The effect of political stability on cross-border
mergers and acquisitions
26 June 2018
Abstract
This paper examines the effect of political stability on the market’s reaction of the
acquirer to cross-border mergers and acquisitions. A sample of 461 deals conducted
from 2008-2018 is used. Two measurements are used to measure the effect of political
stability. These are corruption and if a country is in war or conflict. This paper finds
inconclusive results for the effect of political stability on the markets’ announcement
reaction.
Keywords: Cross-border, M&A, Political stability, Cultural differences, Cross-sectional,
CAR, Announcement effect, Event study
Author: Bastiaan Scholtes Student number: 10519475
Economics and Business, Finance and Organisation, 12 ECTs Bachelor Thesis in Finance
University of Amsterdam, FEB Supervisor: Shivesh Changoer
This document is written by Bastiaan Scholtes who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
1. Introduction
There are many mergers and acquisitions (hereafter M&A) every year, and this number is increasing over time. In 2008 the total M&A deal value was US$2252.43 billion. In 2017 this has become US$2994.42 billion (appendix 1). Cross-border M&A becomes an important strategy to gain market power. Globalization and advancing technology contribute to a growing number of cross-border M&A deals (Erel, Liao and Weisbach, 2012; appendix 2). In 2017 42% of the M&A deals were cross-border (White&Case).
A large number of studies shows that domestic M&A perform better than cross-border M&A. A possible explanation for this could be that cultural distance has a negative effect on the firm’s performance (Bertrand & Zitouna, 2008; Con et al., 2005; Shimizu et al., 2004; Child et al., 2001). While this explanation sounds plausible, there are other possibilities to why cross-border M&A underperformances in comparison to domestic M&A. One of them is political stability. Political instability in a country can cause higher integration costs and loss of efficiency. It is known that markets anticipate to decisions by governments, like the steel import tariffs by Donald Trump (fd, 2018). Also, countries with instable governments are associated with lower economic growth and are less efficient (Alesina & Perotti, 1996; Aisen, Veiga and José, 2013).
To examine this explanation, I examine the effect of political stability on the market reaction of the acquirer to cross-border mergers and acquisitions. To measure political stability I use two proxies, namely corruption and if a country is in war or conflict. For the analysis, this paper uses a sample of 461 cross-border deals ranging from 1-1-2008 until 1-1-2018. I use an event window of 5 days.
I find that political instability leads to a better performance for the acquirer after the announcement of the takeover. This result is inconsistent with the hypothesis that more political stability leads to better performance for the acquirer. In the sensitivity analyses these results are consistent for the event windows of 11 and 21 days. For the regression of the sample ranging from 1-1-2012 until 1-1-2018 and an event window of 5 days, I find that more political instability has a negative effect on the performance after the takeover. This is consistent with the hypothesis. The reason for this could be that the merger wave during the financial crisis affects the results. All findings for corruption and war are statistically insignificant.
My thesis contributes to the literature on the announcement effect of the acquirer in M&A, and in particularly border M&A. Prior research on border M&A mainly focuses on the cross-border effect in the short- and long-term performance for the acquirer and target. The findings for underperformance of cross-border deals are mainly focused on the effect of cultural distance. In my thesis I examine another possible reason for the underperformance of cross-border deals, namely political stability. Where prior research mainly focuses on one acquiring country that conducts
cross-border deals with different target countries, this paper uses data from 48 acquiring countries and 50 target countries and therefore gives a broader perspective on the effect of political differences.
This paper is divided in the following sections. Section 2 gives an extensive literary review of M&A, market reaction to M&A announcements, cross-sectional studies and cross-border M&A. Section 3 gives the hypothesis. Section 4 describes the research design. First it describes the method and then the data. Section 5 analyses the results. Section 6 concludes all findings.
2. Literature
2.1. M&A
According to Berk and DeMarzo (2014), M&A is when two companies integrate into one. Mergers are when two separate companies combine and go on together as one new larger company. Shareholders become shared owners of the new company. Acquisitions are when the acquirer purchases another company, the target, and takes over control. The shares of the target company transfer to the acquiring company.
There are multiple motives for M&A. First, M&A is a tool to gain new market power and eliminate competition. Second, it is a way to create synergy advantages. Examples of synergy advantages are economies of scale and scope, vertical integration, additional expertise, monopoly gains, efficiency gains and tax savings from operating losses. Third, it is a way to diversify and eliminate firm specific risk. Although it is more effective when shareholders diversify themselves because for large conglomerates it is difficult to allocate resources efficiently across multiple divisions themselves (Berk & DeMarzo, 2004, pp. 934-939). And fourth, for managers it is a way to get more important and earn more money. This is called empire building (Berk & DeMarzo, 2004, p. 560 & pp. 939-940). Empire building has a negative effect on the shareholder value. In 75% of the M&A deals where acquiring shareholders lose money, the manager gains money (Harford & Li, 2007).
There are also costs involved with M&A. According to the findings of Eckbo (2005) acquirers pay an average premium of 43% over the initial premerger price. Hence the acquirer needs to realize the estimated savings, otherwise, in the worst-case scenario it will go bankrupt and face additional bankruptcy costs. There are also integration costs involved with M&A. Examples of the integration costs are extra costs because of geographical distance, complexity of the firms or whether the deal is vertical or horizontal (Ahern et al., 2012).
Because it is not clear if savings outrun the costs, it is hard to tell if the deal creates value for the acquiring shareholder. If investors expect that the savings outrun the costs then there is a positive market reaction on the day of the announcement.
2.2. Market reaction to M&A announcements
Because the acquirer expects to gain advantages that outweigh the costs, he is willing to pay an acquisition premium. “This is the percentage difference between the acquisition price and the premerger price” (Berk & DeMarzo, 2014, p. 933). Investors of the target firms usually gain, but the acquiring investors make little to no return.
In a large sample of M&A deals from 1985-2005 with U.S. targets, acquiring shareholders make on average 0.73% (3-day event window). In the run-up leading to the announcement, which is from 41 days prior- until 2 days prior the announcement date, they make 0.49%. 49% of the deals have a negative CAR. Deals with public targets and paid by equity do worse. Target shareholder make on average 14.61% on the announcement compared to 6.8% during the run-up.
Fuller, Netter and Stegemoller (2002) find an average bidder CAR of 1.8% using a sample of 3,135 takeovers from 1990 until 2000. The CAR is -1.0% when the target is public.
Moeller, Frederik and Stulz (2005) found a average bidder CAR of 1.10% using a sample of 12,023 deals with a total deal value of US$3,413,180 million ranging from 1980 until 2001.
2.3. Cross sectional studies
There are also studies that examine the cross sectional effect of certain variables on the market’s reaction to the announcement. They show that size, method of payment, consideration and insider trading laws have an effect on the market reaction.
2.3.1. Size
The effect of size on the announcement effect is studied extensively. Prior research shows that the announcement return for bidder shareholders is higher if the acquirer is small. According to Moeller et al. (2004) acquiring shareholders of small firms earn on average 2 percentage points more. A reason could be that incentives of managers are better aligned in small firms than large firms. According to Demsetz and Lehn (1985) managers of small firms have more control and ownership. They have less incentive to go empire building. Firms that are large are likely to be at the end of their lifecycle or overvalued. This reduces growth opportunities (Moeller et al. 2004). The synergy advantages are also larger and easier to realize for small companies then large companies. Larger companies will face more integration costs. They also tend to overpay. According to the markets reaction the premium they pay will unlikely be outrun by the cost savings. Moeller et al. (2004)
separate small companies from larger ones by using a dummy variable. Small firms are less than 25% of the size of NYSE firms in the same year. In this paper I will use the natural logarithm to measure the size. I expect the coefficient on ln(Size) to be negative, in line with the findings of Moeller et al. (2004) and Demsetz and Lehn (1985).
2.3.2. Method of payment
The method of payment also has an effect on the bidders announcement return. The method of payment is used as a signalling method. Paying with cash is a signal that suggests to the shareholder that the stock is under-priced (Leland & Pyle, 1977; Meyers Majluf, 1984). Signalling is possible when there is information asymmetry.
According to Travlos & Papaioannou (1991) when the level of leverage is kept constant, cash payments always generate higher bidder announcement returns than equity. Another paper by Traylos (1987) shows that acquiring firms’ stockholders experience losses on the announcement date. Acquiring companies that pay by cash can expect normal rates of return. Therefore I expect that the coefficient on the dummy variable Cash will be positive.
2.3.3. Consideration
Hostile bids are occur when a poor performing board of the target does not want the target to be taken over For a hostile takeover to be successful, the acquirer must convince the shareholders of the target to sell their shares. The acquirer is likely to insinuate a proxy fight. When weakens the board it can exercise more control. The target is likely to prevent the takeover by using takeover defences. Examples of takeover defences are poison pills, priority shares, preference shares, a white knight, golden parachutes and recapitalization (Berk & DeMarzo, 2014, p. 948). Goergen and Renneboog (2004) find that hostile bids are significantly higher than friendly bids. All these extra costs will have negative effects on the bidders return. Goergen and Renneboog (2004) found a negative abnormal return for the bidder of -2.5%. Therefore I also assume that the market reaction of the acquirer will be negative towards hostile bids. I expect that the coefficient on Hostile will be negative.
2.3.4. Cross-border M&A
M&A can take place within one country, domestic, or between firms of two separate countries, cross-border. According to most prior research, domestic- has an advantage over cross-border M&A. Due to cultural differences it is more difficult to integrate two companies that culturally differ from each other, especially after the target has been acquired (Child et al., 2001). A study by KPMG shows that only 17% of the cross-border M&A generates shareholder value (The Economist 2017). According to Morosini et al. (1998) difficulties in the integration process and communication that arise have a negative effect on the shareholders wealth. Ahern et al. (2012) show that the volume of cross-border
M&A is less for countries that are more culturally distant. They also state that greater cultural distance in trust and individualism leads to lower combined announcement returns. Thus create less shareholder value (Bertrand & Zitouna, 2008; Con et al., 2005).
There are also some articles that found positive cross-border performance like Morosini et al. (1998) and Chakrabarti et al. (2009). These articles claim that the cultural difference helps the merged companies gain new perspectives that boost efficiency. Page (2007) supports the idea by stating that greater cultural distance leads to more cultural diversity and that this could give new insight into problem solving ideas that facilitate innovation.
To summarize we can state that most prior research proves that cross-border M&A creates less shareholder value than domestic M&A.
3. Hypothesis
Cross-border deals perform worse due to several reasons according to prior research. Because of cultural differences it is more difficult to integrate two companies that culturally differ from each other, especially after the target has been acquired (Child et al., 2001). Some cultures work more in groups, others more individual. Some cultures do not question authority, for others that is more acceptable. Another one is that geographical distance causes extra integration costs. Synergy gains rely on the coordination after the merger or acquisition. When it is difficult to communicate because of the distance and different time zones it causes loss of efficiency (Ahern et al., 2012). Different languages can cause a barrier, especially when the cultural difference is big (Morosini et al., 1998). In an interview I conducted with TMF group Amsterdam, they gave two other examples apart from language, which still causes a lot of miscommunication according to them. The first is the possibility to extract money from the firm to the owner. In some regions it is difficult to pay dividends to the shareholders. The second example concerns property rights. In some regions owners have little to no rights over their possessions. The local government can take their assets. All these difficulties have a negative effect on the shareholders wealth.
Another possibility instead of cultural distance could be that political instability may cause impediment. Political instability is associated with lower economic growth and lowers the rate of productivity (Aisen & Veiga, 2013; Alesina & Perotti, 1996). I measure political stability by two proxies, Corruption and War.
3.1. Corruption
In countries where corruption is an accepted method of getting business done, companies are put into a difficult position. Because firms that are listed in countries where corruption is prohibited and heavily penalized cannot pay bribes, they have difficulties to work efficiently in countries were corruption is accepted. An example of this is when a container ship of Maersk, a Danish company, goes to a South-American port, it often has to pay a bribe to get in the port without having to wait a few days in the harbour. It can’t cooperate and therefore there is a loss of efficiency (Maersk, 2016). This has a negative effect on the earnings and therefore shareholder value is lost.
According to Mauro (1996) corruption is negatively associated with economic growth and the level of investment. He states that a better bureaucratic efficiency causes high investment and growth. To summarize, I assume that investors are more reluctant to M&A deals with countries that are not corrupt.
3.2. War
Another way to measure political stability is to look if a country is in war. War has consequences for the economic performance. According to Koubi (2005) cross-border differences in economic growth are related to the severity and duration of war. The effects of pre- and post war combined have a negative influence on economic performance. Countries that fought a severe and prolonged war have less economic welfare compared to other countries. However in the long run countries that have incurred a war face post war economic growth. According to Yi Feng (2003) war stagnates economic growth, especially in the long run. She proves this by comparing growth rates of countries with comparable income levels in the 1960 and looks at the annual growth rate from 1960 until 1998. She finds 4.4% in Asia, 2.0% in Latin America and -0.5% in sub-Saharan Africa. She claims the lack of economic prosperity to war and conflicts.
Therefore I expect that countries that are in war lack in economic growth, which has a negative effect on the shareholders return. Therefore the market will react negatively on the takeover of a firm that is located in a country that is in war. I assume the coefficient on War will be negative.
3.3. Hypothesis
Based on the above arguments I predict that more political stability creates additional shareholder value. Assuming that investors are rational, I expect a positive relation between political stability and the market reaction.
H0: More political stability does not lead to a higher CAR
4. Research design
4.1. Method
To test this explanation, I use the event study approach as described by de Jong (2007). This approach assumes market efficiency. If this assumption is valid, then the market reaction can be measured by calculating the abnormal return. To calculate the abnormal return I use the market model. This model is defined as follows:
𝑅!" = 𝛼!+ 𝛽!𝑅!"+ 𝜖!" (1) 𝑅!" = Daily realized return ex dividends
𝛼! = Intercept
𝛽! = The slope of the regression, volatility 𝑅!" = Market return, S&P Global BMI 𝜖!" = Error term
After estimating model (1) I calculate the expected returns using model (2).
𝑁𝑅!" = 𝛼!+ 𝛽!𝑅!" (2)
𝑁𝑅!" = Expected return 𝛼! = Estimated alpha
𝛽!= Estimated beta, volatility
𝑅!" = Benchmark market return, S&P Global BMI
Alpha hat and beta hat are estimated by running a regression for each company using the realized returns and the index returns for the given estimation window of 200 days until 10 days prior to the event date. I use the S&P Global BMI as the index for the market return. This index gives a comprehensive, rules-based index measuring global stock market performance. The estimation window is large enough to make a proper estimation of the volatility of the shares. Most studies use between 180 and 200 trading days, ending 10 or 20 days prior the event date. I do not expect information leakage before 10 days prior to the announcement.
Then I calculate the abnormal return by looking at the residual of the market model. This is simply the realized return minus the expected return.
𝐴𝑅!" = 𝑅!" − 𝑁𝑅!" (3) 𝐴𝑅!" = Abnormal return
𝑅!" = Daily realized return ex dividends 𝑁𝑅!" = Expected return
For the five day event window of [-2,+2] days I cumulate the abnormal returns to calculate the CAR. I use a five-day event window (trading week) because we expect the market to be efficient and react close to the announcement date. Also, to minimize noise and share price effects from other information we keep the event window short.
𝐶𝐴𝑅! = 𝐴𝑅!,!!+ ⋯ + 𝐴𝑅!,!! = !! 𝐴𝑅!"
!!!! (4)
𝐶𝐴𝑅! = Cumulative abnormal return
In the sensitivity analyses event windows of [-5,+5] and [-10,+10] will also be used. This is to look at the change of the coefficients for war and corruption.
After we calculate the CARs we can calculate the cumulative average abnormal return. This is done by a linear regression on all CARs.
𝐶𝐴𝐴𝑅 = !
! 𝐶𝐴𝑅!
!
!!! (5)
The p-value of the CAAR is a better test of significance that the t-test because the regression uses robust standard errors that takes heteroskedasticity into account.
4.1.2 Linear regression
To test the effect of political stability on the cumulative abnormal returns this paper uses 2 measurements and 4 control variables. The regression I use is as follows:
(6) 𝐶𝐴𝑅!" = 𝛼!+ 𝛽!∗ 𝐶𝑜𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 + 𝛽!∗ 𝑊𝑎𝑟 + 𝛽!∗ 𝐼𝑛𝑠𝑖𝑑𝑒𝑟𝑇𝑟𝑎𝑑𝑖𝑛𝑔 + 𝛽!∗ 𝐻𝑜𝑠𝑡𝑖𝑙𝑒 + 𝛽!∗
ln 𝑆𝑖𝑧𝑒 + 𝛽!∗ 𝐶𝑎𝑠ℎ + 𝜖!
CAR is the cumulative abnormal return α0 is the intercept of the regression
Corruption is a dummy variable that is 1 if the government is corrupt. To measure this I use a median split. I use the corruption perception index from 2008 until 2018. I calculate the mean score per country and divide the list based on their means into two groups. The corrupt countries receive a 1, the incorrupt countries receive a 0. I then match the list of targets to the list of the average
corruption over the 10 years. I expect Corruption to have a negative affect as discussed in 3.1. Therefore I expect the coefficient on Corruption to be negative.
War is a dummy variable that is 1 if a country is in war or conflict. I use the Global piece index of 2016 to measure if a country is in war. I use a median split to divide the index into countries that are in war, and countries that are not in war. I expect War to have a negative effect on the acquirer’s earnings as discussed in 3.2. Therefore I expect the coefficient on War to be negative. InsiderTrading is a dummy variable that is 1 if a country doesn’t penalize for insider trading. To
measure InsiderTrading I add up al the columns per country from table 5 from Beny (2005). I then set the median at 7.5. All acquiring countries with a value lower than 7.5 have poor insider trading laws and receive a 1 and all others receive 0. I expect the coefficient on InsiderTrading to be negative. I assume that InsiderTrading has a negative effect on the bidders shareholder return because trading start in advance of the event date. Shareholders will sell if they think the market will react negatively. This paper looks at deals with a large deal value. Prior research shows that larger companies have a negative market reaction (Moeller et al., 2004). Therefore I think the coefficient on InsiderTrading will be negative.
Hostile is a dummy variable for consideration that is 1 if the takeover was hostile. ln(Size) is measure by the natural logarithm of the market capitalization.
Cash is a dummy variable that is 1 if the deal is paid for more than 50% by cash. This method of measuring size is used in most prior research.
εt is the error term of the regression
4.2. Data
4.2.1. Cross-border deals
For my analysis I collect cross-border M&A deals from the first of January 2008 until the first of January 2018. I chose a 10-year period to have enough data that is still relevant. Since 2007 the cross-border market has been stable at around 40% of all M&A deals which makes it a consistent period to research. The cross-border deals are obtained from the Thomson One database. The following filters are used. Deals below 50 million US dollars are excluded. This is to eliminate all smaller deals that often are very volatile and can create noise in the dataset. The acquirer has less than 50% of the share beforehand and gains all shares after the deal to make sure that it really is a M&A deal on which the market reacts and not a constitutional owner buying the last set of shares. The acquirer and target
companies are publically listed to ensure that the stock data is available. The status of the deal is completed. These criteria result in 628 cross-border deals. The corresponding share prices are retrieved from Datastream. For a small number of firms there is not enough stock data available to calculate the alpha and beta in the estimation window. There are also deals that have the announcement date on a day that was not a trading day and therefore are not matched in the Stata process to an event window. Those deals are also omitted. This leaves us with 461 deals. The data is Winsorized via Stata. All outliers outside the 1st and 99th percentile have been adjusted to the value of the 1st and 99th percentile value. The consideration, market capitalization and method of payment are
also retrieved from the Thomson One database. To measure corruption I use the corruption indices from 2008 until 2018 of from transparency.org. I then calculate the mean over the 10 years. To measure if a country is in war I use the Global Peace Index of 2016 from economicsandpeace.org. To measure insider trading I use data from the table Formal Insider Trading Law and Enforceability (Beny, 2005, p. 160).
4.3. Descriptive statistic
From the sample of 628 deals 461 remain in the final regression. This is due to the fact that 108 did not have stock data or because the event date did not match a trading day. Another 59 companies miss enough data to match the estimation window of [-10, -200] days. The following deal characteristics are described in table 1. In the sample there are 2 hostile deals, 299 deals are paid with more than 50% cash, 19 deals are done with target countries that have a corruptive government, 15 deals are made with a target country that is in war or conflict and 45 deals occur where the target country has absent or malfunctioning insider trading laws. Table 2 shows the country breakdown of acquiring countries with the corresponding number of deals. Table 3 shows the country breakdown of target countries with the corresponding number of deals. Most of the acquiring and target firms are U.S. based. Also, 52% of the acquiring firms are native English and even 70% for the target countries.
Table 1 Deal characteristics N Mean Percent Hostile 1 0.22% Cash 299 64.86% Corruption 19 4.12% War 15 3.25% InsiderTrading 45 9.76% ln(Size) 7.592 Table 2 Table 3
Country breakdown of acquiring countries Country breakdown of target countries
Target country No. of transactions
United States 155 Canada 74 United Kingdom 47 Australia 45 France 10 Netherlands 9 Bermuda 8 Israel 8 Norway 8 Norway 8 Singapore 8 Hong Kong 7 South Africa 7 Brazil 6 Guernsey 6 Sweden 5 Finland 4 Italy 4 Luxembourg 4 Turkey 4 Chile 3 Denmark 3 Indonesia 3 Thailand 3 Belgium 2 China 2 Ireland-Rep 2 Jersey 2 Others 22 Total 461
Acquirer country No. of transactions
United States 106 Canada 58 United Kingdom 42 Japan 41 France 24 Sweden 16 Switzerland 16 Germany 12 Italy 10 China 10 Ireland-Rep 10 Netherlands 9 Australia 9 Bermuda 8 Spain 7 South Africa 7 Hong Kong 7 Israel 6 India 5 South Korea 5 Belgium 5 Mexico 4 Brazil 3 Denmark 3 Isle of Man 3 Taiwan 3 Others 34 Total 461
5. Results
5.1. Univariate analysis
Table 4
Correlation matrix
This table presents the correlation between the variables. Corruption and war are proxy variables to measure political stability. Hostile, ln(Size), Cash and InsiderTrading are control variables.
VARIABLES Corruption War Hostile ln(Size) Cash InsiderTrading Corruption 1.000 War 0.1395 1.000 Hostile -0.0084 -0.0075 1.000 ln(Size) 0.0074 0.0504 -0.0250 1.000 Cash -0.1588 -0.0436 0.0295 0.3402 1.000 InsiderTrading 0.1121 -0.0541 -0.0133 -0.0272 -0.1504 1.000
Table 4 represents the correlation matrix. It shows that Cash correlates the most with other variables. The relation between ln(Size) and Cash is the strongest correlation. It is positive because larger companies usually have more cash, and thus buy without having to rely on equity. Firms have an incentive to pay cash because issuing stock gives a bad signal. It is a sign of prosperity when a deal is paid with cash. The manager believes its share price is undervalued. When the deal is paid with equity it reflects the management’s uncertainty concerning the feasibility of potential synergy advantages.
The second largest correlation is between Cash and Corruption. It is negative, which means hat cash is used less as method payment in target countries where there is corruption. The reason for this could be that managers are uncertain if they are able to realize the expected synergy advantages. Therefore they prefer to pay with equity.
The third largest correlation is between InsiderTrading and Cash. This is also negative which means that cash is used less as method of payment in countries that do not have insider trading laws or do not prosecute them. It could be that countries that do not have good insider trading laws also do not have an effective juridical system. This could have negative effects on the efficiency of the takeover. The manager is uncertain if the expected synergy advantages can be realized and prefers to pay with equity.
The relation between Corruption and War is positive. This means that they partially measure the same. But the correlation is small and weak. It has less significant effect than expected.
5.2. Multivariate analysis
Table 5
Cumulative average abnormal returns of the acquirer. Using a linear robust regression on all CARs of the event window of [-2, +2]. (1) VARIABLES CAAR Constant 0.0594 (0.246) Observations 461 R-squared 0.000
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Table 6
The effect of political stability on the cumulative abnormal announcement returns.
Dependent variable is the cumulative abnormal announcement return of the cross-border acquirer during the period [-2, +2] days weighted by the Global S&P500. The announcement dates of the deals lie between 2008 and 2018. The regression of the model is CARit = α0 + β1*Corruption + β2*War +
β3*InsiderTrading + β4*Hostile + β5*ln(Size) + β6*Cash + εt . The measurements for political stability are
Corruption and War. It shows a statistically insignificant effect for Corruption at p<0.10, but in combination with the control variables it is statistically significant at p<0.05 The effect of War is statistically insignificant at p<0.10.
(1) (2) (3) (4) (5)
VARIABLES Corruption War Corruption War Model
Corruption 1.362 2.284** 2.244** (0.918) (0.983) (0.984) Hostile 3.744*** 3.749*** 3.750*** (0.347) (0.347) (0.348) ln(Size) -0.541*** -0.526*** -0.543*** (0.136) (0.136) (0.137) Cash 1.768*** 1.613*** 1.775*** (0.604) (0.596) (0.608) InsiderTrading -0.596 -0.493 -0.581 (0.834) (0.850) (0.839) War 0.181 0.659 0.312 (1.038) (1.058) (1.071) Constant 0.00330 0.0535 2.976*** 3.025*** 2.978*** (0.254) (0.252) (0.998) (0.999) (0.999) Observations 461 461 461 461 461 R-squared 0.003 0.000 0.049 0.043 0.049
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 5 presents the cumulative average abnormal return (CAAR). This is the average of all CARs given the event window of 5 days, [-2, +2] days. The constant is positive, which means that bidder shareholders make a profit of 5.94% after the announcement of the takeover. This is more than what prior research shows as discussed in section 2.2. The CAAR is statistically insignificant and therefore the result cannot be compared directly to the findings of prior research.
Table 6 shows the output of the regression of the cumulative abnormal returns, during the five-day period surrounding the event-date, on the measurements for political stability (War and Corruption) and the control variables (Hostile, Cash and InsiderTrading).
The coefficient on Corruption is positive, which means that the market reaction towards deals with target countries where there is corruption is positive. This suggests that investors think that these deals create more value. The coefficient is statistically not significant on its own, but in combination with the control variables it is statistically significant at α=0.05. This finding is not in line with the expectation. I assumed that corruption would raise the cost of integration and therefore lead to destruction of wealth for the acquiring shareholders (Rahahleh & Wei, 2013) and that more bureaucratic efficiency would create more shareholder value (Mauro, 1996). The reason for the positive coefficient can be that countries where there is corruption are usually countries in emerging markets (MSCI market classification). These emerging markets were performing poorly during the financial crisis. Larger conglomerates were able to gain market power by buying these targets from emerging companies at lower costs. Resulting in wealth creation for the shareholder of the acquirer. In the sensitivity analysis the financial crisis is excluded.
The coefficient on War is also positive. This means that the acquirers that buy target firms in countries that are in war generate more shareholder value opposed to countries that are not in war, although the coefficient is statistically insignificant. This finding is against the expectation. Countries that are in war are most of the time unstable and this makes integrating the firms less efficient.
The first control variable is Hostile. The coefficient on Hostile is positive and statistically significant at p<0.01. This is not in line with the expectation that hostile takeovers usually initiate a loss for the bidder shareholder. The reason for the coefficient being positive is that we only have two hostile deals in the sample. One has a very positive CAR and the other one a negative CAR. Therefore we cannot draw any conclusions form this outcome even though it is statistically significant at p<0.01. The second control variable is ln(Size). The coefficient on this variable is statistically significant at p<0.01. The coefficient is negative which means that larger companies cause a less positive market reaction to takeovers. This suggests that investors think large companies create less value. This corresponds to the expectation made in section 2.3.1. that larger firms face higher integration costs and tend to pay higher premiums, which leads to wealth destruction.
The third control variable is InsiderTrading. The coefficient on this variable is negative. This is in line with the expectations explained in section 2.3.4. It is statistically insignificant at p<0.1. In the sensitivity analysis this variable is explained in more detail.
The last control variable is Cash. The coefficient on this variable is positive. This is in line with the expectations of section 2.3.2. that cash deals trigger more positive market reactions. The coefficient is statistically significant at p<0.01.
5.3. Sensitivity analysis
In this section I will perform three different sensitivity tests. The definition of sensitivity analysis is “the study of how the uncertainty in the output of a model can be apportioned to different sources of uncertainty in the model input” (Saltelli, 2002). The first two tests are to find out if the outcomes remain consistent when the event window changes. I am especially interested in the effects of Corruption and War. Also the change in effect on InsiderTrading will be of interest.
Table 7
Sensitivity analysis for effect of political stability on the cumulative abnormal announcement returns. Dependent variable is the cumulative abnormal announcement return of the cross-border acquirer during the period [-5, +5] days weighted by the Global S&P500. The announcement dates of the deals lie between 2008 and 2018. The regression of the model is CARit = α0 + β1*Corruption + β2*War +
β3*InsiderTrading + β4*Hostile + β5*ln(Size) + β6*Cash + εt . The measurements for political stability are
Corruption and War. It shows a statistically insignificant effect for Corruption at p<0.10. The effect of War is statistically insignificant at p<0.10.
(1) (2) (3) (4) (5)
VARIABLES Corruption War Corruption War Model
Corruption 1.231 1.991 2.057 (1.479) (1.540) (1.554) Hostile -4.666*** -4.677*** -4.676*** (0.493) (0.495) (0.496) ln(Size) -0.599*** -0.579*** -0.595*** (0.198) (0.197) (0.199) Cash 1.349* 1.189 1.338* (0.793) (0.779) (0.794) Insidertrading -0.697 -0.641 -0.721 (1.075) (1.108) (1.079) War -0.628 -0.202 -0.520 (1.456) (1.480) (1.527) Constant 0.435 0.506 4.153*** 4.193*** 4.150*** (0.350) (0.349) (1.439) (1.437) (1.439) Observations 461 461 461 461 461 R-squared 0.001 0.000 0.026 0.023 0.026
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 7 looks at the event window of 11 days, [-5, +5]. The coefficient on Corruption is still positive. It is statistically not significant, also not in combination with the control variables. The coefficient on War is negative. It is statistically not significant. The coefficient of War is now in line with the expectation stated in section 3.2. The reason for the change of the coefficient from positive to negative is unclear. Because the coefficient is statistically insignificant there is unclear whether the change is caused by new information.
The coefficients on the variables ln(Size), Cash and InsiderTrading in table 7 remain approximately the same as table 6. Although Cash is only statistically significant at α=0.1. The coefficient on Hostile shifts from positive in table 6, to negative in table 7. This must be because of new information after the event date, which suggests that investors think it will destroy value.
Table 8
Sensitivity analysis for effect of political stability on the cumulative abnormal announcement returns. Dependent variable is the CAR of the cross-border acquirer during the period [-10, +10] days weighted by the Global S&P500. The announcement dates of the deals lie between 2008 and 2018. The regression of the model is CARit = α0 + β1*Corruption + β2*War + β3*InsiderTrading + β4*Hostile + β5*ln(Size) +
β6*Cash + ε. The measurements for political stability are Corruption and War. It shows a statistically insignificant effect for Corruption at p<0.10, but in combination with the control variables it is statistically significant at p<0.05. The effect of War is statistically insignificant at p<0.10.
(1)
(2)
(3)
(4)
(5)
VARIABLES
Corruption
War
Corruption
War
Model
Corruption
1.632
2.975**
3.329**
(1.268)
(1.386)
(1.402)
Hostile
-7.087***
-7.142***
-7.140***
(0.657)
(0.660)
(0.660)
ln(Size)
-0.603**
-0.559**
-0.584**
(0.259)
(0.261)
(0.261)
Cash
2.408**
2.107**
2.349**
(1.076)
(1.046)
(1.073)
Insidertrading
-2.065
-2.071
-2.200*
(1.288)
(1.315)
(1.292)
War
-2.670
-2.296
-2.810
(2.470)
(2.428)
(2.483)
Constant
-0.0907
0.0638
3.089
3.146
3.074
(0.461)
(0.452)
(1.994)
(1.986)
(1.993)
Observations
460
460
460
460
460
R-squared
0.001
0.002
0.027
0.026
0.030
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Model 8 examines the event window of 21 days, [-10, +10]. The effect of corruption is still positive and in combination with the control variables statistically significant at α=0.05. The effect of war is increasingly negative compared to table 6 and 7. A possible explanation could be that investors are unaware of the situation at first, but later on expect the negative influence that war causes and the market reacts.
Comparing table 6, 7 and 8 the effect of InsiderTrading is more significant in a longer event-window. The reason for this could be that insider traders trade in advance of the event date and so the effect becomes more significant once you include more trading days prior to the event date.
Next I examine the effect of political stability on cross-border performance, but excluding the merger wave peak in the beginning of the financial crisis (appendix 2). During and just after the financial crisis investors are willing to invest in politically risky securities, but will only invest at high lemon discounts (Hill, 1998). This means that CARs are unstable in the period up to 2011. To be sure I exclude this period, I look at the cross-border deals from 1-1-2012 until 1-1-2018. Appendix 2 shows that from 2012 onwards the M&A market is stable.
The output of the regression is presented in table 9. The coefficient on Corruption is negative which is in line with the expectation stated in section 3.1. It is statistically insignificant at p<0.1. The coefficient on War is negative when regressed solely, but positive in combination with the control variables. It is statistically insignificant.
The control variable Hostile is not included in the model because the two hostile deals in the dataset occurred before 2012. The coefficient on ln(Size) is negative which is in line with the expectation. It is statistically significant at p<0.01. The coefficient on Cash is positive, which is in line with the expectation. It is statistically significant at p<0.05. The coefficient on InsiderTrading is negative which is in line with the expectation. It is statistically insignificant.
Table 9
The effect of political stability on the cumulative abnormal announcement returns.
Dependent variable is the CAR of the cross-border acquirer during the period [-2, +2] days weighted by the Global S&P500. The announcement dates of the deals lie between 2012 and 2018. The regression of the model is CARit = α0 + β1*Corruption + β2*War + β3*InsiderTrading + β4*Hostile + β5*ln(Size) +
β6*Cash + ε. The measurements for political stability are Corruption and War. It shows a statistically
insignificant effect for Corruption at p<0.10. The effect of War is statistically insignificant at p<0.10.
(1)
(2)
(3)
(4)
(5)
VARIABLES
Corruption
War
Corruption
War
Model
Corruption
-1.377
-0.851
-0.844
(1.291)
(1.168)
(1.174)
ln(Size)
-0.604***
-0.610***
-0.606***
(0.188)
(0.191)
(0.192)
Cash
1.716**
1.734**
1.721**
(0.811)
(0.814)
(0.818)
Insidertrading
-1.016
-1.020
-1.010
(1.206)
(1.219)
(1.215)
War
-0.695
0.204
0.181
(1.169)
(1.363)
(1.366)
Constant
0.0980
0.0882
3.710**
3.721**
3.723**
(0.332)
(0.334)
(1.446)
(1.450)
(1.452)
Observations
276
276
276
276
276
R-squared
0.001
0.000
0.054
0.053
0.054
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
6. Conclusion
This paper analyses the effect of political stability on the acquirers shareholder return during the period of 2008 until 2018. I have studied 461 cross-border deals from all over the world. To measure political stability I used two proxies, namely War and Corruption. The control variables size, method of payment, consideration and insider trading laws are used. To look at the announcement effect I performed an event study with an event window of five days, [-2, +2]. I calculated the abnormal returns for each firm and cumulated them for the given event window. Then I performed a regression to look at the effects of the proxies that measure political stability.
The results showed that more corruption and war caused a positive market reaction. The reason for this could be that the countries that were in war or had corruption are emerging markets with high potential growth. Also the fact that during the financial crisis many firms were bought at ‘lemon’ prices could relate to the fact that countries with war and corruption cause a positive market reaction. The results of the control variables ln(Size), Hostile and Cash were in line with the expectations. To answer the research question: I do not have enough evidence to state that more political stability leads to a positive market reaction.
I did find that for the deals from 2012 until 2018 corruption and war have negative effects, which is in line with the hypothesis. But these findings were not statistically significant.
Limitations to the model I used are that there might be omitted variables because I used only two proxies to measure political stability. Also I had limited deals. Gathering more data from more different countries would make the model more precise. Another limitation is the fact that I used dummy variables. This means that countries that are just on the edge of being corrupt or not corrupt are measured as either corrupt or not corrupt, the same counts for War and InsiderTrading. The last limitation to the model is that I used the peace index that corresponds only to 2016 to measure War. Although most wars last for a long time, this still can cause erroneous results.
Further research on this topic can be done with a larger sample that contains more countries that have corruption or are in war. Also a model with better proxies to measure political stability adds value to this topic. Another area to investigate is to examine if there are other variables that have on effect on cross-border M&A that have not been studied yet.
7. References
Accio, M. and Masulis, R. W. (2005), The Choice of Payment Method in European Mergers and Acquisitions. The Journal of Finance, 60: 1345-1388.
Ahern, Daminelli, Fracassi, (2012). Lost in translation? The effect of cultural values on mergers around the world. Journal of Financial Economics. 117(2015) pp. 165-189.
Aisen, A. & Veiga, F.J. (2013). How does political instability affect economic growth? European Journal of Political Economy, Vol.29, pp.151-167.
Al Rahahleh, N. & Wei, P.P. (2013). Frequent cross-border acquirers from emerging countries and cultural distance: Does the cultural difference of the initial deal matter? Journal of Multinational Financial Management. Vol.(23) pp. 356-373.
Alesina, A. & R. Perotti. (1996). Income distribution, political instability, and investment. European Economic Review, 40(6) pp. 1203-1228.
Baynes, T. (2016). No threats and a lot less shouting. Retreived on June 13th 2018, of https://www.maersk.com/stories/no-threats-and-a-lot-less-shouting
Beny, Laura Nyantung, (2005). Do Insider Trading Laws Matter? Some Preliminary Comparative Evidence. American La wand Economics Review. Vol(7) pp. 144-183.
Berk, J. & DeMarzo, P. (2014) Corporate Finance (3rd edition). Pearson.
Bertrand,O. & Zitouna, H. (2008). Domestic verus cross-border acquisitions: Which Impact on the Target Firms’ Performance? Applied Economics, Taylor and Francis Journals, Vol 40(17), pp. 2221-2238.
Child, J., Falkner, D. & Pitkethly, R. (2001). The Management of International Acquisitions. Oxford University Press, Oxford, UK.
De Boer, M. (2018). Markten anticiperen op een handelsoorlog. Retrieved on June 20th 2018, of https://fd.nl/economie-politiek/1258819/markten-anticiperen-op-een-handelsoorlog
Demsetz, H., Lehn, K., (1985). The structure of corporate ownership: causes and consequences. Journal of Political Economy, pp. 1155-1177.
Eckbo, B. E. (2008). Handbook of Corporate Finance: Empirical Corporate Finance. Elsevier/North-Holland Handbook of Finance Series, Vol. 2 pp. 363-368.
Fuller, K., J. Netter and M. Stegemoller, (2002). What do returns to Acquiring Firms Tell Us? Eidence from Firms that Make Many Acquisitions, Journal of Finance, 57, pp. 1763-1793.
Goergen, M. & Renneboog, L. (2004), Shareholder Wealth Effects of European Domestic and Cross -border Takeover Bids. European Financial Management, 10: pp. 9-45.
Hill, Claire A., (1998). How Investors React to Political Risk. Duke Journal of Comparative and International Law, Vol. 8, No. 2.
Institute for Economics and Piece. (2017) Global Piece Index 2016. Retrieved on June 10th 2018, of http://economicsandpeace.org/wp-content/uploads/2016/06/GPI-2016-Report_2.pdf
Leland, H., and D. Pyle, (1977). "Information Asymmetries, Financial Structure and Financial Intermediaries," Journal of Finance, 32, pp. 371-387.
M. Paolo, (1995). Corruption and growth. Quarterly Journal of Economics, Vol.110(3), p.681(32). Meyers, S.C., and Ν J. Majluf, (1984). Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have. Journal of Financial Economics, Vol.13, pp. 187- 221 Moeller, S. B., Schlingemann, F. P. and Stultz, R. M. (2005). Wealth Destruction on a Massive Scale? A Study of Acquiring-Firm Returns in the Recent Merger Wave. The Journal of Finance, 60: 757-782.
Nickolaos G. Travlos and George J. Papaioannou. (1991). Corporate Acquisitions: Method of
Payment Effects, Capital Structure Effects, and Bidding Firms Stock Returns. Journal of Business and Economics, Vol.30(4) pp. 3-22.
Saltelli A., (2002). Sensitivity Analysis for Importance Assessment, Risk Analysis, 22(3): pp. 579-590.
Sara B Moeller, Frederik P Schlingemann, René M Stulz, (2004). Firm size and the gains from acquisitions, Journal of Financial Economics, Volume 73, Issue 2, pp. 201-228.
Transparancy International. (2018) Corruption perception index. Retrieved on June 10th 2018, of https://www.transparency.org/news/feature/corruption_perceptions_index_2017
Traylos, Nickolaos. (1987). Corporate takeover bids, methods of payments and budding firms’ stock return. Journal of Finance. Vol.42(2) pp. 943-963.
Van Rijn, E. (2018). Geopolitiek laat de belegger koud. Retrieved on June 16th 2018, of https://fd.nl/beurs/1258203/geopolitiek-laat-de-belegger-koud
Vally Koubu, (2005). War and Economic Growth. Journal of Piece Research. Vol.42(1) pp. 67-82. Yi Feng, (2003). Democracy, Governance, and Economic Performance: Theory and Evidence. By Yi Feng. Cambridge, Mass.: MIT Press. Pp. xiii+383.
8. Appendix
1. M&A activity
