Policy uncertainty effect on cross-border mergers and acquisitions

(1)

Policy Uncertainty Effect on Cross-Border

Mergers and Acquisitions

Master Thesis Inge Dontje (10783474) University of Amsterdam Amsterdam Business School

MSc Finance: Duisenberg Honours Programme in Corporate Finance & Banking Supervisor: Jeroen Ligterink

June 2018

Abstract

Using the Policy Uncertainty Index from Baker, Bloom and Davis (2016) as a measure of policy uncertainty, this study examines the effect of policy uncertainty on cross-border Mergers & Acquisitions (M&A) between 1995 and 2016. The results provide evidence of a negative relation between policy uncertainty and cross-border M&A. An increase in policy uncertainty in the acquirer country significantly decreases outbound M&A. This supports the notion that the value of the option to delay the investment increases. Furthermore, an increase in policy uncertainty in the target country deceases inbound M&A because foreign firms are deterred to engage in inbound M&A. To mitigate endogeneity concerns, this study controls for various economic conditions and other macroeconomic uncertainty. This study also includes an instrument for policy uncertainty where we use the Partisan Conflict Index from Azzimonti (2018).

(2)

2

Statement of originality

This document is written by Student Inge Dontje 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.

Inge Dontje June 2018

(3)

3

Acknowledgement

I would first like to thank my thesis supervisor, Mr. Jeroen Ligterink, of the University of Amsterdam, Amsterdam Business School. His door was always open whenever I ran into a trouble spot or had questions about my approach, research or writing. He consistently allowed this paper to be my own work, but steered me in the right direction whenever he thought I needed it.

The internship opportunity I had with Ernst & Young was a great chance for learning and professional development. I consider myself as a lucky individual as I was provided with an opportunity to be part of the Ernst & Young organization during my internship period.

Last but not least, to my family, thank you for encouraging me in all my pursuits and inspiring me to follow my dreams. I am especially grateful to my parents, who supported me

(4)

4

1. Introduction

The election of Trump in the United States (US), the Brexit referendum in the United Kingdom (UK) and the newly elected government of Italy are recent shown cases of the importance of political uncertainty on the global economy. For example, the Brexit decreased the UK M&A activity in 2016 by 20% (Deloitte, 2016). The increased policy uncertainty

negatively influences the global economy (Baker, Bloom and Davis, 2016; Bloom et al., 2014). Especially the uncertainty around taxes, government spending and monetary & regulatory policy influences the business environment of firms (Bonaime, Gulen and Ion, 2017). Hence, policy uncertainty has real implications.

This study focuses on the impact of policy uncertainty on cross-border Mergers and

Acquisitions (M&A) activity. The border M&A activity is growing worldwide. The cross-border M&A activity increased from 23% of the total M&A activity in 1998 (Erel, Liao and Weisbach, 2012) to 38% of the total M&A activity in 2016 (Thomson Reuters, 2016). It is interesting to analyze the impact of policy uncertainty on cross-border M&A. Policy uncertainty increases the uncertainty about the value of the merged firm and the value of the synergies which makes policy uncertainty an important risk. Regarding the increasing importance of cross-border M&A activity and the concerns about the increased policy uncertainty, this study tries to get a better understanding of how policy uncertainty effects the decision to engage in cross-border M&A.

This study will analyze the effect of policy uncertainty measured by the Policy Uncertainty Index from Baker, Bloom and Davis (2016) on cross-border M&A. The Policy Uncertainty Index from Baker, Bloom and Davis (2016) measures the policy-related economic uncertainty in a country. This study includes an empirical analysis with 207,890 deals between 1995 and 2016. The results show negative effects of policy uncertainty on real economic outcomes. The results contribute to the literature since it provide evidence of a negative relation between the policy uncertainty in the acquirer country and outbound M&A as well as the policy uncertainty in the target county and inbound M&A. The negative effect of policy uncertainty in the acquirer country on outbound M&A supports the notion that the value of the option to delay the investment increases. Moreover, the negative effect of policy

(7)

7

uncertainty in the target country on inbound M&A supports the notion that foreign firms are deterred to engage in inbound M&A. A potential concern about the Policy Uncertainty Index from Baker, Bloom and Davis (2016) is endogeneity concerns due to omitted variable bias. The Policy Uncertainty Index may capture other macroeconomic conditions which influences M&A activity but which are not observed in this study. To mitigate omitted variable bias, this study includes various macroeconomic control variables. Furthermore, this study includes an extra macroeconomic control analysis. The Macroeconomic Uncertainty Index from Jurado, Ludvigson and Ng (2015) is used as a measures of the macroeconomic uncertainty in the United States. To further mitigate endogeneity concerns, this study includes an instrument for policy uncertainty where we use the Partisan Conflict Index from Azzimonti (2018) to obtain an exogenous measure of policy uncertainty. This index measures the disagreement of politicians about the policy.

The results contribute to two strands in the literature, namely the determinants of cross-border M&A and the effect of policy uncertainty on economic outcomes. This study

contributes to the literature on the determinants of cross-border M&A since it will focus at another determinant, namely policy uncertainty. This is often ignored in other empirical analyses. For example in the studies by Erel, Liao and Weisbach (2012) and Rossi and Volpin (2004). Furthermore, this study contributes to the literature of policy uncertainty on

economic outcomes. The economic outcome that will be examined in this study is the investment level. Typically, M&A is an important and risky investment which is hard to reverse (Nguyen and Phan, 2017). This study will contribute to the literature since it provide evidence on negative effects of policy uncertainty on real economic outcomes.

There are some other recent papers by Bonaime, Gulen and Ion (2017), Cao, Li and Liu (2017), Chen et al (2017) and Nguyen and Phan (2017) that analyze the effect of policy uncertainty on M&A activity. There are two main types of policy uncertainty measures used in those studies. The first one is the policy uncertainty around national elections and the second one is the Policy Uncertainty Index from Baker, Bloom and Davis (2016). This study contributes to the recent papers by Bonaime, Gulen and Ion (2017), Cao, Li and Liu (2017), Chen et al (2017) and Nguyen and Phan (2017) since it provides evidence of a negative effect of the policy uncertainty, measured with the Policy Uncertainty Index from Baker, Bloom and

(8)

8

Davis (2016) in the acquirer country on outbound M&A as well as the policy uncertainty in the target county on inbound M&A. This study contributes to the literature on outbound M&A since it will include acquirers inside the United States as well as acquirers outside the United States. Therefore, the sample is a better proxy for the global M&A activity which provides better insides on the effect of policy uncertainty on cross-border M&A activity. Furthermore, this study will be the first study examines the effects of policy uncertainty, measured by the Policy Uncertainty Index from Baker, Bloom and Davis (2016) on inbound M&A. The study by Cao, Li and Liu (2017) examines the effect of policy uncertainty measured by national elections on inbound M&A. However, the Policy Uncertainty Index from Baker, Bloom and Davis (2016) measures the policy-related uncertainty also outside the national elections years which could be important due to the high variation in M&A activity outside the national election years (Bonaime, Gulen and Ion, 2017).

The remainder of the study proceeds as follows: Section 2 discusses the literature and the derived hypotheses. Section 3 explains the methodology. Section 4 describes the data and presents the descriptive statistics. Section 5 presents the results of the analysis. Section 6 presents the robustness check and additional results. Eventually, Section 7 concludes the study which includes the discussion, the limitations and the implementation.

(9)

9

2. Literature review

This section discusses the literature and the derived hypotheses. This study contributes to two main theories in the literature, namely the determinants of cross-border M&A and the effect of policy uncertainty on economic outcomes.

2.1 Determinants of M&A activity

There is growing literature that studies the determinants of M&A activity. Understanding the determinants of M&A activity is important since M&A activity plays an important role in the efficiently allocation of capital (Bonaime, Gulen and Ion, 2017). There is evidence that M&A activity enhances innovation (Bena and Li, 2014), synergies and shareholder value (Devos, Kadapakkam & Krishnamurthy, 2009; Sheen, 2014). However, there is also evidence that M&A activity is driven by CEO preferences (Jenter and Lewellen, 2015). Compared with domestic M&A, there are other important determinants which plays a role in the decision to engage in cross-border M&A.

The difference between domestic and cross-border M&A is that cross-border M&A is influenced by both the environment in the acquirer country as well as the environment in the target country. There is evidence that cultural differences play a role in cross-border M&A (Ahern et al., 2015). The cultural differences can create barriers to efficiently work with each other. Therefore, the differences in culture between the acquirer county and target country increase the cost of a border M&A. An increase in cost will lead to fewer cross-border M&A between the acquirer country and target country (Ahern et al., 2015; Lim, Makhija & Shenkar, 2016). Also, the geography plays a role in cross-border M&A. The distance between the acquirer and the target may create a barrier to efficiently work together. When the distance between the acquirer and target increases, the cost of cross-border M&A will increase which leads to less cross-cross-border M&A (Bick et al., 2017; Erel, Liao and Weisbach, 2012).

There is also further evidence that relative values plays an important role in M&A (Shleifer and Vishney, 2003; Rhodes-Kropf and Viswanathan, 2004; Rhodes-Kropf, Robinson and Viswanthan, 2005; Dong et al., 2006). Firms which are overvalued are more likely to be the

(10)

10

acquirer, while firms which are undervalued, or relatively less overvalued, are more likely to be the target. Therefore, the stock market plays an important role in M&A activity (Shleifer and Vishney, 2003). The currency also influence the relative values in cross-border M&A. The firms in countries with increased stock market or appreciated currency are acquirers, while the companies in the countries with decreased stock market values or depreciated currency are more often targets. In the situation where the currency appreciates, which is unrelated to firms profitability, the firm may find targets in other countries relatively cheaper.

Therefore, the currency appreciation, decreases the cross-border target valuation, which increases the number of profitable cross-border M&A’s (Erel, Liao and Weisbach, 2012).

There also is evidence that regulation plays an important role in M&A activity (Rossi and Volpin, 2004; Karolyi and Taboada, 2015). Rossi and Volpin (2004) provide evidence that better accounting standards and stronger shareholder protection are important

determinants of cross-border M&A. Furthermore, they state that governance could play an important role in cross-border M&A. They provide evidence that acquirers are typically from countries with stronger investors protection than their targets.

There is also literature which provides evidence that economic and regulatory shocks affect M&A activity (Harford, 2005). Bonaime, Gulen and Ion (2017) support these findings and state that uncertainties related to taxes, government spending and monetary & regulatory policy are important factors for M&A activity. Duchin and Schmidt (2013) and Grafinkel and Hankins (2011) provide evidence of a positive relation between uncertainty and M&A activity. An increase in uncertainty increases M&A activity which are driven by risk

management motives (Grafinkel and Hakins, 2011). The risk management theory predicts that an increase in uncertainty will increase the M&A activity. The policy uncertainty in the acquirer country negatively influences the environment of the acquirer. Since the cross-border target is less influenced by the policy uncertainty in the acquirer country, the acquirer could hedge against policy uncertainty by engaging in outbound M&A.

This study will contribute to the literature since it will focus on another determinant, namely policy uncertainty. Especially, the uncertainty around taxes, government spending and monetary & regulatory policy (Bonaime, Gulen and Ion, 2017).

(11)

11

2.2 Policy uncertainty on economic outcomes

There is growing literature that studies the effect of policy uncertainty on economic outcomes (Bloom et al., 2014; Baker, Bloom and Davis, 2016; Julio and Yook, 2016).

The literature provides evidence of a negative relation between policy uncertainty and macro-economic performance (Baker, Bloom and Davis, 2016; Bloom et al., 2014; Julio and Yook, 2016). The literature shows that policy uncertainty affects the business cycle, where a policy uncertainty shock leads to a large drop in GDP and slower recovery (Bloom et al., 2014). This indicates that higher policy uncertainty is associated with lower economic growth and slower recovery. Baker, Bloom and Davis (2016) showed that an increase in policy

uncertainty decreases the macroeconomic performance. This can be explained by the real option theory. The real option theory predicts that an increase in policy uncertainty will decrease investments because the option to delay the investment becomes more valuable. There is also evidence that policy uncertainty affects cross-border capital flows. An increase in policy uncertainty leads to a decrease in Foreign Direct Investment (FDI), while a decrease in policy uncertainty leads to an increase in FDI. These findings are in line with the theory that increased policy uncertainty deters foreign investments (Julio and Yook, 2016). The deterrence theory predicts that an increase in policy uncertainty will decrease investments because higher policy uncertainty increases the uncertainty of investments returns which deters foreign investors.

There is also literature that studies the effect of policy uncertainty on asset prices. The literature shows that policy uncertainty decreases stock prices (Liu, Shu and Wei, 2017; Pastor and Veronesi, 2012) and increases stock prices volatility (Baker, Bloom and Davis, 2016; Pastor and Veronesi, 2012; Pastor and Veronesi, 2013). Furthermore, policy uncertainty increases risk premia (Pastor and Veronesi, 2013) and is priced in the equity option market (Kelly, Pastor and Veronesi, 2016).

This study is mostly in line with the literature that studies the effect of policy uncertainty on corporate finance decisions such as the decisions to invest or to issue equity. The literature shows that policy uncertainty negatively influences firms investment levels (Baker, Bloom and Davis, 2016; Julio and Yook, 2012; Gulen and Ion, 2015; Jens, 2017) and R&D expenses

(12)

12

(Atanassov, Julio and Leng, 2016). Policy uncertainty also negatively influences the decision to issue equity, like initial public offerings (IPO) (Colak, Durev and Qian, 2017) and seasoned equity offerings (SEO) (Jens, 2017).

2.3 Policy uncertainty and M&A

There are some recent papers that analyze the effect of policy uncertainty on M&A activity. In the literature, there are two types of policy uncertainty measures being used. Namely, the policy uncertainty around national elections and the Policy Uncertainty Index from Baker, Bloom and Davis (2016). The Policy Uncertainty Index from Baker, Bloom and Davis (2016) measures the policy-related economic uncertainty in a country. It captures the policy uncertainty not only in the national elections years but also the policy uncertainty in other years which could play an important role in the decisions to engage in cross-border M&A.

The study by Nguyen and Phan (2017) studies the effect of policy uncertainty which is measured by the Policy Uncertainty Index from Baker, Bloom and Davis (2016) on M&A activity. They focused on M&A activity, the times it takes to complete the M&A, the payment method used in the M&A and the acquirer and target value. They provide evidence that policy uncertainty has a negative effect on M&A activity and the time it takes to complete the M&A. Furthermore, higher policy uncertainty increases the attractiveness to use stock as a payment method and decreases bid premiums.

Bonaime, Gulen and Ion (2017) also find the negative relation between the Policy

Uncertainty Index from Baker, Bloom and Davis (2016) and M&A activity. The most important effects of policy uncertainty are the increased uncertainty on taxes, government spending and monetary & regulatory policies. In addition, they include a part which mainly focuses on outbound M&A. They provide evidence that policy uncertainty increases the likelihood to engage in outbound M&A which is in line with the risk management theory. Yet Bonaime, Gulen and Ion (2017) only look at the policy uncertainty in the acquirer country on outbound M&A while this study also look at the policy uncertainty in the target country on inbound M&A.

(13)

13

Chen et al. (2017) studies the effect of policy uncertainty measured by uncertainty around national elections on M&A activity. They provide evidence that the M&A activity decreases when there is high policy uncertainty. However, they didn’t make a distinction between domestic and cross-border M&A.

Cao, Li and Liu (2017) focused on the effect of policy uncertainty which is measured by the uncertainty around national elections on cross-border M&A. In cross-border M&A, both the policy uncertainty in the acquirer country and the policy uncertainty in the target country influence the outcomes of the M&A. Therefore, they studied the effect of policy uncertainty on outbound M&A as well as on inbound M&A. They provide evidence that policy

uncertainty in the acquirer country increases outbound M&A which is in line with the risk management theory. On the other hand, the policy uncertainty in the target country decreases inbound M&A which is in line with the deterrence theory. To the best of my knowledge, the study by Cao, Li and Liu (2017) is the first study which examine policy uncertainty on inbound M&A activity.

This study will analyze the effect of policy uncertainty measured by the Policy Uncertainty Index from Baker, Bloom and Davis (2016) on cross-border M&A. Since M&A activity is influenced by both the policy uncertainty in the acquirer and the target country, this study examines the effect of policy uncertainty on outbound as well as on inbound M&A.

This study will contribute the literature with regards to the effect of policy uncertainty on cross-border M&A. The effect of policy uncertainty measured by the Policy Uncertainty Index from Baker, Bloom and Davis (2016) on outbound M&A will include acquirers inside the United States as well as acquirers outside the United States. Bonaime, Gulen and Ion (2017) and Nguyen and Phan (2017) only include acquirers from the United States. However, this is unrepresentative for the global M&A activity since the majority of the global M&A activity involves countries outside the United States (Erel, Liao and Weisbach, 2012). Since this study includes countries from different continents and with high M&A activity, our sample is a better proxy for the global M&A activity which provides better insides on the effect of policy uncertainty on cross-border M&A activity. To the best of my knowledge, this study will be the first study that studies the effects of policy uncertainty, measured by the Policy Uncertainty

(14)

14

Index from Baker, Bloom and Davis (2016) on inbound M&A. The study by Cao, Li and Liu (2017) examines the effect of policy uncertainty measured by national elections on inbound M&A. However, the Policy Uncertainty Index from Baker, Bloom and Davis (2016) measures the policy-related uncertainty also outside the national elections years which could be important due to the high variation in M&A activity outside the national election years (Bonaime, Gulen and Ion, 2017).

2.4 Hypotheses

There are three main theories about the effect of policy uncertainty on cross-border M&A.

The first theory argues that policy uncertainty in a country influences the business

environment which increases the risk of the firms (Baker, Bloom and Davis, 2016; Bloom et al., 2014). To reduce this risk, there is evidence that firms engage in international

diversification (Fatemi, 1984; Rugman, 1976). The literature shows that an increase in policy uncertainty increases the risk of the firm and increases foreign investments (Lensink, Hermes and Murinde, 2000). Grafinkel and Hankins (2011) provide evidence that uncertainty

increase M&A activity which is in line with the risk management theory. The acquirer can hedge against the policy uncertainty risk by acquiring targets which are less influenced by policy uncertainty. Bonaime, Gulen and Ion (2017) and Cao, Li and Liu (2017) also provide evidence in line with the risk management theory. The higher policy uncertainty in the acquirer country will increase the risk on the acquirer’s business environment. It is likely that the cross-border targets are less influenced by the policy uncertainty risk in the acquirer country. Therefore, the acquirer could hedge against this risk by engaging in outbound M&A. The risk management hypothesis predicts that an increase in policy uncertainty will increase outbound M&A to hedge against the risk.

The second theory relates to the real option theory. Cross-border M&A is an important and risky investment which is hard to reverse (Nguyen and Phan, 2017). Irreversibility is

important because it increases the value of the information to delay the investment (Gulen and Ion, 2016). There is evidence that policy uncertainty increases the incentives to delay the investments which are hard to reverse (Rodrik, 1991; Bernanke, 1983). The recent literature provides evidence of a negative relation between policy uncertainty and M&A activity which

(15)

15

is in line with the real option theory (Bonaime, Gulen and Ion, 2017; Chen et al., 2017; Nguyen and Phan, 2017). Since cross-border M&A is an important and risky investment which is hard to reverse, the acquirer could prefer to delay the investment. Therefore, the real option hypothesis predicts that an increase in policy uncertainty decreases the outbound M&A because the option to delay the investment becomes more valuable.

The third theory is the deterrence theory. Col, Durnev and Molchanov (2018) provide evidence that the policy uncertainty of the foreign target country influences the risk of the acquirer. Higher policy uncertainty in the target country could lead to a higher likelihood of expropriation (Julio and Yook, 2016). The higher likelihood of expropriation includes direct and indirect effects. The direct effect of expropriation is the seizes of private assets by the government. The indirect effects of expropriation are excessive taxation, overregulation, manipulated exchange rates or capital controls (Cao, Li and Liu, 2017). These factors increase the risks of the merged firm. Cao, Li and Liu (2017) provide evidence which is in line with the deterrence theory. Since M&A is influenced by both the policy uncertainty in the acquirer and target country, the increase in policy uncertainty in the target increases the risk of the merged firm. Policy uncertainty increases the uncertainty about the value of the merged firms and the value of the synergies. Therefore, the deterrence hypothesis predicts that an increase in policy uncertainty in the target country decreases inbound M&A because the increased risk will deter foreign acquirers.

The three hypotheses that will be tested in this study are the risk management hypothesis, the real option hypothesis and the deterrence hypothesis.

Hypothesis 1: an increase in policy uncertainty in the acquirer country increases outbound M&A activity.

The first hypothesis that will be tested is the risk management hypothesis. The coefficient of policy uncertainty in the acquirer country is used to test this hypothesis. The risk

management hypothesis state that the coefficient is positive which means that an increase in policy uncertainty increases the outbound M&A activity.

(16)

16

Hypothesis 2: an increase in policy uncertainty in the acquirer country decreases outbound M&A activity.

The second hypothesis that will be tested is the real option hypothesis. The real option hypothesis state that the coefficient of policy uncertainty in the acquirer country is negative which means that an increase in policy uncertainty in the acquirer country decreases the outbound M&A activity.

Hypothesis 3: an increase in policy uncertainty in the target country decreases inbound M&A activity.

Furthermore, the third hypothesis is the deterrence hypothesis. The deterrence hypothesis state that the coefficient of policy uncertainty in the target country is negative which means that an increase in policy uncertainty in the target country decreases the inbound M&A activity. The three hypotheses are summarized in Table 1.

Table 1: the three hypotheses

Policy uncertainty ↑ Policy uncertainty ↑

Outbound M&A ↑ Risk management

hypothesis

Inbound M&A ↑ Outbound M&A ↓ Real option

hypothesis

Inbound M&A ↓ Deterrence hypothesis

(17)

17

3. Methodology

This paragraph explains the empirical models to study the effect of policy uncertainty on cross-border M&A.

3.1 Outbound M&A

The following model is used to examine the effect of policy uncertainty in the acquirer country on the outbound M&A activity of the acquirer country:

𝐿𝑛(𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑢𝑡𝑏𝑜𝑢𝑛𝑑 𝑀&𝐴 𝑝𝑒𝑟 𝐴𝑐𝑞𝑢𝑖𝑟𝑒𝑟 𝐶𝑜𝑢𝑛𝑡𝑟𝑦)_𝑖,𝑡

= 𝛼 + 𝛽₁𝑃𝑜𝑙𝑖𝑐𝑦 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝐴𝑐𝑞𝑢𝑖𝑟𝑒𝑟 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝑖,𝑡 + 𝛾 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠_𝑖,𝑡+ 𝛿_𝑡+ 𝜀_𝑖,𝑡

where the dependent variable is the natural logarithm of one plus the number of outbound M&A. This dependent variable is commonly used in the literature to measure M&A volume, for example by Ahern et al. (2015), Cao, Li and Liu (2017) and di Giovanni (2005). The independent variables are the natural logarithm of one plus the average Policy Uncertainty Index in the acquirer country and the control variables of the acquirer country. Based on the literature, this study controls for level of shareholder protection, economic conditions, risk of investments, quality of institution and trade openness. Furthermore, the model will control for the year before an election since Cao, Li and Liu (2017) provide evidence that the year before an election influences outbound M&A. Including the year before an election into the model allow us to measure the overall policy uncertainty. The Policy Uncertainty Index will not only explain the effect of the uncertainty around national elections but provide evidence for the broader measure of policy uncertainty, including events such as the Brexit. Since it is likely to have unobserved differences between the countries, this study follows Cao, Li and Liu (2017) which includes the lagged value of dependent variable to control for unobserved differences. Moreover, the model includes time fixed effects and the standard errors are clustered at country and year level.

The coefficient of policy uncertainty in the acquirer country is used to test the effect of policy uncertainty in the acquirer country on the outbound M&A activity of the acquirer country.

(18)

18

The risk management hypothesis predicts that the coefficient is positive which means that an increase in policy uncertainty leads to an increase in outbound M&A activity. In contrast, the real option hypothesis predicts that the coefficient is negative which means that an increase in policy uncertainty leads to a decrease in outbound M&A activity.

This study includes the cross-border ratio as the dependent variable. The cross-border ratio includes the volume of the domestic and cross-border M&A activity. This helps to control for unobserved indicators which influences the volume of both the domestic and cross-border M&A (Erel, Liao and Weisbach, 2012; Rossi and Volpin, 2004). The cross-border ratio is also commonly used in the literature to measure M&A activity, for example by Erel, Liao and Weisbach (2012) and Rossi and Volpin (2004). The cross-border ratio is calculated as the number of outbound M&A divided by the domestic number of M&A plus the outbound M&A of the acquirer. In line with Cao, Li and Liu (2017), this study includes the one year difference between the dependent variable and independent variable.

3.2 Inbound M&A

The next model is used to study the effect of policy uncertainty in the target country on the inbound M&A:

𝐿𝑛(𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑖𝑛𝑏𝑜𝑢𝑛𝑑 𝑀&𝐴 𝑝𝑒𝑟 𝑇𝑎𝑟𝑔𝑒𝑡 𝐶𝑜𝑢𝑛𝑡𝑟𝑦)_𝑗,𝑡

= 𝛼 + 𝛽₁𝑃𝑜𝑙𝑖𝑐𝑦 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑇𝑎𝑟𝑔𝑒𝑡 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝑗,𝑡 + 𝛾 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠_𝑗,𝑡 + 𝛿_𝑡+ 𝜀_𝑗,𝑡

where the dependent variable is natural logarithm of one plus the average number of inbound M&A. The independent variables are the natural logarithm of one plus the average Policy Uncertainty Index in the target country and the control variables of the target country. Based on the literature, this study controls for the level of shareholder protection, economic conditions, risk of investment, quality of institution and trade openness. The model also includes a control variable which controls for the one year before an election since Cao, Li and Liu (2017) provide evidence that the one year before an election influences inbound M&A. The model includes time fixed effects and the standard errors are clustered at country and year level.

(19)

19

The coefficient of policy uncertainty in the target country is used to test the effect of policy uncertainty in the target country on inbound M&A activity. The delay hypothesis predicts the coefficient of policy uncertainty in the target country is negative which means that an

increase in policy uncertainty decreases the inbound M&A activity.

This study will include the cross-border ratio of Erel, Liao and Weisbach (2012) and Rossi and Volpin (2004) as the dependent variable. The cross-border ratio is calculated as the number of inbound M&A divided by the number of domestic M&A plus the number of inbound M&A of the target country. Including the domestic number of M&A controls for unobserved indicators which influences the volume of both the domestic and cross-border M&A activity (Erel, Liao and Weisbach, 2012; Rossi and Volpin, 2004). In line with Cao, Li and Liu (2017), this study includes the one year difference between the dependent variable and

independent variable.

3.3 Country pair analysis

This study includes a country pair analysis which is in line with Cao, Li and Liu (2017), Erel, Liao and Weisbach (2012) and Rossi and Volpin (2004). The country pair analysis contributes to the previous models since the country pair analysis allows to study the difference between the acquirer country and the target country. For example, the country pair analysis allows the study to control for the difference between shareholder protection (Rossi and Volpin, 2004) and cultural differences (Ahern et al., 2015) which plays a role in cross-border M&A activity. The data is used to produce a matrix of (18 x 18) ordered country pairs. The M&A activity between the country pairs are shown in appendix Table 4. The following model is used:

𝐿𝑛(𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑟𝑜𝑠𝑠 − 𝑏𝑜𝑟𝑑𝑒𝑟 𝑀&𝐴 𝑝𝑒𝑟 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑝𝑎𝑖𝑟)_𝑗,𝑡 = 𝛼 + 𝛽₁𝑃𝑜𝑙𝑖𝑐𝑦 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝐴𝑐𝑞𝑢𝑖𝑟𝑒𝑟 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝑖,𝑡 + 𝛽₂𝑃𝑜𝑙𝑖𝑐𝑦 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑇𝑎𝑟𝑔𝑒𝑡 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝑗,𝑡 + 𝛾 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠_{𝑖−𝑗,𝑡} + 𝛿_𝑡+ 𝜀_𝑗,𝑡

where the dependent variable is the natural logarithm of one plus the number of cross-border M&A of a country pair. The independent variable is the natural logarithm of the

(20)

20

average Policy Uncertainty Index in the acquirer country, the average Policy Uncertainty Index in the target country and the control variables of the country pairs. The control variables are the difference between the level of shareholder protection, economic

conditions, risk of investments, quality of institution and trade openness of the acquirer and the target. Based on Erel, Liao and Weisbach (2012) this study controls for the valuation effect. Other control variables are regarding cultural differences. Furthermore, the regression includes time fixed effects and the standard errors are clustered at country pair and year level.

The coefficient of policy uncertainty in the acquirer country and the coefficient of policy uncertainty in the target country are used to test the effect of policy uncertainty on the cross-border M&A activity of the country pair. The risk management hypothesis predicts that the coefficient is positive which means that an increase in policy uncertainty leads to an increase in outbound M&A activity. In contrast, the real option hypothesis predicts that the coefficient is negative which means that an increase in policy uncertainty leads to a decrease in cross-border M&A activity. Furthermore, the deterrence hypothesis predicts that the coefficient is negative which means that an increase in policy uncertainty leads to a decrease in inbound M&A.

Again, the model will include the border ratio as a dependent variable. The cross-border ratio is calculated as the number of cross-cross-border M&A divided by the number of domestic M&A in the target country plus the number of cross-border M&A between the country pair (Erel, Liao and Weisbach, 2012; Rossi and Volpin, 2004). Also, the one year difference between the dependent and independent variable will be included (Cao, Li and Liu, 2017).

3.4 Potential problems with the empirical methods

This section discusses the potential problems with the empirical methods.

There is a potential problem of omitted variable bias. There could be unobserved factors which influences the cross-border M&A activity. To mitigate the omitted variable bias, this study includes various control variables which are based on the literature. Furthermore, this

(21)

21

study includes the lagged dependent variable. The lagged dependent variable is an easy way to account for omitted variable bias. This method may not be perfect, however, it can help to provide better results of the effect of policy uncertainty on cross-border M&A activity

(Wooldridge, 2013). In addition, this study includes an extra macroeconomic uncertainty control variable analysis to make sure that the results are driven by policy uncertainty instead of macroeconomic uncertainty. The extra macroeconomic uncertainty control variable is the Macroeconomic Uncertainty Index from Jurado, Ludvigson and Ng (2015). Furthermore, there is a potential problem of measurement error. A potential concern about the Policy Uncertainty Index from Baker, Bloom and Davis (2016) is that the index may measures the effect of other macroeconomic uncertainty. The instrument variable approach is used to further mitigate endogeneity concerns. The instrument is used to obtain an

exogenous measure of policy uncertainty. The instrument is the Partisan Conflict Index from Azzimonti (2018) which measures the disagreement about the policy. The Partisan Conflict Index is relevant since higher disagreement about the policy should be related to higher policy uncertainty. However, the Partisan Conflict Index is exogenous since the disagreement about the policy should not affect the M&A activity other than through an increase in policy uncertainty (Bonaime, Gulen and Ion, 2017). Another potential problem is sample selection bias. This study only includes announced deals. However, this may not be a random sample from the population of firms (Bonaime, Gulen and Ion, 2017).

(22)

22

4. Data and descriptive statistics

This paragraph describes the measure of policy uncertainty, the M&A data and the control variables.

4.1 Measure of policy uncertainty

This study used the Policy Uncertainty Index from Baker, Bloom and Davis (2016) to measure the policy-related economic uncertainty. This index is based on newspaper coverage

frequency and is available at their website1_.

There are two main measures of policy uncertainty in the literature, namely the policy uncertainty around national elections and the Policy Uncertainty Index from Baker, Bloom and Davis (2016). The policy uncertainty around national elections is a very broad measure since it measures uncertainty around changes in government policy but also measures the uncertainty around changes in the composition of the government (Julio and Yook, 2016). Furthermore, the national elections only capture the policy uncertainty around the national elections years and not the time in between these years. The time in between the national elections years could be important due to the high variation in M&A activity (Bonaime, Gulen and Ion, 2017). The figures in the appendix plot the Policy Uncertainty Index from the Baker, Bloom and Davis (2016) per country together with the national election years in that country. During the years in between the national election years there could be other important policy uncertainty, such as the Brexit referendum in the United Kingdom which could influence the decisions to engage in cross-border M&A. For this reason, this study used the Policy Uncertainty Index from Baker, Bloom and Davis (2016). The Policy Uncertainty Index captures the policy uncertainty not only in the national elections years but also the policy uncertainty in other years which could play an important role in the decisions to engage in cross-border M&A.

The Policy Uncertainty Index is available for Australia, Brazil, Canada, Chile, China, France, Germany, Hong Kong, India, Ireland, Italy, Japan, Mexico, Netherlands, Russia, Singapore,

(23)

23

Spain, Sweden, UK and USA. However, this study will exclude China and Hong Kong since there is no data on national elections. Therefore, this study will be limited to these countries.

The summary statistics of the policy uncertainty are given in Table 2. The average policy uncertainty value is 112. This average is somewhat larger than the average policy uncertainty value of 100 by Baker, Bloom and Davis (2016) and the value of 107 by Bonaime, Gulen and Ion (2017). The average policy-related uncertainty in all the countries between 1995 and 2016 is 112. This indicates that values below 112 are times with relatively low policy uncertainty, while values above 112 are times with relatively high policy uncertainty. For example, Sweden had an average policy uncertainty of 107 in 2016 which indicates that there is relatively low policy-related economic uncertainty. Before the Brexit, the United Kingdom had an average Policy Uncertainty Index of 204 in 2015. However, the United Kingdom had an average Policy Uncertainty Index of 542 in 2016 which is even higher. The uncertainty around the Brexit causes a relatively high policy-related economic uncertainty in the United Kingdom.

Table 2: Policy uncertainty summary statistics

Mean Standard deviation Median

Policy uncertainty 112 50 101

4.2 M&A data

To examine the cross-border M&A activity, this study creates a large sample of domestic and cross-border M&A using data from the Thomson One database. The sample will cover the period from 1995 to 2016. The sample is restricted to the countries which have a Policy Uncertainty Index and national election data available. Therefore, the sample only includes acquirers and targets from Australia, Brazil, Canada, Chile, France, Germany, India, Ireland, Italy, Japan, Mexico, Netherlands, Russia, Singapore, Spain, Sweden, United Kingdom and United States.

(24)

24

In line with Bonaime, Gulen and Ion (2017) and Nguyen and Phan (2017), the sample will only include M&A which are worth more than 1 million dollar2_{. The deal types minority stake} purchases, acquisition of remaining interest, privatizations, leveraged buyouts, spinoffs, recapitalizations, self-tender offers, exchange offers and repurchases are excluded from the sample (Erel, Liao and Weisbach, 2012). Moreover, the sample only includes M&A where the acquirer is acquiring 50% or more shares of the target, which makes the acquirer below 50% owner before the acquisition and above 50% owner after the acquisition. These restrictions are applied because this study only wants to capture important deals where the M&A affects the ownership of the firm and where the M&A is an important financial agreement.

The summary statistics of the M&A data are given in Table 3. The sample includes 207,890 deals with an average value of 263 million dollar. The distribution is skewed because the median is 20 million dollar. From these observations, 176,991 deals are domestic and 30,899 deals are cross-border. This means that 85% of the total deals are domestic M&A with an average value of 252 million dollar and a median of 19 million dollar, while 15% of the total deals are cross-border M&A with an average value of 322 million dollar and a median of 29 million dollar. This means that the average value of cross-border M&A is higher than the average value of domestic M&A.

Table 3: M&A summary statistics

Observations Mean deal value

(in $ million)

Median deal value (in $ million)

All deals 207,890 263 20

Domestic deals 176,991 252 19

Cross-border deals 30,899 322 29

Figure 1 plots the average number of outbound M&A together with the average Policy Uncertainty Index in the acquirer country. Figure 2 plots the average number of inbound

(25)

25

M&A and the average Policy Uncertainty Index in the target country. Both figures show the same pattern which indicates that the average number of outbound and inbound M&A peaked when the average Policy Uncertainty Indexes in the acquirer county and in the target country were low. These findings are in line with Bonaime, Gulen and Ion (2017) and Nguyen and Phan (2017) which found the same pattern for the effect of policy uncertainty on M&A activity. The inverse relationship between the Policy Uncertainty Index and cross-border M&A activity is in line with the real option hypothesis and the deterrence hypothesis which indicates that an increase in policy uncertainty decreases cross-border M&A activity.

Figure 1: average number of outbound M&A and average Policy Uncertainty Index in the acquirer country 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 1995 2000 2005 2010 2015 year

(26)

26

Figure 2: average number of inbound M&A and average Policy Uncertainty Index in the target country

4.3 Control variables

This study includes various control variables. The control variables are obtained from various databases. Appendix Table 1 includes a detailed description about the variables used in this study.

Erel, Liao and Weisbach (2012) and Rossi and Volpin (2004) provide evidence that economic conditions in a country influence M&A activity. It is important that the effect of policy uncertainty on cross-border M&A activity holds even after controlling for economic conditions. Furthermore, the control variables of economic conditions are included to mitigate endogeneity concerns due to omitted variable bias. It is important to control for economic conditions since there is evidence that economic conditions drive M&A activity (Erel, Liao and Weisbach, 2012; Rossi and Volpin, 2004). However, there is also evidence that policy uncertainty influences economic conditions (Bloom et al., 2014). Therefore, the Policy Uncertainty Index from Baker, Bloom and Davis (2016) could capture the effect of lower economic conditions (Gulen and Ion, 2016). The control variables are GDP per capita and GDP growth to control for the economic conditions within a country. To control for the trade

1 0 0 2 0 0 3 0 0 4 0 0 1995 2000 2005 2010 2015 year

(27)

27

openness within a country, the trade to GDP is included. The trade to GDP is both the import and the export in a country divided by the GDP. The GDP per capita, GDP growth and trade to GDP are available at the World Bank database3_.

Following the study by Cao, Li and Liu (2017), this study includes the control variables business environment and quality of intuition. The business environment measures the risk of investments in a country. The business environment is measured with the Investment Profile Index from the International Country Risk Guide (ICRG). This Investment Profile Index measures the risks to invest that are not influenced by other political, economic and financial risk components. The quality of institution is based on the study by Bekaert, Harvey and Lundblad (2005). It is measured by three components from the ICRG, namely corruption, law and order, and bureaucratic quality to create an index which measures the overall quality of the institution.

This study controls for policy uncertainty around national elections since Cao, Li and Liu (2017) provide evidence that policy uncertainty around national elections effects cross-border M&A activity. The reason to control for policy uncertainty around national elections is to make sure that the Policy Uncertainty Index not only measures the policy uncertainty around national elections but also the overall policy-related economic uncertainty.

Therefore, this study will contribute to the literature on the effect of policy uncertainty on M&A activity. The pre-election variable is a dummy variable which equals one if the year is one year before an election year. The data of election year is from the Database of Political Institutions (DPI)4_.

This study also controls for the level shareholder protection since Rossi and Volpin (2004) provide evidence that better shareholder protection results in more M&A activity. La Porta et al. (1998) show that common law countries have better shareholder protection than civil law countries. Therefore, the level of shareholder protection is measured with a dummy variable

3_{Available at: https://data.worldbank.org/}

(28)

28

equals one if the legal origin is common law and zero otherwise. The legal origin data is obtained from La Porta et al. (1998)5_.

In line with the study by Cao, Li and Liu (2017), this study includes the lagged dependent variable to mitigate endogeneity problems due to omitted variable bias. Although this study includes various amounts of control variables, it is likely that there are unobserved

differences between the countries. This study includes the lagged dependent variable to control for the unobserved differences. The lagged dependent variable is an easy way to account for omitted variables that causes differences in the dependent variable. Adding the lagged dependent variable to control for these omitted variables may not be perfect, however, it can help to provide better results of the effect of policy uncertainty on cross-border M&A activity (Wooldridge, 2013).

The country pair analysis enables to study the difference between the acquirer country and the target country. Therefore, the control variables are based on the difference between the acquirer country and the target country. The control variables are the differences between the level of shareholder protection, economic condition, risk of investment, quality of institution and trade openness. Other control variables are related to cultural differences since there is evidence that cultural differences play a role in cross-border M&A (Ahern et al., 2015; Lim, Makhija & Shenkar, 2016). The variables language, religion and continent are included to control for cultural differences. These variables are dummy variables which equals one if the acquirer country and target country share the same language, religion or continent and zero otherwise. The information is available at the World Factbook6_.

Furthermore, Erel, Liao and Weisbach (2012) provide evidence of the valuation effects. Therefore, the country pair analysis controls the difference between the exchange rate and the market-to-book ratio. The data on the national exchange rate and the value-weighted market-to-book ratio are obtained from DataStream.

As mentioned in section 3.4, this study includes an instrument variable analysis and an extra macroeconomic control variable analysis. The instrument variable is the Partisan Conflict

5_{Available at: http://faculty.tuck.dartmouth.edu/rafael-laporta/research-publications/} 6_{Available at: https://www.cia.gov/library/publications/the-world-factbook/}

(29)

29

Index from Azzimoti (2018) which is available at the Federal Reserve Bank of Philadelphia7_. The extra control variable is the Macroeconomic Uncertainty Index from Jurado, Ludvigson and Ng (2015) which is available at their website8_.

7_{Available at: https://www.philadelphiafed.org/research-and-data/real-time-center/partisan-conflict-index/} 8_{Available at: https://www.sydneyludvigson.com/data-and-appendixes/}

(30)

30

5. Results

5.1 Univariate analysis

Figure 3 shows the average number of outbound and inbound M&A in times of high policy uncertainty and low policy uncertainty. The figure shows that the average number of cross-border M&A differs in times with high policy uncertainty compared to low policy uncertainty. In contrast with the literature, the outbound and inbound M&A activity show the same pattern. The average number of outbound M&A is lower in times with relatively high policy uncertainty. These findings on outbound M&A are not in line with Cao, Li and Liu (2017). They have the same figure which shows that when the acquirer country is in the year before an election, which is their measure for high policy uncertainty, there is a higher average number of outbound M&A. The findings of Cao, Li and Liu (2017) are in line with the risk management hypothesis which predicts that acquirers will use outbound M&A to hedge against high policy uncertainty in their country. In contrast, the findings in figure 3 are in line with the real option hypothesis which predicts that acquirers will delay their investment due to high policy uncertainty. The findings on inbound M&A are in line with Cao, Lu and Liu (2017). The average number of inbound M&A is lower in times with relatively high policy uncertainty which is in line with the deterrence hypothesis which predicts that acquirers are deterred due to the high policy uncertainty in the target country.

Figure 3: the average number of outbound and inbound M&A in high and low policy uncertainty

High Low High Low 0 1 0 0 2 0 0 3 0 0

Average number of cross-border M&A

(31)

31

The two-sample t test is used to test if there is a significant difference between the average number of cross-border M&A in high and in low policy uncertainty. The average number of outbound M&A in high policy uncertainty is 261 and the average number of outbound M&A in low policy uncertainty is 290. The difference between the number of outbound M&A in high and low policy uncertainty is statistically significant. The average number of outbound M&A decreases by 29 deals, or 10% when the policy uncertainty in the acquirer country increase from low policy uncertainty to high policy uncertainty. The results on inbound M&A are the same. The average number of inbound M&A in high policy uncertainty is 227 and the average in low policy uncertainty is 256. The results show that the difference is statistically significant. The average number of inbound M&A decreases by 29 deals, or 12% when the policy uncertainty in the target country increase from low policy uncertainty to high policy uncertainty.

Table 4: difference between the number of cross-border M&A in high and in low policy uncertainty

The standard errors are reported in parentheses where *, ** and *** denote statistical significance at the 1%, 5% and 10% level, respectively.

High policy uncertainty Low policy uncertainty Difference Outbound M&A 261.044 290.001 -28.965*** (0.572) (0.588) (0.822) Inbound M&A 226.887 256.491 -29.603*** (0.493) (0.582) (0.766) 5.2 Multivariate analysis 5.2.1 Outbound M&A

The results of the effect of policy uncertainty in the acquirer country on outbound M&A are shown in Table 1. Columns 1 and 2 show the results on the number of outbound M&A and the cross-border ratio. Columns 3 and 4 show the results of the one year difference of the number of outbound M&A and the cross-border ratio.

(32)

32

The coefficient of policy uncertainty is negative and in three out of four regressions significant. In Column 1, the coefficient of policy uncertainty in the acquirer country is negative and significant at the 5% level. So, our results suggest a negative effect of policy uncertainty in the acquirer country on the average number of outbound M&A. This economic magnitude indicates that, ceteris paribus, a 1% increase in the Policy Uncertainty Index decrease the average number of outbound M&A with 11.5%. In Column 3, the negative effect is even larger. The economic magnitude indicates that, ceteris paribus, an annual difference of 1% of the policy uncertainty decrease the annual difference of the number of outbound M&A with 22.4%. Both findings indicate that the outbound M&A decreases significantly. Therefore, the effect of policy uncertainty on outbound M&A is statistically as well as economically significant. These findings are in line with the real option hypothesis. An increase in policy uncertainty will increase the value of the option to delay the investment (Bernanke, 1983), especially when the investment is hard to reverse (Gulen and Ion, 2016). These findings are in line with Bonaime, Gulen and Ion (2017), Chen et al. (2017) and Nguyen and Phan (2017) that found a negative relation between policy uncertainty and M&A activity. However, these findings are in contrast with the literature which mainly focused on cross-border M&A. The studies by Bonaime, Gulen and Ion (2017) and Cao, Li and Liu (2017) provide evidence for the risk management hypothesis which predicts an increase in outbound M&A when policy uncertainty is high.

The results on the other variables are in line with the literature. The firms in countries with better shareholder protection, better economic conditions, lower risk of investments and better quality of institution engage in more outbound M&A (Erel, Liao and Weisbach, 2012; Cao, Li and Liu, 2017; Rossi and Volpin, 2004). However, the results on the trade openness are much smaller than expected. Furthermore, the pre-election dummy is positive for three out of the four columns which is in line with Cao, Li and Liu (2017). Appendix Table 2 includes the results with and without the Policy Uncertainty Index. The results without the Policy Uncertainty Index show that the pre-election dummy is positive and not significant for all columns. This suggest that the policy uncertainty around national elections has a positive effect on outbound M&A which is in line with Cao, Li and Liu (2017). However, our results with the Policy Uncertainty Index provide evidence that the overall policy uncertainty has a negative effect on outbound M&A.

(33)

33

Table 5: the effect of policy uncertainty on outbound M&A

This table presents the results from the panel regression of the volume of outbound M&A by country and year. Columns 1 and 2 show the estimation of the average volume of outbound M&A, where the dependent variable is the natural logarithm of one plus the average number of outbound M&A and the cross-border ratio. Columns 3 and 4 show the estimation of the one year difference between the average number of outbound M&A and the cross-border ratio. All models include time-fixed effects and the standard errors are clustered at country and year level. The standard errors are reported in parentheses where *, ** and *** denote statistical significances at the 1%, 5% and 10% level, respectively.

(1) (2) (3) (4)

Ln(Number of outbound M&A)

Cross-border ratio ∆Ln(Number of outbound M&A) ∆Cross-border ratio Policy Uncertainty -0.115** -0.014 (0.054) (0.009) ∆Policy Uncertainty -0.224** -0.035* (0.106) (0.018) Legal 0.208** 0.012 0.023 0.010 (0.087) (0.008) (0.052) (0.008) GDP growth 0.008 -0.000 0.013 -0.000 (0.010) (0.001) (0.009) (0.002) GDP 0.100* 0.008 0.013 0.002 (0.053) (0.006) (0.042) (0.006) Business environment -0.040 -0.014** 0.010 -0.002 (0.047) (0.006) (0.041) (0.005) Quality of institution 0.028 0.007** 0.001 0.000 (0.019) (0.003) (0.016) (0.003) Trade to GDP -0.001** -0.000*** 0.000 0.000* (0.000) (0.000) (0.000) (0.000) Pre-election 0.000 0.009 -0.004 0.006 (0.048) (0.008) (0.049) (0.008) Lagged D.V. 0.856*** 0.845*** -0.187*** -0.348*** (0.037) (0.033) (0.050) (0.073) Constant 0.154 0.061 -0.223 -0.013 (0.355) (0.061) (0.240) (0.044)

Year-fixed effects YES YES YES YES

Observations 316 316 298 298

(34)

34

5.2.2. Inbound M&A

The results of the effect of policy uncertainty in the target country on inbound M&A are shown in Table 2. Columns 1 and 2 show the results on the number of inbound M&A and the cross-border ratio. In addition, Column 3 and 4 show the results of the one year difference of the number of inbound M&A and the cross-border ratio.

Columns 1 and 2 shows that the coefficient of policy uncertainty is positive and not significant. In contrast, Column 3 and 4 show that the coefficient of policy uncertainty is negative and not significant. This sign of the coefficient changes when the results are based on the one year difference between the Policy Uncertainty Index and the one year difference between the number of inbound M&A. These findings are not in line with Cao, Li and Liu (2017) which provide evidence for a negative relation between policy uncertainty and the number of inbound M&A as well as between the annual difference in policy uncertainty and the annual difference in number of inbound M&A. This change in the sign of the coefficient could be due to nonstationary data. Nonstationary data means that both the dependent variable and the independent variable have a trend over time (Stock and Watson, 2011). This could be the case since the cross-border M&A increased over time as well as the policy uncertainty. Taking the one year difference could mitigate the problem of nonstationary panel data (Stock and Watson, 2011) which shows a negative relationship between policy uncertainty and inbound M&A. The findings of the negative effect of policy uncertainty and inbound M&A are in line with Cao, Li and Liu (2017) which provides evidence of the

deterrence hypothesis. The deterrence hypothesis indicates that higher policy uncertainty in the target country deters foreign acquirers due to higher uncertainty about the value of the merged firm and the value of the synergies.

The results on the other variables are in line with the literature. The countries with better shareholder protection, lower risk of investment and better quality of institution attract more inbound M&A. There is no effect of the economic conditions in the target country on inbound M&A which is in line with Cao, Li and Liu (2017). Again, the findings on the trade openness is smaller than expected. The pre-election dummy is positive in Column 1 and 4 but negative in Columns 2 and 3. Appendix Table 3 includes the results with and without the Policy Uncertainty Index. The results without the Policy Uncertainty Index show that the

(35)

pre-35

election dummy is negative for two out of four columns. Based on the findings by Cao, Li and Liu (2017), this study expected the pre-election dummy to be negative in all columns.

However, our results with the Policy Uncertainty Index provide evidence that there is a negative effect of the overall policy uncertainty on inbound M&A.

(36)

36

Table 6: the effect of policy uncertainty on inbound M&A

This table presents the results from the panel regression of the volume of inbound M&A by country and year. Columns 1 and 2 show the estimation of the average volume of inbound M&A, where the dependent variable is the natural logarithm of one plus the average number of inbound M&A and the cross-border ratio. Columns 3 and 4 show the estimation of the one year difference between the average number of inbound M&A and the cross-border ratio. All models include time-fixed effects and the standard errors are clustered at country and year level. The standard errors are reported in parentheses where *, ** and *** denote statistical significances at the 1%, 5% and 10% level, respectively.

(1) (2) (3) (4)

Ln(Number of inbound M&A)

Cross-border ratio ∆Ln(Number of inbound M&A) ∆Cross-border ratio Policy Uncertainty 0.033 0.001 (0.048) (0.019) ∆Policy Uncertainty -0.056 -0.009 (0.088) (0.030) Legal 0.163** -0.002 -0.034 0.027 (0.064) (0.019) (0.044) (0.019) GDP growth 0.007 -0.004 0.012 -0.005* (0.008) (0.003) (0.007) (0.003) GDP 0.061 -0.009 -0.008 0.005 (0.039) (0.019) (0.032) (0.018) Business environment -0.058* -0.018 0.011 -0.011 (0.034) (0.015) (0.032) (0.013) Quality of institution 0.021 0.010** -0.002 0.006 (0.013) (0.005) (0.013) (0.005) Trade to GDP -0.001*** -0.000*** 0.000 0.000 (0.000) (0.000) (0.000) (0.000) Pre-election 0.002 -0.004 -0.015 0.006 (0.038) (0.012) (0.037) (0.011) Lagged D.V. 0.869*** 0.769*** -0.284*** -0.379*** (0.029) (0.063) (0.070) (0.131) Constant 0.123 0.221* -0.000 -0.014 (0.286) (0.117) (0.203) (0.100)

Observations 323 323 305 305

(37)

37

5.2.3 Country pair analysis

Table 7 shows the results of the effect of policy uncertainty in the acquirer and target country on the number of cross-border M&A between country pairs. The country pair

analysis contributes to the previous results since the country pair analysis allows to study the differences between the acquirer country and the target country.

Column 1 and Column 2 show that both the policy uncertainty in the acquirer country and target country have a positive but not significant effect on the volume of cross-border M&A between the country pairs. In contrast, Column 3 and Column 4 show that both the one year difference of the policy uncertainty in the acquirer country and in the target country have a negative effect on the one year difference of the volume of cross-border M&A between country pairs. These findings are the same as in Table 6. There could be problems of

nonstationary in the panel data. The one year difference mitigates this problem which shows a negative and significant effect.

In Column 4, the coefficient of the one year difference of the target policy uncertainty is negative and statistically significant at the 10% level. The economic magnitude indicates that, ceteris paribus, an annual increase of 1% in the policy uncertainty will decrease the annual difference of the cross-border ratio with 0.00015. The economic effect of policy uncertainty is small. Therefore, the effect of policy uncertainty on inbound M&A is statistically significant but not economically significant. The results of the negative relationship between policy uncertainty and inbound M&A are in line with Cao, Li and Liu (2017) which provides evidence of the deterrence hypothesis where an increase in the policy uncertainty in the target

country increase the risk about the value of the target and the value of the synergies.

In line with Rossi and Volpin (2004), the results show that better shareholder protection increases the cross-border M&A activity. The firms in countries with higher shareholder protection are more likely to acquire the firms in countries with lower shareholder

protection. In line with the literature, the firms from countries which have higher economic conditions are more likely to acquire firms form countries which have relatively lower economic conditions (Cao, Li and Liu, 2017; Erel, Liao and Weisbach, 2012; Rossi and Volpin, 2004). Furthermore, the results show that the coefficient of institutional quality is negative

(38)

38

and significant which is in line with Erel, Liao and Weisbach (2012). The coefficient of business environment is positive and significant which is in line with Cao, Li and Liu (2017). However, both the economic significance of the effect of better institution quality and better business environment on cross-border M&A is small. In contrast with Cao, Li and Liu (2017) and Erel, Liao and Weisbach (2012), this study provides no evidence that firms with higher level of trade openness are more likely to engage in cross-border M&A.

Consistent with the literature, these results show that firms in countries which have the same language and are in the same continent are more likely to engage in cross-border M&A. However, the firms in countries which have the same religion are less likely to engage in cross-border M&A (Cao, Li and Liu, 2017; Erel, Liao and Weisbach, 2012; Rossi and Volpin, 2004).

In contrast with Erel, Liao and Weisbach (2012), this study provides no evidence of a valuation effect. In Column 2, the coefficient of the difference between the average real exchange rate is negative and statistically significant. However, the economic significance is low. Moreover, the coefficients of the differences between the average market-to-book ratios are for three out of four regressions negative and not significant. Based on the literature, the expectation is that both the valuation coefficients are positive and significant which indicates that firms in countries with appreciated currency and relatively high market-to-book ratios are more likely to acquire firms in countries with depreciated currency and relatively low market-to-book ratios.

(39)

39

Table 7: the effect of policy uncertainty on cross-border M&A

This table presents the results from the panel regression of the volume of cross-border M&A by country pair and year. Columns 1 and 2 show the estimation of the average volume of inbound M&A, where the dependent variable is the natural logarithm of one plus the average number of cross-border M&A and the cross-border ratio. Columns 3 and 4 show the estimation of the one year difference between the average number of cross-border M&A and the cross-border ratio. All models include time-fixed effects and the standard errors are clustered at country pair and year level. The standard errors are reported in parentheses where *, ** and *** denote statistical significances at the 1%, 5% and 10% level, respectively.

(1) (2) (3) (4)

Ln(Number of cross-border M&A)

Cross-border ratio ∆Ln(Number of cross-border M&A) ∆Cross-border ratio Acquirer Policy Uncertainty 0.045 0.007 (0.031) (0.006)

Target Policy Uncertainty 0.015 0.001

(0.027) (0.005) ∆Acquirer Policy Uncertainty -0.011 -0.001 (0.039) (0.011) ∆Target Policy Uncertainty -0.025 -0.015* (0.038) (0.008) Legal 0.023 0.012*** 0.002 -0.006 (0.023) (0.004) (0.021) (0.004) GDP growth -0.003 0.000 -0.003 0.002*** (0.003) (0.000) (0.003) (0.001) GDP -0.002 0.006*** -0.006 -0.002 (0.014) (0.002) (0.013) (0.003) Business environment 0.007 0.004** 0.003 0.003 (0.013) (0.002) (0.012) (0.002) Quality of institution -0.001 -0.002*** 0.001 -0.002* (0.005) (0.001) (0.005) (0.001) Trade to GDP 0.000 -0.000*** 0.000 -0.000 (0.000) (0.000) (0.000) (0.000) Same language 0.195*** 0.020*** 0.010 0.005 (0.027) (0.005) (0.022) (0.005) Same religion -0.053*** -0.001 -0.000 0.001 (0.020) (0.004) (0.020) (0.004) Same continent 0.079*** 0.020*** -0.016 0.001 (0.021) (0.004) (0.020) (0.005) Currency 0.000 -0.000*** 0.000 0.000 (0.000) (0.000) (0.000) (0.000) Market-to-Book ratio -0.011 0.006 -0.001 -0.005 (0.017) (0.004) (0.016) (0.006) Pre-election -0.010 0.002 0.003 0.003 (0.015) (0.002) (0.014) (0.003) Lagged D.V. 0.824*** 0.414*** -0.452*** -0.274*** (0.010) (0.059) (0.018) (0.093) Constant -0.014 -0.027 0.014 -0.005** (0.190) (0.034) (0.013) (0.002)

Observations 2,893 2,894 2,624 2,771

Policy uncertainty effect on cross-border mergers and acquisitions