Cross-border M&A ownership strategies of SWFs and SOEs and the role of administrative distance and nationalist feelings of a nation

(1)

Cross-border M&A ownership strategies of

SWFs and SOEs and the role of

administrative distance and nationalist

feelings of a nation.

Student: Robert van Leeuwen (10009019)

Thesis supervisor: Vittoria Scalera

MSc Business Administration, International Management

University of Amsterdam

26-1-2017

(2)

1

STATEMENT OF ORIGINALITY

This document is written by Student Robert van Leeuwen who declares to take full

responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is

(3)

2

ABSTRACT

This research focuses on the relation between two different types of government controlled organizations, respectively state-owned enterprises (SOEs) and sovereign wealth funds (SWFs), and their cross-border ownership strategies. The nationalistic feelings of the target’s country and the administrative distance (AD) between acquiror and target country are

hypothesized and found to influence this relation. Data on acquisitions was extracted from the Zephyr database and merged with data from Murtina and Scalera’s (2016) research. The nationalistic feelings construct was subdivided in two different factors, nationalism and constructive patriotism. The ISSP National Social Survey Programme (Gesis, 2013) provided data on these factors. In addition the CAGE Comparator (Pankaj Ghemawat, 2016) database was used to extract data on AD. By merging all the different datasets, statistical analysis were conducted and resulted in the presentation of three different models. The first model is a world-wide applicable model that present a curvilinear U-shaped relation between AD and share of acquisitions where SOEs take a bigger share than SWFs, with the exception of the low-medium administrative distance. Furthermore there is also an interaction between the level of nationalism in the target country and type of organization on the share of acquisition, which is only significant for target countries where the level of nationalism is relatively high. In these cases, SOEs take a bigger share than SWFs. The second model is a world-wide applicable model with the exception of Norwegian organizations. Norwegian organizations were proofed to take significantly smaller shares in acquisitions compared to organizations from other countries. The second model presents an direct effect of AD and type of

organization on the share of acquisition. In this model there is a curvilinear U-shaped relation between AD and the share of acquisition where SOEs take a greater share compared to SWFs when making acquisitions. The last model presented is an model solely applicable on

Norwegian SOEs. This model also presents a curvilinear U-shaped relation between AD and the share of acquisition.

Keywords: SOE; SWF; cross-border ownership strategies; share of acquisition;

(4)

3

LIST OF TABLES AND FIGURES

TABLE 1. CHARACTERISTICS SOES AND SWFS 14

FIGURE 1. CONCEPTUAL MODEL 20

TABLE 4. DEFINITIONS OF VARIABLES 23

TABLE 5. ANOVA OUTPUT ORGANIZATION AND SHARE OF ACQUISITION 26

TABLE 6. DESCRIPTIVE STATISTICS OF SHARE OF ACQUISITION PER ORGANIZATION 27

FIGURE 2. SCREE PLOT FACTOR ANALYSIS 28 TABLE 7. FACTOR ANALYSIS CONSTRUCTIVE PATRIOTISM AND NATIONALISM 28

TABLE 8. DESCRIPTIVE STATISTICS AND CORRELATION MATRIX 29

TABLE 9. FACTORIAL ANOVA OF AD, ORGANIZATION AND SHARE OF ACQUISITION 31 TABLE 10. DESCRIPTIVE STATISTICS FACTORIAL ANOVA OF AD, ORGANIZATION AND SHARE OF

ACQUISITION 31

FIGURE 3. CHART OF INTERACTION BETWEEN AD AND TYPE OF ORGANIZATION 32 TABLE 11. HAYES MODERATOR SPSS MACRO OUTPUT OF NATIONALISM, ORGANIZATION AND

SHARE OF ACQUISITION. 33

FIGURE 4. MODERATING EFFECT NATIONALISM 34 TABLE 12. JOHNSON-NEYMAN OUTPUT NATIONALISM, CONDITIONAL EFFECT OF ORGANIZATION

ON SHARE OF ACQUISITION AT VALUES OF THE LEVEL OF NATIONALISM 34 TABLE 12. HAYES MODERATOR SPSS MACRO OUTPUT OF NATIONALISM, ORGANIZATION AND

SHARE OF ACQUISITION AND CONTROL VARIABLES. 35 TABLE 13. HAYES MODERATOR SPSS MACRO OUTPUT OF CONSTRUCTIVE PATRIOTISM,

ORGANIZATION AND SHARE OF ACQUISITION. 36

FIGURE 4. MODEL WORLDWIDE 37

TABLE 14. ACQUIROR COUNTRIES AND SHARE OF ACQUISITION 37

TABLE 15 ONE-WAY ANOVA OF THE RELATION BETWEEN ORGANIZATION AND SHARE OF

ACQUISITION WITH THE EXCLUSION OF NORWEGIAN ORGANIZATION. 39

FIGURE 5. AD DIRECT INFLUENCE WHEN EXCLUDING NORWEGIAN ORGANIZATIONS. 39 TABLE 16. FACTORIAL ANOVA OF AD, ORGANIZATION AND SHARE OF ACQUISITION EXCLUDING

NORWEGIAN ORGANIZATIONS 39

TABLE 17. DESCRIPTIVE STATISTICS FACTORIAL ANOVA OF AD, ORGANIZATION AND SHARE OF

ACQUISITION EXCLUDING NORWEGIAN ORGANIZATIONS. 40 TABLE 18. HAYES MODERATOR SPSS MACRO OUTPUT OF NATIONALISM, ORGANIZATION AND

SHARE OF ACQUISITION WITH THE EXCLUSION OF NORWEGIAN ORGANIZATIONS. 41 FIGURE 6. MODERATION TEST OF NATIONALISM ON THE RELATION BETWEEN ORGANIZATION AND

SHARE OF ACQUISITION 41

TABLE 19. HAYES MODERATOR SPSS MACRO OUTPUT OF NATIONALISM, ORGANIZATION AND

SHARE OF ACQUISITION AND CONTROL VARIABLES WITH THE EXCLUSION OF NORWEGIAN

(6)

5

TABLE 20. HAYES MODERATOR SPSS MACRO OUTPUT OF CONSTRUCTIVE PATRIOTISM,

ORGANIZATION AND SHARE OF ACQUISITION AND CONTROL VARIABLES WITH THE

EXCLUSION OF NORWEGIAN ORGANIZATIONS. 43 TABLE 21. REGRESSION OUTPUT OF BEING A NORWEGIAN SOE OR NOT AND THE RELATION WITH

SHARE OF ACQUISITION 44

FIGURE 7. SHARE OF ACQUISITION BY NORWEGIAN SOES 44 TABLE 22. ONE-WAY ANOVA OF THE RELATION BETWEEN AD AND SHARE OF ACQUISITION FOR

NORWEGIAN SOES 45

TABLE 23 DESCRIPTIVE STATISTICS ONE-WAY ANOVA BETWEEN AD AND SHARE OF ACQUISITION

FOR NORWEGIAN SOES 45

FIGURE 8. SHARE OF ACQUISITION BY NORWEGIAN SOES PER AD 46 TABLE 24. REGRESSION ANALYSIS NORWEGIAN SOES AND THE RELATION BETWEEN NATIONALISM

AND CONSTRUCTIVE PATRIOTISM AND THE SHARE OF ACQUISITION 46

TABLE 25. ANOVA OF REGRESSION ANALYSIS 47

FIGURE 9. MODEL FOR NORWEGIAN SOES AND THEIR SHARE OF ACQUISITION 47

TABLE 26. SECTORS AND SHARE OF ACQUISITION 48

FIGURE 10. TARGET SECTORS’ AND TYPE OF ORGANIZATION’S INTERACTION ON SHARE OF

ACQUISITION 49

TABLE 27. YEAR AND SHARE OF ACQUISITION 50

TABLE 28. AMOUNT OF ACQUISITIONS PER YEAR 50

FIGURE 11. YEAR, ORGANIZATION AND SHARE OF ACQUISITION 51 TABLE 29. INTERACTION BETWEEN ACQUIROR AND TARGET TRADE BLOCS AND TYPE OF

ORGANIZATION 52

FIGURE 12. ACQUIROR TRADE BLOCS ORGANIZED PER ORGANIZATION AND THEIR AVERAGE

SHARE OF ACQUISITION. 53

TABLE 30. SHARE OF ACQUISITION PER TARGET TRADE BLOC COMPARED TO THE EU 54 TABLE 2. FREQUENCY OF AND PRESENTATION OF COUNTRIES INCLUDED IN THE ISSP MODULE 72 TABLE 3. DESCRIPTIVE STATISTICS OF THE TARGET COUNTRIES’ LEVEL OF NATIONALISM AND

(7)

6

1. INTRODUCTION

Recently experts have stated their concerns about state-owned organizations’ acquisitions, as an example the article in the Financial times, written by Kynge (2016), about Chinese groups’ acquisitions in Europe can be taken. The Financial times dedicated an article to the Chinese state-owned enterprises (SOEs) where they let experts express their opinion on these enterprises and their acquisition strategy. Self-made billionaire Wang Jianlin, chairman of Dalian Wanda group, states that these Chinese companies invested a record amount in Europe but concerns grows due to their management cycles. Their long cycles for getting approval makes these companies inflexible according to him. In addition these SOEs are urged by their government to undertake foreign direct investments (FDI) in ‘good project which make money’ due to the disappointing results of these acquisitions. These first mentioned are not the only concerns expressed about Chinese SOEs operating in Europe. In the same article, Haremann & Hudori, experts in that area, state that the asymmetry of bilateral market access is another European concern. “Chinese interest is particularly growing in sectors that are restricted to foreign investors in China, amplifying the political salience of unequal market access”. Besides this article in the Financial Times, scientific articles have also expressed their concerns about the way SOEs differ from privately-owned enterprises (POEs) in their foreign strategies (Rudy, Miller & Wang, 2016; Bruton, Ahlstrom, Stan, Peng & Xu, 2015, Muchassio & Lazzarini, 2014; Inoue, Lazzarini & Musacchio, 2013).

In addition to the SOEs, Bremmer (2010) argues that next to these enterprises governments possess different intermediary institutions where through they are able to

undertake FDI. Sovereign Wealth Funds (SWFs) are among those discussed. Others discussed are privately owned champions. This article will only focus on the SOEs and SWFs because, in line with Rudy et al. (2016), the role of FDI for ‘national champions’ must be better understood but needs to be the core focus of an article. SWFs are state owned funds without explicit pension liabilities and typically pursue long-term investment strategies. These are argued to have a faster growth rate than any other institutional investor (Aizenman & Glick, 2008). Both these state-owned investors are by definition under control of their domestic government and may have more goals than just profit maximization (Bremmer, 2010). Due to their relation to the government, the political setting is argued to play a role in strategic decisions for SOEs (Shi, Hoskisson & Zhang, 2016; Bruton et al, 2015; Ahlstrom & Bruton, 2010) as state ownership dictates that SOEs represent nations interests and strategic

(8)

7 also argued to play a role in strategic decisions for SWFs as well (Bortolotti, Fotak, &

Megginson, 2015; Bernstein, Lerner & Schoar, 2013; Dewenter, Han & Malatesta 2013; Johan, Knill & Mauck, 2013; Megginson, You & Han, 2013; Knill, Lee & Mauck, 2012).

Not only the proposed article in the Financial Times (Kynge, 2016) is giving

additional attention to Chinese SOEs, even the scientific article focus foremostly on Chinese SOEs (Bruton et al., 2015). And not do they only focus on the Chinese SOEs, within the variations, they tend to focus on the Wholly Owned SOEs (WOSOEs), or at least they see the SOEs from a ‘black-and-white’ perspective where an enterprise is either state-owned or private-owned. They fail to address the hybrid structures and therefor, the implications for cross-border M&A strategy as well. Nevertheless, Musacchio, Lazzarini and Aguilera (2015), address some of these hybrid structures while they question how they relate to other state institutions. Hybrid structures can be enterprises majority/majority owned by the state, sometimes related to as ‘new varieties of state capitalism’. Complementary, Rudy et al. (2016) argue that the various hybrid structures are an interesting opportunity for future research as well. Next to the literature on SOE, literature on SWFs addresses some gaps as well. As some argue that the literature on private investors may not be used on SWFs (Aguilera, Capapé & Esade, 2016; Murtina & Scalera, 2016; Johan, Knill & Mauck, 2013; Dixon & Monk, 2012), they all address the importance of the SWFs for the global economy since they possess huge amounts of capital. This recent trend of attention for SWFs presents a lot of opportunities for research. As both literature on SWFs (e.g. Aguilera et al., 2016) and literature on SOEs (e.g. Rudy et al. 2016) addresses the ‘political missions’ they might have, this article would like to follow Musaccio et al.’s (2015) lead, to address the relation between SOEs and SWFs and focus on their cross-border M&A ownership strategy. The reason the focus is on M&A ownership strategy is because it is raising recent concerns, as discussed earlier (Kynge, 2016). In addition, this article wants to include the nationalist feelings of the population in the host-country. The salience of nationalist politics (SNP) are the ‘tactics and maneuverings used by domestic political leaders to arouse or appease nationalistic feelings’ (Shi, Hoskisson and Zhang, 2016: 22), this article would like to focus on the nationalistic feelings of the host-country population because this is the extent of the SNP. Shi et al.’s article on SOEs, proposes that the SNP is a determinant of the relation between nations states and influences the opposition faced in a target state. Previous literature also states that the political setting plays a role in strategic decisions (Bueno de Mesquita, 2002). This article

(9)

8 would like to test how the nationalist feelings influences the cross-border M&A ownership strategies of both SOEs as SWFs.

In addition, the literature states that international investment can and should be

thought of as two separate decisions, where and how much (Shi et al., 2016; Knill et al., 2012; Biglaiser & DeRouen, 2007; Reed, 2000). Location should not just be considered as a

economic but also as an political concept. Ghemawat provides an CAGE model wherein he describes how distance may impact the location choice (Ghemawat, 2001). The CAGE model presents the importance of Cultural, Administrative, Geographical and Economic distance between countries and industries. Since the topic of this article is related to nationalistic feelings, the main focus is going to be on the administrative distance (AD) since it relates the most to this topic.

The difference between SOEs and SWFs and their ownership strategies will be researched using different datasets on acquisitions by SOEs and SWFs gathered from Zephyr and data on SWFs from Murtina and Scalera (2016). The nationalist feelings variable will be based on information from the ISSP National Identity Module (Gesis, 2016). The ISSP National Identity Module is a cross-national survey that deals with issues regarding respondents’ national identification, national pride, support for their own nation, attitudes towards national and international issues, attitudes towards foreigners and foreign cultures, etc. In addition, the data on AD will be gathered from Ghemawat’s online CAGE comparator (Pankaj Ghemawat, 2016). The data on attitudes and nationalistic feelings of a population will be combined with the data on acquisitions by SOEs and SWFs and data on AD to create a complete dataset which will address the gap in the literature.

By using the method described in the previous paragraph, this article is going to provide a framework of how SWFs and SOEs differ in their cross-border M&A ownership strategy and how the nationalist feelings of a population influences this relation. As described in the article of the Financial Times (Kynge, 2016), state-owned investors investing in foreign countries raise concern by the home governments. Not only is the performance and

governance of these state-owned investors not efficient, the political setting of the home- and host- country is described as being of influence on the potential investment strategy as well. This raises the question for home governments how to deal with these foreign investors and for foreign investors how to invest in a foreign country. Some states, like China even close their border for foreign investors in some industries. ‘The state perceives more and more sectors to be of strategic importance and deters foreign companies from entering them.

(10)

9 Indeed, the rise of state capitalism in some of the world’s most important emerging markets has shifted the tectonic plate. Globalization now comes with new costs and risks’ (Bremmer, 2014: 1). This article will address how different state-owned investors differ in their M&A ownership strategy, so home- and host-country governments will know how to bring a merger or acquisition to a good end. The political setting plays an important role in the explanation of this relation and therefor the nationalistic feelings of the population is included. Not only does this article gives recommendations for managerial implications, it also contributes to the literature on different areas. This article combines international business, finance, strategy and politics, whereby it is relevant to the present-day business context since they are heavily investing.

The rest of the paper is organized as follows. Section 2 develops a theoretical

framework, section 3 describes the hypotheses development, section 4 the data, methodology and descriptive statistics, section 5 reports the results, section 6 is the discussion and the last section, section 7 concludes and describes directions for future research.

2. THEORETICAL FRAMEWORK

2.1 Trends

We are entering a new area of ‘state capitalism’ where governments share ownership with private owners and/or provide strategic support trough credit and/or state protection

(Musacchio & Lazzarini, 2014; Bruton, Ahlstrom, Stan, Peng & Xu, 2015). These different ownership configurations have implications on the performance on the international market (Cuervo & Villalonga, 2000; Ramamurti, 2000). Not only are these SOEs changing the international market, SWFs are gaining ground as well. According to the Sovereign Wealth Fund Institute, the market size of SWF investments has doubled from September 2007 to September 2014, with a growth rate faster than any other institutional investor (Aizenman & Glick, 2008). Both these institutions are under government control, which influences their cross-border M&As ownership strategies. The following part will address the rise of state capitalism.

2.2 State capitalism

Recent transformation of government involvement in the market has been argued to derive from the crisis. Bremmer (2010) argues the crisis set foot to ‘state capitalism’, a transfer of market power from capitals of finance to capitals of political power. In his article he makes two different distinctions between the free market and state capitalism. Firstly, the state sees

(11)

10 state capitalism as a long-term policy choice instead of a temporary state of policy to rebuild economy. Secondly, state capitalists see the market solely as a tool that serves the state and use it to extend their political grasp both within society as on the international stage. Bremmer argues that the rise of state capitalism is problematic for capitalism efficiency because

governments only maximize economic performance within their country and promote their own political goals and political dominance. These same governments have considerable control over intermediary institutions to exert this control over the market to and use it for their own goals. The most important of these are national oil (and gas) corporations (NOCs), other state-owned enterprises (SOEs), privately owned national champions (POEs backed by the state), and sovereign wealth funds (SWFs). In the following parts the focus will be on SOEs and SWFs, their performance and M&A strategies.

Another view on the rise of state capitalism is that it didn’t derive from the recent crisis but that it co-emerged with the liberalization and privatization reform period that lifted of in the 80’s (Mussacio, Lazzarini & Aguilera, 2015). Governments around the world have transformed the state capitalism model, under which they controlled wholly-owned SOEs (WOSOEs) (Ahroni, 1986; Ramamurti & Vernon, 1991; Trebat, 1983), into new models in which the government cooperates with domestic and foreign private investors to develop new strategic capabilities using new governance arrangements, thereby creating hybrid

organizations (Musacchio et al., 2015). WOSOEs lack profit-only investors and transparency and are therefore damaged in their performance.’ Such negative effects occur through various channels often referred to a as ‘liability of stateness’: poor selection of managers,

low-powered incentives, and weak monitoring (managerial agency); conflicting goals clashing with profitability (social view); and use of WOSOEs for political gain and bailout (political view)’ (Mussachio et al., 2015: 120). The formation of these hybrid organizations are a response to the ‘liability of stateness’ of the WOSOEs while at the same time preserving the benefits of having the state as a stakeholder. In these hybrid organizations, states own either majority or minority stakes in POEs, which makes them SOEs, whereas each of these forms have different implications for performance (Bruton et al., 2015; Mussachio et al., 2015) and the difference in level of state control can influence SOEs strategic choices (Ioue, Lazzarini, & Mussachio, 2013, Musacchio & Lazzarini, 2012, 2014).

(12)

11

2.3 SOEs

By definition, hybrid organizations are subject to various levels of state control, where the pursuit of national and political interests and may supersede the goal of financial profit maximization (Contractor, 2013; Knutsen, Rygh, and Hveem, 2011; Zif, 1983). SOEs pursue multiple goals, based on both the state’s social, national and political objectives as the profit objectives of the organization itself. This leads to the suggestion that SOEs behave differently from POEs, not only on domestic but also foreign terrain. The literature has addressed this feature. Most research has been done on the inefficiencies that create a performance gap between SOEs relative to POEs (Dharwadkar, George, & Brandes, 2000; Doh, Teegen, & Mudambi, 2004; Megginson & Netter, 2001; Ramamurti & Vernon, 1991; Rudy et al, 2016; Shi, Hoskisson & Zhang, 2016; Spencer, Murtha, & Lenway, 2005; Trebat, 1983; Yiu, Bruton, & Lu, 2005), where these inefficiencies often relate to the “liability of stateness.” (Mussachio et al., 2015). In the little literature there is, only a few have focused on the

determinants of this performance gap, either by looking at different ownership configurations (Cuervo & Villalonga, 2000; Ramamurti, 2000) or by outlining institutional conditions that may lead to smaller performance gaps (Bartel & Harrison, 2005; Mahmood & Rufin, 2005; Vickers & Yarrow, 1988). In sum, the dominant view is that SOEs underperform POEs under almost any circumstance (Mussachio et al, 2015). Not only do their performance differ from POEs, their M&A strategies differ as well. This will be discussed in the next section.

As a result of their soft budget constraints and tolerance for long-term investment, the risk appetite for SOEs is likely to be greater than that of POE (Rudy et al., 2016). The

negative side to this soft budget constraint is that it may lead to overinvesting and dedicating to the wrong technology (Bruton et al., 2015). These same soft budget constraints and long-term investment motives do set basis for asset-replicating FDI strategy (Rudy et al., 2016). Asset replicating FDI is when a SOE’s acquisition of a foreign firm motive is to replicate its strategic assets to benefit the state as a whole. Asset-replicating FDI and its FSA flows are less likely to occur with POEs. Another strategy differentiating SOEs from POEs is the supply-controlling FDI strategy, ‘which involves the state’s objective to control the production and distribution of critical inputs for the SOE or an entire industry in the home country’ (Rudy et al, 2016: 74). In fulfilling the social, political or national objectives for the state, the SOE seeks to reduce uncertainty and cross-border dependency by procuring a constant supply of the input, while gaining non location-bound FSAs from the acquired foreign firm for exploitation in the home country. When states are in transition, developing capabilities are even argued to be one of the most important SEO strategies (Peng, 2001).

(13)

12 While Rudy et al. (2016) is not the only one in the literature to exclusively focus on the

WOSOEs, nevertheless they are congruent with the view that SOEs pursue more goals than just maximizing profit (Contractor, 2013; Knutsen, Rygh, and Hveem, 2011; Zif, 1983). Literature on SOEs from emerging markets argue that cross-border M&A can be used as a springboard (Luo & Tung, 2007). It is argued they do this to acquire strategic resources, reduce market constraints at home, reduce institutional constraints at home and by using aggressive risk-taking acquisitions they try to overcome latecomer disadvantage. But beyond SEOs, governments have other institutions under their wing, SWFs among them. In the next section SWFs and their nature will be discussed.

2.4 SWFs

Next to the emergence of the earlier described hybrid organizations, another important state-owned institution gained attention in both literature as practice. These institutions, SWFs, are a class of funds ‘blurring the lines between finance and politics’ (Aguilera, Capapé & Esade, 2016:5). Sovereign wealth funds (SWFs) are funds owned and/or controlled by the state and invest assets in both domestically as foreign markets(Johan, Knill & Mauck, 2013), they don’t have explicit pension liabilities, mostly pursue long-term investment strategies, tend to be internationally focused and manage multibillion-dollar assets. This definition is excluding SOEs from SWFs. In sum, SWFs are unique investors due to their size, central involvement in global finance, systemic power, geopolitical influence and growth rate. Examples of SWFs are Singapore’s Temasek, China Investment Corporation (CIC), Norway’s Pension Fund Global(NPFG) and United Arabic Emirates’s Abu Dhabi Investment Authority (ADIA). The given examples demonstrates the heterogeneity of the SWFs. They differ in size, home-country, source of funding but all posses an extraordinary amount of financial capital to invest. Although these funds are state-owned, they do not differ from other institutional investors in their capacity to invest in POEs. Nevertheless, their motives to invest are likely to be different (Aguilera et al, 2016) and a ‘one size fits all’ approach to evaluating SWFs may not even be appropriate (Knill, Lee & Mauck, 2012). These state owners adopt a unique type of governance. ‘They are equity owners that cannot exercise sovereign regulatory or

supervisory powers in the organizations in which they invest. SWFs are simply one investor among many, and they have a fiduciary duty to the state (or, ultimately, the citizens of a given country). Some researches argue SWFS are still under pressure to achieve the same financial efficiency as other institutional investors in the race to become global players’ (Aguilera et

(14)

13 al., 2016: 7). Their challenge is to balance both state’s as organization’s objectives (Shleifer & Vishny, 1994; Megginson & Netter, 2001; Dixon & Monk, 2010; Johan et al., 2013, Aguilera et al., 2016), while others argue they are already acting as regulator investors (Kotter & Lel, 2010; Megginson, You & Han, 2013). In the following section the performance and

internationalization strategies of SWFs will be discussed.

In SWFs, there has to be balance between financial efficiency and political

effectiveness. The active presence of politicians in SWFs might lead to investment behaviors that do not maximize shareholder value because of political and strategic objectives (Murtina & Scalera, 2016; Chhaochharia & Laeven, 2008; Bernstein, Lerner & Schoar, 2010). In addition, other researchers state there are no political objectives behind the SWFs (Balding, 2008; Karolyi & Liaou, 2010). Nevertheless Kotter and Lel (2011) find that SWFs appear to be passive shareholders, as target firms’ performance and governance do not change

significantly in the long run relative to a control sample. The magnitude of the market reaction is similar to the announcement effects of investments by institutional investors on stock returns for a comparable event window (e.g., Brav, Jiang, Partnoy, and Thomas, 2008; Wahal,1996), indicating that SWF investments a positive signal to market participants about the future risk adjusted returns of target firms. The reasons underlying the positive market reaction, they find that firms with a high likelihood of future distress and low cash reserves experience higher excess returns, suggesting that SWF investments are particularly valuable for firms facing financial difficulties and serve as a certification of the firm’s long-term economic viability. Their results also show that transparent SWFs are more likely to prefer firms facing financial difficulties than opaque SWFs and the short-run market reaction to announcements of SWF investment is positive and is an increasing function of the degree of SWF transparency and target firms’ financial difficulties. Results show that market

participants react favorably SWF investments, suggesting that they view transparent SWFs as profit-oriented investors. These findings are consistent with Megginson et al. (2013). Their results support the investment facilitation hypotheses that indicate that SWFs act principally as commercial investors in their cross-border transactions concerning stock acquisitions. In contrast, others argue that SWFs are not perceived as passive stakeholders by host-countries but have potential political objectives (Bortolotti, Fotak & Megginson, 2015; Bernstein et al, 2013; Dewenter et al., 2013) and where opaqueness leads to a higher perceived risk in the target country (Gieve, 2008, Johnson, 2007, Summers, 2007). Therefor SWFs have different investment methods compared to POEs, for example trough investment vehicles to minimize target country’s risk perception and when acquiring majority stakes (Murtina & Scalera, in

(15)

14 press). The decision making by SWFs with regard to investing in private equity differs from other institutional investors as well (Johan, et al. 2013). As demonstrated there is no

consensus on whether SWFs are different in their internationalization strategies as compared to other institutions. But there can be noted that the political objectives play a role when assessing potential treats. An overview of what has been discussed to so far is presented in table 1. In the next section the importance of nationalist feelings is going to be addressed in both SWFs as POEs.

Table 1. Characteristics SOEs and SWFs

SOEs SWFs

 Various level of state control (hybrid structures)

 Full state control  Domestic/internationally focused  Internationally focused  Soft budget constraints and tolerance

for long-term investment

 Extraordinary amount of financial capital

 Multiple goals (profit & state)  Multiple goals (profit & state)  Competitor in foreign markets  Investor in foreign markets  Liability of stateness  Liability of stateness

*Asset-replicating FDI *Active presence of politicians

*Supply-controlling FDI *Investment vehicles to minimize target country's risk perception

2.5 Cross-border M&A ownership strategies

As stated in the previous section host-countries may perceive a SWF hostile when they opaque because they can have hidden political objectives. The same concern is expressed for SOEs (Mussachio et al., 2015). Wherefore the WOSEOs, the hybrid organizations are given as a solution for their opaqueness, perceived potential political hazard and bailout (political view) (Mussachio et al., 2015), transparency and passivity are given as a solution for the SWFs to handle hostility. But not only the construction of the venture is of importance when dealing with risks, also the share of ownership when (partially) acquiring a foreign venture is of great importance. The share of ownership has implications for among others risk, return and control (Anderson & Gatignon 1986; Luo, 2001).

Entry modes have been classified along a continuum ranging from non-ownership modes or non-equity modes like contractual agreements and export to complete ownership or equity modes like wholly owned subsidiaries (Pan & Tse 2000; Gatignon & Andersen, 1988). The factors that influence the choice are based on different country and industry based factors that relate to uncertainties and risks (Pan & Tse, 2000). The rule of thumb for these entry

(16)

15 modes is the higher the risk and uncertainty in a host country, the lower the commitment the venture is going to make. A low commitment is demonstrated as non-ownership entry modes like export and contractual agreements. Among the entry modes are also JVs. These are a preferred when one wants to partner up with a local organization to spread the risk and uncertainty it is going to face in a host country. This risk may come from different sources, for example host country restrictions and information asymmetries.

Acquisitions under the 100% have been categorized in different research articles as a JV. This should not always be the case (Chen & Hennart, 2004). Where traditional JVs are a greenfield entry mode since they enter the market slowly and are ventures on its own, entry through a partial acquisition is not a greenfield entry mode and is a much quicker move (Brouthers & Hennart, 2007). This study will distinguish entry trough partial acquisitions from entry trough JVs since these are different.

Information asymmetry sets the base for share of ownership (Malhotra & Gaur, 2014). As the distance increases, so does the information asymmetry. As a result ventures have preferred shared ownership to minimize the potential hazards the asymmetry can inflict. The primary source demonstrated of exogenous uncertainty, and there for a lower share of ownership, is country risk (Chari & Chang, 2009). Country risk departs specifically from economic, financial and political risk (Barkema & Vermeulen 1998, Gatignon & Andersen, 1988, Pan, 1996). Share of ownership is also demonstrated to be influenced by the sectors the organization is originally from and acquiring in. When the two differ, the share of ownership is lowered (Malhotra & Gaur, 2014). This may be a result from information asymmetry.

2.5 Distance

The literature states that international investment can and should be thought of as two separate decisions. Firstly the where has to be considered and secondly, how much (Shi et al., 2016; Knill et al., 2012; Biglaiser & DeRouen, 2007; Reed, 2000). They suggest location should not just be considered as an economic but also as an political concept. Gemawhat’s (2001)

dimensions of the distance concept describe how distance may impact location choice for enterprises. The CAGE model expresses the importance of a distinction between Cultural, Administrative, Geographical and Economic distance between countries and industries.

The cultural dimension relates to the cultural distance which consists of differences in languages, ethnicities, religions and social norms (Ghemawat, 2001). In addition Hofstede (1994) proposed a more complex and complete demonstration of the cultural dimension. He

(17)

16 divided culture into six dimensions. Since these cultural aspects are being researched by a fellow researcher at the same time as this research is taking place there will be no focus on the cultural aspects in this article. The A-dimension of the CAGE model relates to the

administrative distance and consist of colonial ties, trade relations, currency, legal origin and transparency international corruption perception between the host- and home country. Institutional frameworks are therefor of great importance and influence the organization’s strategic behavior (Ghemawat, 2001; Peng, 2002; Hutzschenreuter et al. 2015). Frameworks consist of two different constraints, formal and informal and may influence entry modes directly (Ghemawat, 2001; Hutzschenreuter et al. 2015; North, 1990; Peng, 2002; Scott, 1995). When a firm has to work with an informal-based country, and thus has to build trust, interpersonal relationships and mutual understanding, the best strategy is to create networks and alliances. When competing in a formal based country, growth strategies can therefore be acquisitions (Peng, 2002). Geographical distance is the G-dimension and represents the physical distance and geographical overlapping features (e.g. the presence of a river or sea, climate, etcetera). There is found to be U-shaped relationship between share of acquisition and geographical distance (Malhotra & Gaur, 2014). The E-dimension represents the Economic distance. This referrers to the difference in income, wealth, infrastructure and knowledge between the population of the host- and home country. Ghemawat (2001) states that these four distances need to be considered when evaluating risk and uncertainty of

entering a host country and the importance of them may differ per situation. Since the topic of this article is related to SOEs, SWFs and nationalistic feelings, the main focus is going to be on the administrative distance (AD) since it relates the most to this topic.

The distance-organizational outcome can be aggraviated or alleviated based on industry characteristics and firm characteristic (Hutzschenreuter et al., 2015). One of the firm’s characteristics, experience, is found to substantially weaken the negative effects of cultural distance (Cho & Pamanabhan, 2005). Organizations may overcomes distance and thus uncertainty, with FSA’s (Ghemawat, 2001). Nevertheless prior presence in a target country is also found to not influence the share of ownership (Chari & Chang, 2009). In addition, Hutzschenreuter (2015) addresses the more complexness and versatile aspects of distance that influences the costs and risk of doing business abroad. As distance increases firms tend to favor non-equity modes (low-commitment) over equity modes (high-commitment). Nevertheless different dimensions of distances may have different effects (Arslan and Limo, 2011).

(18)

17

2.6 Nationalism and Constructive Patriotism.

Shi et al (2016) state that the salience of nationalist politics weakens the negative influence of geographic distance, religious (dis)similarity; political regime (dis)similarity; and resource complementarity on the level of opposition that an SOE faces in a target state. Because geopolitical theory mainly concerns international relations among states and political setting, the importance of politics is stressed and this can be related to theory on both SWF’s and SOE’s investment strategies.

In addition Zaheer (1995) demonstrates that there is another disadvantage of doing business abroad due to the liability of foreigness (LoF). Zaheer demonstrates that costs arise directly associated with spatial distance, firm-specific costs related with the host location, costs resulting from the home-country and costs resulting from the host-country. Hereby there is a clear distinction between accidental and discriminatory LoF. Accidental LoF is something that can be overcome resulting from experience, discriminatory LoF however comes from explicit regulations, prejudices and nationalism. This level of nationalist feelings in a host-country therefore is of great concerns when choosing how and where to enter a new market. Nationalism is connected to, but not the same as constructive patriotism (Davidov, 2009). Both represent a different set of feelings. Nationalism relates to the intensity of uncritical loyalty in feelings and closeness towards one’s nation while constructive patriotism relates to the questioning constructive critical feelings and closeness towards one’s nation (Schatz, Schaub & Lavine, 1999). Where nationalism is connected to, among others, perception of foreign threat and political disengagements, constructive patriotism is related to political involvement and interest.

To summarize, there are different state owned investors (SOEs and WF’s), with different motives/determinants to invest in home- and host- location, with different investment strategies and different influences of the salience of nationalist politics and distance. The literature up till now lacks a clear distinction between how SOEs and SWFs differ in their cross-border M&As ownership strategies. Besides this, on an international level the sovereignty is playing a bigger role than ever but the literature is lacking knowledge about the influence of nationalist feelings of a host-country population on government owned investors’ cross-border ownership strategies. Anticipating risks in foreign markets and developing creative strategies how to manage government owned institutions will become increasingly important capabilities over the next decade (Bremmer, 2014). For this reason this thesis would like to address the following research question:

(19)

18 'How do the nationalistic feelings of the target country population and the

administrative distance between acquiror and target country influence the cross-border M&As ownership strategies of SWFs and SOEs and how do SWFs and SOEs differ?'

3. HYPOTHESES DEVELOPMENT

3.1. Administrative distance

To build further on the Ghemawat’s CAGE model, the administrative distance stresses the importance of the political situation in a target country. The less one knows about the way business is done in a country, the greater information asymmetry and potential hazard to doing business is going to be (Ghemawat, 2001). As SWFs and SOEs are different players playing a different game they are tend to be influenced by the dimensions of distance

differently. SWF’s enter the market as an investor instead of a competitor, therefor they may not face the same problems as SOEs. When taking cultural distance and geographical distance there can be demonstrated that they affect SWFs in a different way (Chhaochharia & Laeven , 2008). They demonstrate that cultural aspects, notably similarity in religion, drive the

allocation of investment of many but not all SWFs. In addition partial equity acquisitions are preferred over full acquisitions when making cross-border acquisitions when there is a greater cultural distance (Chari & Chang, 2009; Chi, 1994; Pisano, 1989).

When choosing location, SWFs are likely to invest in countries sharing the same culture, and investment value will be higher if the bilateral trade between the acquirer and target countries is higher (Rossi & Volpin ,2004; Ferreira, Massu & Matos, 2010; Megginson et al, 2013). In contrast, others argue that there is a negative relation between political

relations and where/how much to invest (Knill et al., 2012), creating a disagreement on the subject. Nevertheless as discussed previously, FSA’s may help an organization to overcome distances.

As distance increases, the information asymmetry increases which can cause potential harm to the organization. As a result ventures have preferred shared ownership to minimize the potential hazards the asymmetry can inflict (Malhotra & Gaur, 2014). However working together when separated by administrative distance can cause problems, especially with integrating and cooperating. Deriving from this there can be stated that SOEs are more

vulnerable than SWFs. SOEs are operational companies and therefor much more vulnerable to these managerial problems. The difference between publicly and privately owned

(20)

19 organizations are already demonstrated due to the difference in risk appetite between the two (Malhotra & Gaur, 2014). If the acquiring is publicly, it has an ability to take greater risks and assume more control in the distant host locations.

For this reasons AD impacts the SOEs more heavily than SWFs. The following hypotheses derives from this:

H1: AD between home- and target country has a negative influence on the share of acquisition for both SOEs and SWFs, but is stronger for SOEs.

3.2 Nationalistic feelings

As discussed previously, nationalism and constructive patriotism are related to two different feelings (Schatz et al., 1999). Where nationalism relates to the uncritical loyalty and closeness towards one’s nation, constructive patriotism relates to the constructive critical and closeness towards one’s nation. These different sets of feelings relate to other related constructs. Nationalism is closely connected to, among others, perception of foreign threat and political disengagements, while constructive patriotism is related to political involvement and interest. Deriving from this there can be stated that in the higher ranks and decision making part of the population constructive patriotism plays a bigger role that nationalism. Related to this, SWFs are mainly concerned with investing and thus the decision makers and SOEs are related to operating on the market and thus the consumers. In addition, since the SOEs are more

involved with the operational part of doing business than the SWFs, which tend to be passive investors (Kotter and Lel, 2011). Therefore, SOEs are much more vulnerable to LoF, both discriminatory and accidental. Acquisitions made by SWFs are less vulnerable to LoF, since most of the company culture stays the same. The following hypotheses derive from the fact that SOEs and SWFs deal with different issues on different levels.

H2: The level of nationalism in the target country has a negative influence on share of acquisition for both SOEs and SWFs but the relation is stronger for SOEs.

H3: The level of constructive patriotism in the target country has a negative influence on share of acquisition for both SOEs and SWFs but the relation is stronger for SWFs.

To answer the research question, the hypotheses that have to be researched are demonstrated in figure 1.

(21)

20 Figure 1. Conceptual model

4. DATA AND SAMPLE

The difference between SOEs and SWFs and their ownership strategies will be researched using different datasets on acquisitions by SOEs and SWFs gathered from Zephyr. The nationalist feelings variable will be based on information from the ISSP National Identity Module (Gesis, 2016). The ISSP National Identity Module is a cross-national survey that deals with issues regarding respondents’ national identification, national pride, support for their own nation, attitudes towards national and international issues, attitudes towards foreigners and foreign cultures, etc. The data on attitudes and nationalistic feelings of a population will be combined with the data on acquisitions by SOEs and SWs to create a complete dataset which will address the gap in the literature. In addition the distance between countries will be calculated using Ghemawat’s online CAGE distance calculator (Pankaj Ghemawat, 2016). Together, these quantative results will be combined and the hypotheses will be analyzed using SPSS. The home-country, year of acquisition and industry of the SWFs/SOEs will be used as control variables to eliminate possible alternative explanations of the relationship. After the analysis the results will be demonstrated using tables.

The variables gathered directly from Zephyr on both SWFs and SOEs are: Deal number, announced date, completed date, cross border (Y/N), acquirer name, acquirer country code, parent acquirer, acquirer country code, acquiror world region, target major sector, target name, target country code, target world region, amount of shareholders target firm, deal type, acquired stake, final stake, deal status, deal value, deal headline, acquiror advisor name. The acquisitions before 2008 where excluded in the data due to the usability of the results, since the crisis the scene has changed. The gathered data therefor has a timespan of 1/1/2008 up to 31/12/2013. In addition, SOEs are only included when coming from a nation possessing a SWF.

The variables gathered from the ISSP National Identity Module include v19 ’The world would be a better place if people from other countries were more like the [Country

(22)

21 Nationality of Respondent’, v20 ‘Generally speaking, [Respondent's Country] is a better country than most other countries’, v25 ‘How proud are you of [Respondent's Country] in the way democracy works?’, v28 ‘How proud are you of [Respondent's Country] in its social security system?’ and v34 ‘How proud are you of [Respondent's Country] in its fair and equal treatment of all groups in society?’. According to Davidov (2009) nationalism and

constructive patriotism emerge as two distinct constructs. In some countries however, some items measured both constructs. Variables 19 and 20 respectively represent the ‘nationalism’ construct while v25, v28 and v34 represent the ‘constructive patriotism’ construct. As described earlier, the two constructs describe a different set of feelings (Schatz, Schaub & Lavine, 1999). The feelings and closeness towards one’s nation also has been labeled as political national pride (Hjerm, 1999). For this reason the blind loyalty or nationalism in political national pride is the best construct to use when looking for an extent of the salience of nationalist politics. Both constructs, nationalism and constructive patriotism, where tested for internal validity. To demonstrate the variety in and differences between the countries in the amount of nationalistic and constructive patriotic feelings their population has, table 2 and 3 are added in the appendix.

Acquisitions are only included in the final dataset, when the inhabitants of the target country also participated in the ISSP National Identity Module (Gesis, 2013). The target countries included are the following: Belgium (BE); Switzerland (CH); Czech Republic (CZ); Germany (DE); Denmark (DK); Spain (ES); Finland (FI); France (FR); Great Britain (GB-GBN); Georgia (GE); Hungary (HU); Ireland (IE); Israel (IL); India (IN); Japan (JP); Korea, Republic of (KR), South; Mexico (MX); Norway (NO); Philippines (PH); Portugal (PT); Russian Federation (RU); Sweden (SE); Turkey (TR); United States (US); South Africa (ZA).

4.1 Dependent variable

The dependent variable in this research is the share of acquisition. In this research, the focus will be on cross-border deals, so excluding domestic acquisitions. The deals included are acquisitions. The data is gathered from the Zephyr database and in this database the values are given in percentages. The percentages range from 0.006 % up till 100% and are copied to the working file. The share of acquisitions is measured in percentage of total ownership.

4.2 Independent variables

The four independent variables included in this research are level of nationalism in the target country, level of constructive patriotism in the target country, AD and type of organization.

(23)

22 There are two types of organizations that make acquisitions, SOEs and SWFs. The variable is measured by a dummy variable where SWFs are given the value 0 and SOEs are given the value 1.

The variable AD is segmented in different parts (Pankaj Ghemawat, 2016). The different segments are Trade Bloc, Currency, Colony/colonizer, Legal origin and Index score on Transparency International Corruption Perception Index Score. For each different segment a dummy variable is created, the variable will take the value 1 when there is a match and it will take 0 when there is no match. After doing this a categorical variable is created for the amount of matching segments (0 matches, 1 match, 2, matches, 3 matches, 4 matches, 5 matches). The more matches, the lower the AD.

Both nationalism as constructive patriotism variables come from a secondary database (Gesis, 2016). The items predicted to create the nationalism variable are: v19 ’The world would be a better place if people from other countries were more like the [Country Nationality of Respondent’, v20 ‘Generally speaking, [Respondent's Country] is a better country than most other countries’. Both items where measured on a 5-point liker scale ranging from ‘agree strongly’ to ‘disagree strongly’. The lower the participants score on the scale, the more

nationalistic feelings the population of that country has. Due to valence, the scale was reversed during the analysis. The ‘can’t choose’ and ‘no answer’ options were reported missing.

As noted in the previous part, the constructive patriotism variable comes from a secondary database as well. The items predicted to be part of this construct are v25 ‘How proud are you of [Respondent's Country] in the way democracy works?’, v28 ‘How proud are you of [Respondent's Country] in its social security system?’ and v34 ‘How proud are you of [Respondent's Country] in its fair and equal treatment of all groups in society?’. These items where measured on a four-point Likert scale ranging from ‘very proud’ to ‘not proud at all’. The lower the participants score on the scale, the more constructive patriotism feelings the population of that country has, for this reason the scores were reversed. Again, the ‘can’t choose’ and ‘no answer’ options were reported missing.

4.3 Control variables

In the analysis, there is controlled for a number of factors that may have influenced the share of acquisition. Firstly, sector of target company. As demonstrated, the influence of the

(24)

23 target companies are dived up in one of the nine sectors from the Zephyr database.

Respectively machinery, banks, construction, gass/ water/electricity, Transport, wholesale and retail, primary sector, metal and metal products and other services, Secondly, the completed date can be of influence, specifically the year. After the crisis of 2008 the business world has changed. This is in line with findings from Morosini and Singh (1988), who state that it is important to control for the year of the acquisition since national and international economic and financial conditions change yearly.

The acquiror country has been included as a control variable as well. This is due to the different concerns risen against international operations of both SOEs as SWFs

(Chhaochharia & Laeven, 2008; Bernstein, Lerner & Schoar, 2010; Kyge, 2016, Rudy et al., 2016). Where some are concerned about the political influences it is interesting to see if some countries indeed have a higher commitment to making cross-border investments. Not only the concerns about strategical investments concerning politics are the reason behind this. Some researchers have already proven that investments from countries may significantly differ, for example the Norway’s SWF never makes big share of acquisitions (Dewenter et al, 2010).

The definition and the source of the data is demonstrated per variable in table 4. Table 4. Definitions of variables

Independent variables

Organization Dummy variable that equals 0 when the organization is a SWF and 1 when the organization is a SOE.

Zephyr (2016), Murtina & Scalera

(2016)

AD

The different segments of AD are

Pankaj Ghemawat (2016)  Trade Bloc  Currency   Colony/colonizer   Legal origin

 Index score on Transparency International Corruption Perception Index Score The variable will be constructed around the acquiror and target country having matching segments. Each matching segment will add a value of 1 to the AD construct, e.g. if the acquiror country and target country have a matching trade bloc and a legal origin, the value of AD will be 2.

Nationalism

Scale constructed from two items which are measured on a five-point Likert scale ranging from ‘agree strongly’ to ‘strongly disagree’. The scale has been reversed due to the valence. The original items are:

Gesis (2016)  ‘The world would be a better place if

people from other countries were more like the [Country Nationality of Respondent’ 

 ‘Generally speaking, [Respondent's Country] is a better country than most other countries’

(25)

24

Constructive Patriotism

Scale constructed from three items which are measured on a four-point Likert scale ranging from ‘very proud’ to ‘not proud at all’. The scale has been reversed due to the valence. The original items are:

Gesis (2016)  ‘How proud are you of [Respondent's

Country] in the way democracy works?’  ‘How proud are you of [Respondent's Country] in its social security system?’   ‘How proud are you of [Respondent's Country] in its fair and equal treatment of all groups in society?’

Control variables

Sector

Target company sector. Dummy’s where created to do the analysis. The following sectors where distinguished, whereas Machinery was taken as the baseline group:

(2016)  Banks 

 Chemicals, rubber, plastics, non-metallic products

 Construction  Education, Health  Food, beverages, tobacco  Gas, Water, Electricity  Hotels & restaurants  Insurance companies

 Machinery, equipment, furniture, recycling

 Metals & metal products   Other services 

 Post and telecommunications 

 Primary Sector (agriculture, mining, etc.)   Publishing, printing

 Transport 

 Wholesale & retail trade   Wood, cork, paper   X (non-known) Year

Year of completed transaction. Dummies were created to do the analysis. For the years reaching from 2008 to and including 2012 dummies where made, 2013 was taken as baseline group.

(2016)

Acquiror country

Home country of acquiring organization. Dummies were created to do the analysis. The following acquiror countries where distinguished, whereas the USA was taken as the baseline group: United Arab Emirates, Azerbaijan, Canada, China, Ireland, Korea, Kuwait, Libya, Malaysia, the Netherlands, Norway, New-Zealand, Oman, Qatar, Russia, Saudi Arabia, Singapore, the United States of America.

(2016)

4.4 Methodology

The share of acquisition that SWFs and SOEs make in relation to the other variables is tested using different models. The first hypothesis states there is an interaction effect between type of organization and AD in relation to the share of acquisition. Since share of acquisition is

(26)

25 measured on an interval level and type of organization has been recoded into a dichotomous variable, these can be tested both on interval level. In contrast, AD is a categorical variable which is predicted to interact with type of organization. Deriving from the nature of the measurements a factorial ANOVA has to be run. The model is as follows:

DV(Share of acquisition)= β0 +β1Organization + β2AD + β3Interaction + ε1

The second hypothesis is related to the interaction between type of organization, level of nationalism in the target country and share of acquisition, which are all interval variables. In addition, constructive patriotism is predicted to act as a moderator on the relation between type of organization and share of acquisition as well. To predict the outcome of the

dependent variable from the predictor variables, the moderating variable and the interactions, one would expect to run two regression analyses (OLS) wherefore the model would be as follows:

DV(Share of acquisition)= β0 +β1Organisation + β2Nationalism + β3InteractionNationalism

Con1TargetSector+ Con2Year + Con3AcquirorCountry + ε1

DV(Share of acquisition)= β0 +β1Organisation +β2ConstructivePatriotism +

β3InteractionContructivePatriotism +Con1TargetSector+ Con2Year + Con3AcquirorCountry

+ ε1

This research has chosen not to follow this method, but has chosen to split the model. The reason why this has been done is because with the SPSS macro of Hayes (2012), the Johnson-Neyman method is presented. This method gives an detailed breakdown of the boundaries of the significant zones of the interaction effects. The macro computes confidence intervals for the indirect effect of type of organization on share of acquisition but a big

limitation of this macro is that it can only handle one moderating effect at a time. Since there has been limited research on nationalism and constructive patriotism in relation to share of acquisition, this research prioritized finding any significant results related to these moderating variables in contrast to researching the roles of the control variables. Therefore, two separate model are going to be run. One without the control variables, and if any significant results are found a second will be run. The second will investigate if the significant results found will hold when controlling for the influence of control variables. The tests without the control variables follow the next methods:

(27)

26 DV(Share of acquisition)= β0 +β1Organisation + β2ConstructivePatriotism+ β3Interaction +

ε1

After the test has been run, there will be checked if the (possible) significance holds against a Hayes’s macro regression analysis involving the control variables as well. Only if there are significant interaction effects in the first method, the secondary test will be run. The methodology for these analyses are:

DV(Share of acquisition)= β0 +β1Organisation + β2Nationalism + β3Interaction +

DV(Share of acquisition)= β0 +β1Organisation + β2ConstructivePatriotism+ β3Interaction +

The role of the control variables will be tested separately as well to look for spurious relations and any other significant influences on the share of acquisition.

4.5 Descriptive Statistics

In total 1179 acquisitions were included in the original database. After checking for the availability of data on Nationalism and Constructive Patriotism in the target countries the database shrunk to 561 acquisitions. Of these acquisitions 63% were made by SOEs and 37% were made by SWFs. Of these acquisitions the average was an 30.7% share of ownership but with a median of 7.00%. Both variables, type of organization and share of acquisition have normal skewness and kurtosis. An one-way ANOVA was conducted to look for differences in means between SOEs and SWFs and their acquisitions. There is no significant difference between SOEs(M=31.6, SD=40.44, 95%CI[27.21, 35.00] and SWFs (M=28.86, SD=35.30, 95%CI[23.33, 34.39] in just looking at their share of acquisition according to the one-way ANOVA (F(1, 485)=.535 with p= .465. This is visualized in table 5 and table 6.

Table 5. ANOVA output organization and share of acquisition

SS DF MS F p

Organization 806,41 1 806,41 0,535 0,465 Error 731569,27 _{485 1508,39}

(28)

27 Table 6. Descriptive statistics of share of acquisition per organization

Organization Mean SD N SWF 28,86 35,30 159 SOE 31,60 40,44 328 Total 30,70 38,82 487

4.5.1 Factor Analysis

The lead was followed from earlier research that proofed nationalism and constructive

patriotism emerge as two distinct constructs (Davidov, 2009). As described in the independent variable section (4.3) there is a prediction which items will make up what scale, since there are multiple items concerning nationalism and constructive patriotism (visualized in table 7). Nevertheless, since the constructs are closely related a direct oblimin-rotation factor analysis with principal axis factoring method was conducted while looking for two factors. The sample size is extremely large (N=7660) wherefore a Kaiser-Meyer-Olkin measure of sampling adequacy must be conducted. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the analysis, KMO =.708 and therefor indicates that the patterns of correlations are relatively compact and the analysis is yielding distinct and reliable factors. Bartell’s test of sphericity also provides significant results χ2 _{(10)= 39948.52, p<.001, stating that the}

correlation between variables are statistically significantly different from zero. According to Keiser’s criteria, factors must have an eigenvalue above one, as is twice the case in this

analysis. Two components have eigenvalues over Kaiser’s criterion of one and in combination explained 69.36% of the variance. In agreement with Kaiser’s criterion, examination of the scree plot revealed a levelling off after the second factor, as visualized in figure 2. Table 7 shows the factor loadings after rotation. The items that cluster on the same factors suggest that factor 1 represents constructive patriotism and factor 2 nationalism, as was expected due to prior research (Davidov, 2009). In addition, there was proven by the factor correlation matrix ,r=.453, that indeed there is a correlation between the two constructs. To test the reliability a Cronbach’s alpha was conducted for both scales and presented an value of at least .68 for both, presented in table 5 as well together with the descriptive statistics. The alpha is not as high as one may wish, however this research states that both scales are reliable. There are some issues with interpreting alpha since, as it is dependent on the number of items on the scale (Cortina, 1993; Pedhazur & Schmelkin, 1991). In the case of both nationalism and constructive patriotism there is a low number of items, respectively two and three. Research has showed that as the number of items on the scale increases, alpha will increase. As such, a

(29)

28 low number of items with a relatively high correlation between the items can result in the same alpha as a large number of items with an average correlation. In the case of the nationalism and constructive patriotism there is such a small base of items for the scales, therefore this research argues that they are reliable.

Figure 2. Scree plot factor analysis

Table 7. Factor Analysis Constructive patriotism and Nationalism

Constructive

Patriotism Nationalism The world would be a better place if people from other

countries were more like the [Country Nationality of

Respondent]. (reversed) -,061 ,767

‘Generally speaking, [Respondent's Country] is a better

country than most other countries (reversed) ,097 ,655

How proud are you of [Respondent's Country] in the way

democracy works? (reversed) ,720 -,003

How proud are you of [Respondent's Country] in its social

security system (reversed) ,725 -,057

How proud are you of [Respondent's Country] in its fair and

equal treatment of all groups in society (reversed) ,603 ,093

Eigenvalue 2,36 1,11

% of Variance 47,11 22,25

Cronbach's Alpha ,728 ,680

(30)

29 SD ,76 ,97 Min 1 1 Max 4 5 N 37660 37660 4.5.2 Correlation matrix

To test if there are any correlations among the dependent variable and the independent

variables a correlation matrix has been conducted. The results are presented in table 8 together with the descriptive statistics of the relevant variables. Note that AD has been separated into dummy variables. Due to the categorical nature of the variable, dummies were created to conduct the correlation matrix.

Table 8. Descriptive Statistics and Correlation Matrix

Variables M SD min max 1. 2. 3. 4. 5. 6. 7. 8. 9.

1. Organization ,70 ,47 0 1 1 2. Share of acquisition 31,00 38,80 0 100 ,033 1 3. Nationalism 3,30 0,32 0 5 ,028 -,024 1 4. Constructive patriotism 2,70 ,26 0 4 ,040 -,023 ,468 ** ₁ 5. 0 matching AD segments ,20 ,40 0 1 -,047 ,080 -,162 **_-,260** ₁ 6. 1 matching AD segment ,40 ,49 0 1 ,241 **_-,137** _-,085 _{-,062 -,409}** ₁ 7. 2 matching AD segments ,24 ,43 0 1 -,124 ** _-,097* _,097* _,205**_-,284**_-,467** ₁ 8. 3 matching AD segments ,11 ,31 0 1 -,089 ,071 ,190 ** _,181**_-,176**_-,290**_-,201** ₁ 9. 4 matching AD segments ,04 ,20 0 1 -,089 * _,267** _,024 _-,053 _-,106*_-,174**_-,121**_{-,075 1}

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). N=487 (listwise exclusion of missings)

5. RESULTS

The next sections will elaborate on the different variables included in the hypotheses and will discuss the results. Firstly, it will start off with discussing the different statistical tests that are conducted.

5.1 Statistical Tests

The data was collected from the Zephyr Database, Ghemawat’s online CAGE comparator (Pankaj Ghemawat, 2016) and the ISSP module (Gesis, 2016). The data was merged into Excel and after that copied into SPSS. The variables were tested for skewness, kurtosis and

(31)

30 normality tests. All the variables were normally distributed. Only Constructive Patriotism had a small positive kurtosis but this can be induced by the Likert scale used, with a neutral option in the middle.

To test the first hypothesis an factorial ANOVA is conducted to present the moderation effect of AD on the relation between type of organization and their share of acquisition. This is due to the categorical nature of the AD variable. To be more specific, there is the possibility to have 0, 1, 2, 3, 4 or 5 matching segments resulting in six categories. Since there is the need to know if the share of acquisition of these groups differ from each and if there is an interaction effect, an factorial ANOVA is the preferred test.

Hypothesis 2 and 3 both ask for a moderating analysis. As explained in section 4.4 (methodology) this research does not follow the regular regression (OLS) analysis for

interaction effects. This research has chosen not to follow this method, but has chosen to split the model. Since this research is rather exploring the implications that nationalism and

constructive patriotism have in relation to the share of acquisition that SOEs and SWFs make. The SPSS macro of Hayes (2012) is a model that is able to look at the direct and moderating effect of its variables. The reason why this choice has been made is because the Johnson-Neyman method is able to be conducted, in contrast to a regular regression analysis (OLS). The Johnson-Neyman method gives an detailed breakdown of the boundaries of the

significant zones of the possible interaction effects. Implications for this may be for example that on the first hand it may look like there is no interaction effect, but when looking at the Johnson-Neyman method there may be found that there is one but only in the smaller

segments. This may have great implications for managers. For this reason two separate Hayes macro’s are conducted for nationalism and constructive patriotism. This will investigate the direct and interaction effects of the variables on the relation between type of organization and share of acquisition. To start off with the interaction effects, first the involved variables have to be standardized to make any results interpretable. When conducting the moderating test, the confidence intervals have to be set on 95% and there must be resampled 5000 times in both cases. If in the first round of moderation analysis statistical results have been, a second test will be conducted. Control variables will be included in the second analysis, to see if the statistical significant results will hold when controlling for spurious relations.

Cross-border M&A ownership strategies of SWFs and SOEs and the role of administrative distance and nationalist feelings of a nation