.0%, indicating there might be some significant variables missing from the current regression. Furthermore, when looking into how different variables affect the dependent variable CAR, it’s noteworthy that variables such as same first-two SIC code, transaction value, payment method, relative size of transaction value to target’s market value, target’s market value, and acquisition attitude all have statistically significant effect on CAR (-10, 10). Except for transaction value and relative size, all the other factors have negative effects on the CAR (-10, 10). For further robustness check, more event windows are included for regression analysis, all sharing the same results mentioned above.
regression is also enhanced, and there are more variables taken into regression consideration comparing to the regular OLS regression model. While in contrast with the bidder’s stepwise regression outcome, the impact of transaction value, the target’s market value and relative size on the cumulative abnormal return at all event windows are all statistically positive. Overall, the target firms which is the friendly acquisition type, which have the same first two SIC code with bidders, whose payment method is either share payment or other payment have statistically adverse effect on the cumulative abnormal returns at CAR (-10, 10), CAR (-1, 1), CAR (-5, 5). Besides, the relative size of the transaction value to the target’s market value has a significant but positive effect on the cumulative abnormal return in all event windows at 5% significant level.
Overall, the abnormal return results for bidder firm and target firm in the short term share similar conclusion with Leggio and Lien (2000), namely that for the merger and acquisition in the electric utility industry, the target firm experience positive announcement effect while for bidders, the announcement effect depends on if the target industry is regulated or not. If it is a merger or acquisition within the electric utility, the bidders obtain negative abnormal returns comparing to the target in the non-regulated industry which earns no abnormal returns. Additionally, the results that the cash payment method instead of stock payment for mergers and acquisitions would contribute to bidder’s higher abnormal returns are consistent with Walker (2000)’s conclusion.
Table 4 Descriptive analysis of variables for short-term performance regression
Panel A shows the descriptive statistics of continuous independent variables for short-term performance regression for both targets and bidders. Panel B presents the descriptive statistics of categorical independent variables for short-term performance regression for both target and bidder.
Panel A. Descriptive analysis of continuous variables Variables
Acquire sample Target sample
Mean Std.
Dev. Min Median Max
Mean Std. Dev.
Min Median Max CAR (-10, 10) -0.013 0.173 -0.866 -0.010 2.136 0.139 0.271 -1.447 0.126 1.571 CAR (-5, 5) -0.004 0.153 -0.477 -0.009 2.293 0.132 0.221 -1.266 0.125 1.510 CAR (-1, 1) -0.005 0.084 -0.405 -0.006 0.524 0.121 0.193 -1.357 0.097 1.702 CAR (-1, 10) -0.016 0.146 -0.756 -0.016 2.106 0.121 0.231 -1.447 0.113 1.628 CAR (-5, 1) -0.004 0.100 -0.376 -0.007 0.633 0.134 0.207 -1.324 0.114 1.576 CAR (-1, 5) -0.005 0.142 -0.373 -0.012 2.279 0.119 0.205 -1.299 0.107 1.636 Ln of transaction value 5.403 2.247 -2.659 5.334 11.016 5.413 2.211 -0.462 5.381 11.219 Relative size 0.655 1.770 0.001 0.311 25.783 1.698 3.549 0.001 1.299 50.340 Ln of market value 6.894 2.477 -0.315 7.082 13.382 5.272 2.304 -1.022 5.218 11.744
Variables Dummy variable in
the regression Freq. Percent Freq. Percent Same first-2 SIC
code not=0 44 6.57% 53 7.15%
same=1 626 93.43% 688 92.85%
Payment method cash=1 137 20.45% 164 22.13%
shares=2 226 33.73% 230 31.04%
other=3 307 45.82% 347 46.83%
Cross-border N=0 548 81.79% 602 81.24%
Y=1 122 18.21% 139 18.76%
Master deal type MA=0 259 38.66% 265 35.76%
IMA=1 411 61.34% 476 64.24%
Acquisition attitude other=0 24 3.58% 26 3.51%
Friendly=1 646 96.42% 715 96.49%
Acquisition type RR=1 36 5.37% 41 5.53%
RT=2 15 2.24% 26 3.51%
TR=3 14 2.09% 25 3.37%
TT=4 605 90.3% 649 87.58%
Table 5 Short-term performance regression results
Panel A gives the short-term performance regression results of bidders considering all the independent variables, while Panel B gives the short-term performance regression results of targets. Panel C and Panel D present the stepwise regression results for both bidders’ and targets’ short-term abnormal returns. Panel A. Bidder’s short-term performance regression results when considering all the variables
Bidder sample
Variables CAR(-10, 10) CAR(-5, 5) CAR(-1,1) CAR(-1,10) CAR(-5,1) CAR(-1,5)
b/t b/t b/t b/t b/t b/t
Same first-2 SIC code
(Same VS Not) -0.007 -0.002 -0.002 -0.001 -0.003 -0.001 (-0.252) (-0.066) (-0.130) (-0.034) (-0.168) (-0.029) Ln of transaction value -0.015*** -0.015*** -0.014*** -0.015*** -0.013*** -0.015***
(-2.765) (-3.026) (-5.276) (-3.304) (-4.226) (-3.367) Payment (Share VS Cash) -0.021 -0.014 -0.015 -0.018 -0.016 -0.013
(-1.002) (-0.765) (-1.603) (-1.074) (-1.387) (-0.786) Payment (Other VS Cash) -0.021 -0.012 0.003 -0.01 -0.003 -0.005
(-1.111) (-0.703) (0.394) (-0.609) (-0.264) (-0.350) Cross-border (Y vs N) -0.011 0.008 0.021*** 0.004 0.021** 0.008 (-0.604) (0.493) (2.434) (0.252) (2.014) (0.537) Relative size 0.005 0.001 0.003 0.006 0.002 0.001 (1.080) (0.162) (1.276) (1.595) (0.796) (0.356) Ln of market value 0.001 -0.003 -0.001 0.001 -0.001 -0.003 (0.152) (-0.727) (-0.312) (0.251) (-0.446) (-0.654) Master deal type
(-1.246) (-2.387) (-3.484) (-1.642) (-3.024) (-2.481) Acquisition type (RT Vs RR) 0.027 0.009 0.013 0.038 0.002 0.019 (0.506) (0.200) (0.509) (0.848) (0.079) (0.452) Acquisition type (TR Vs RR) 0.048 0.027 0.018 0.052 0.019 0.026 (0.870) (0.567) (0.721) (1.121) (0.621) (0.595) Acquisition type (TT Vs RR) 0.007 -0.01 -0.018 0.007 -0.013 -0.015 (0.227) (-0.361) (-1.273) (0.289) (-0.771) (-0.586) Acquisition attitude (Friendly VS Other) -0.047 -0.066** -0.032* -0.023 -0.060** -0.039 (-1.248) (-2.014) (-1.823) (-0.720) (-2.827) (-1.264) Cons 0.135** 0.198*** 0.134*** 0.093* 0.164*** 0.169*** (2.263) (3.802) (4.844) (1.854) (4.908) (3.486) F 2.047** 3.34*** 7.582*** 2.512*** 5.606*** 3.568*** P 0.019 0.000 0.000 0.003 0.000 0.000 𝑅2 1.8% 4.0% 10.6% 2.6% 7.6% 4.4% N 670 670 670 670 670 670
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively.
Note: For each variable, the first line number is the corresponding coefficients in the regression model, the second line number with brackets is the t-value that test if the variable in the regression is significant or not.
Panel B. Target’s short-term performance regression results when considering all the variables Target sample
Variables CAR(-10,10) CAR(-5,5) CAR(-1,1) CAR(-1,10) CAR(-5,1) CAR(-1,5)
b/t b/t b/t b/t b/t b/t
Same first-2 SIC code (Same VS Not) -0.082** -0.070** -0.065** -0.062* -0.074*** -0.061** (-2.172) (-2.320) (-2.456) (-1.956) (-2.632) (-2.161) Ln of transaction value 0.055*** 0.045*** 0.037*** 0.035*** 0.049*** 0.034*** (4.135) (4.266) (4.006) (3.145) (4.920) (3.415) Payment (Share VS Cash) -0.089*** -0.112*** -0.110*** -0.111*** -0.099*** -0.123*** (-3.161) (-5.009) (-5.568) (-4.652) (-4.746) (-5.847) Payment (Other VS Cash) -0.080*** -0.078*** -0.089*** -0.079*** -0.083*** -0.084*** (-3.157) (-3.854) (-4.962) (-3.639) (-4.374) (-4.417) Cross-border (Y vs N) 0.036 0.013 0.004 0.015 0.024 -0.007 (1.385) (0.618) (0.221) (0.690) (1.231) (-0.358) Relative size 0.007** 0.009*** 0.006*** 0.009*** 0.007*** 0.008*** (2.083) (3.667) (2.746) (3.339) (2.975) (3.528) Ln of market value -0.061*** -0.046*** -0.040*** -0.040*** -0.051*** -0.035*** (-4.695) (-4.491) (-4.349) (-3.692) (-5.295) (-3.605) Master deal type
(IMA vs MA)
(0.042) (0.537) (-0.021) (-0.693) (0.638) (-0.082) Acquisition type (RT Vs RR) 0.046 0.042 0.024 0.045 0.03 0.036 (0.704) (0.817) (0.520) (0.816) (0.619) (0.743) Acquisition type (TR Vs RR) 0.028 0.023 0.005 0.034 -0.001 0.029 (0.431) (0.446) (0.118) (0.617) (-0.011) (0.598) Acquisition type (TT Vs RR) -0.011 0.006 0.009 0.014 -0.002 0.017 (-0.249) (0.183) (0.285) (0.395) (-0.079) (0.542) Acquisition attitude (Friendly VS Other) -0.157*** -0.116*** -0.031 -0.054 -0.098** -0.049 (-2.963) (-2.742) (-0.837) (-1.213) (-2.495) (-1.227) Cons 0.443*** 0.348*** 0.274*** 0.300*** 0.351*** 0.271*** (5.585) (5.499) (4.910) (4.456) (5.936) (4.568) F 7.481*** 10.347*** 8.850*** 7.831*** 10.674*** 8.951*** P 0.000 0.000 0.000 0.000 0.000 0.000 𝑅2 9.5% 13.2% 11.3% 10.0% 13.6% 11.4% N 741 741 741 741 741 741
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively.
Note: For each variable, the first line number is the corresponding coefficients in the regression model, the second line number with brackets is the t-value that test if the variable in the regression is significant or not.
Panel C. Bidder’s short-term performance stepwise regression
Bidder sample
Variables CAR(-10,10) CAR(-5,5) CAR(-1,1) CAR(-1,10) CAR(-5,1) CAR(-1,5)
b/t b/t b/t b/t b/t b/t Ln of transaction value -0.012*** -0.016*** -0.014*** -0.011*** -0.013*** -0.016*** (-4.025) (-5.910) (-8.833) (-4.515) (-7.259) (-6.039) Acquisition attitude (Friendly VS Other) -0.075** -0.068*** (-2.410) (-3.397)
Master deal type
(IMA vs MA) -0.032** -0.024*** -0.028*** -0.029**
(-2.496) (-3.406) (-3.247) (-2.433) Acquisition type
(TT Vs RR) -0.028***
(-2.584)
Payment (Share VS Cash) -0.018***
𝑅2 2.2% 5.0% 10.6% 2.8% 7.9% 4.9%
N 670 670 670 670 670 670
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively.
Note: For each variable, the first line number is the corresponding coefficients in the regression model, the second line number with brackets is the t-value that test if the variable in the regression is
significant or not.
Panel D. Target’s short-term performance stepwise regression
Target sample
Variables CAR(-10,10) CAR(-5,5) CAR(-1,1) CAR(-1,10) CAR(-5,1) CAR(-1,5)
b/t b/t b/t b/t b/t b/t
Same first-2 SIC code (Same VS Not)
-0.089** -0.071** -0.067*** -0.070** -0.076*** -0.064** (-2.399) (-2.412) (-2.593) (-2.236) (-2.750) (-2.298) Ln of transaction value 0.053*** 0.043*** 0.037*** 0.036*** 0.046*** 0.034***
(4.087) (4.196) (4.078) (3.278) (4.798) (3.509) Payment (Share VS Cash) -0.101*** -0.116*** -0.114*** -0.121*** -0.105*** -0.125***
(-3.743) (-5.397) (-6.083) (-5.371) (-5.246) (-6.304) Payment (Other VS Cash) -0.089*** -0.081*** -0.092*** -0.086*** -0.087*** -0.086*** (-3.593) (-4.089) (-5.328) (-4.160) (-4.722) (-4.709) Acquisition attitude (Friendly VS Other) -0.160*** -0.119*** -0.104** (-3.048) (-2.840) (-2.653) Relative size 0.007** 0.009*** 0.006*** 0.009*** 0.007*** 0.008*** (2.128) (3.674) (2.770) (3.415) (2.978) (3.553) Ln of market value -0.057*** -0.044*** -0.039*** -0.040*** -0.049*** -0.035*** (-4.506) (-4.384) (-4.389) (-3.693) (-5.131) (-3.670) Cons 0.453*** 0.371*** 0.257*** 0.264*** 0.371*** 0.243*** (6.945) (7.128) (7.712) (6.554) (7.647) (6.852) F 12.358*** 17.566*** 17.646*** 15.237*** 17.924*** 17.622*** P 0.000 0.000 0.000 0.000 0.000 0.000 𝑅2 9.7% 13.5% 11.9% 10.3% 13.8% 11.9% N 741 741 741 741 741 741
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively.
Note: For each variable, the first line number is the corresponding coefficients in the regression model, the second line number with brackets is the t-value that test if the variable in the regression is
significant or not.
5.2 Empirical results for long-term performance analysis
This section presents the empirical results buy-and-hold abnormal returns for long-term performance analysis in Table 6. Table 7 is the descriptive analysis of independent variables for the regression. The outcomes of buy-and-hold abnormal return regression determinants are presented in Table 8.
The long-run stock return performance use monthly return for estimation and the event window’s unit are all in one month. According to Table 6, I may safely conclude that BHAR (0), BHAR (6), BHAR (12), BHAR (18), BHAR (24), BHAR (30) are all statistically significant, which means the bidders’ mean BHARs differ from zero in these event windows. In addition, we can conclude that the merger and acquisition events have a positive effect on the buy-and-hold abnormal returns at least at the 10% significant level for the event window mentioned above. Likewise, for target firms, the conclusion can be reached that BHAR (0) (BHAR = 21.36%), BHAR (6) (BHAR = 23.20%), BHAR (12) (BHAR = 22.82%), BHAR (18) (BHAR = 22.14%), BHAR (24) (BHAR = 14.02%) are also statistically significant at least at 10% significant level and all of them are positively associated with the merger and acquisition announcement events.
Table 6 T-test results for bidder’s buy-and-hold abnormal return (BHAR) for every month Panel A and Panel B show the t-test results for bidder’s and target’s buy-and-hold abnormal returns for every month. Also, Figure 9 and Figure 10 are the corresponding line trend graphs.
Panel A. T-test results for bidder’s buy-and-hold abnormal return (BHAR) for every month
One-Sample T Test
Test Value BHAR (Bidder) = 0
Event window (Months) Average buy-and-hold abnormal return t-value (Bidder) Number of
observation Sig. p-value (2-tailed)
10 12.86% 2.769*** 825 0.006 11 11.55% 2.667*** 825 0.008 12 10.60% 2.488** 825 0.013 13 11.77% 2.530** 825 0.012 14 11.85% 2.488** 825 0.013 15 12.91% 2.534** 825 0.011 16 13.93% 2.646*** 825 0.008 17 13.73% 2.539** 825 0.011 18 16.02% 2.751*** 825 0.006 19 14.76% 2.580*** 825 0.010 20 15.24% 2.598** 825 0.010 21 17.18% 2.726*** 825 0.007 22 15.35% 2.409** 825 0.016 23 15.82% 2.434** 825 0.015 24 15.35% 2.240** 825 0.025 25 14.75% 2.192** 825 0.029 26 13.99% 2.070** 825 0.039 27 14.30% 2.125** 825 0.034 28 13.79% 1.874* 825 0.061 29 15.14% 1.929* 825 0.054 30 17.01% 1.810* 825 0.071 31 16.15% 1.703* 825 0.089 32 17.69% 1.609 825 0.108 33 15.36% 1.530 825 0.126 34 16.02% 1.474 825 0.141 35 15.20% 1.356 825 0.176
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 9 Line graph of bidder’s mean BHAR
Panel B. T-test results for target’s buy-and-hold abnormal return (BHAR) for every month
One-Sample T Test
Test Value BHAR (Target) = 0
Event window (Months) Average buy-and-hold abnormal return t-value (Target) Number of
observation Sig. p-value (2-tailed)
Note: *, **, ***stand for statistically significant at the 10%,5% and 1% level, respectively. Figure 10 Line graph of target’s mean BHAR
5.2.2 Determinants of long-term performance
When coupled with the information in Table 7 and Table 8, we can assess the determinant variables for the long-turn stock performance. In Table 7, we can see that the informative details we can gain from this table share the same insights as table 4, so there are no apparent outliers within all the independent variables for the long-term stock return performance.
For bidders
According to Panel A in table 8, the F-value is 1.388, p-value = 0.167 > 0.05, which means this model does not fit well with the current variables at BHAR (12). Thus, due to prudent consideration, the BHAR (24), BHAR (35) have been regarded as the dependent variable in the regression for robustness check, nevertheless, the regression results still fall outside of the expectation.
The stepwise regression is applied to see if the conclusion for bidder’s long-term performance analysis will change or not. From Panel A in Table 8, the dependent variable of BHAR (12), BHAR (24), BHAR( 35) have all been included as the dependent variables. Only at BHAR(12), the F-value (p = 0.001 < 0.05) is statistically significant. The adjusted 𝑅2 has risen from 0.76% to 2%, suggesting this linear regression model fits the data better than the previous one. Besides, for BHAR (12), the share payment method has a statistically significant and negative influence on a significant level of 1%. On the other side, for master deal type, the IMA type has positively significant coefficients for BHAR (10) at the 5% significant level.
For target
From Panel B in Table 8, the dependent variable of BHAR (12), BHAR (24), BHAR (35) have all been included as the dependent variables. However, only when BHAR (35) is the dependent variable for the regression, the F-value (p = 0.03 < 0.05) is statistically significant. At the same
