Sovereign wealth funds : the effect of investment vehicles on the performance of target firm

(1)

Sovereign Wealth Funds:

The effect of Investment Vehicles on the Performance of

Target Firm

Yifan Liu

Student ID: 11377275

Supervisor: Dr. Vittoria Scalera

Date of Submission: 23

rd

_{, June, 2017}

University of Amsterdam

MSc Business Administration: International Management

Master Thesis – International Management track

(2)

Statement of originality

This document is written by Student Yifan Liu who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents

(3)

Abstract

Sovereign wealth funds (SWFs) have been important players in the global financial market since decades ago. Their tremendous monetary value and nature have drawn the attentions from both political and academic worlds. This thesis intends to investigate how SWFs invest cross-border through vehicles affect the performance of the target firms; and how bilateral agreement and strategic industry affect the performance of target firms. Return on assets (ROA) and return on equity (ROE) are used as the measurements of performance. Cultural and geographical distances between home and host countries, Gross Domestic Product (GDP) and corruption perceptive index (CPI) of SWF home country, GDP of target firm’s country, and the year of acquisition are all controlled variables of this research. The final sample covers 140 investments from 24 parent SWFs, and 16 countries. Whereas the data sample is obtained from published literature and official websites. The result of this study presents a statistically significant and negative relation between the use of vehicle and performance of target firm. However, the strategic industry positively moderates the performance of the target firms.

Keywords: Sovereign wealth funds, performance, vehicle, bilateral agreement,

(5)

1. Introduction:

Sovereign wealth funds (SWFs) are funds owned and/or controlled by sovereign governments, aiming at maximizing long-term risk-adjustment returns (Johan, Knill, and Mauck, 2013). The monetary value of SWFs is tremendous, usually with multibillion dollar assets invest in both domestic and foreign markets through various channels. Up until 2016, SWFs hold $7.4 trillion in total assets. Incredibly, 67.10% of SWFs were established after the year of 2000 (Sovereign Wealth Funds Ranking, 2016). The rapid growth of SWFs in number, size, and nature of the investments has drawn considerable attentions in the last decades (Aguilera, Capapé, and Santiso, 2016; Johan et al., 2013). However, due to the sovereign ownership, and the lack of transparency, SWFs often attract hostilities from host countries (Johan et al., 2013; Tassell and Chung, 2007; Knill, Lee, and Mauck, 2009). Being controlled by sovereign states indicates that politics might affect the SWF’s strategy and investment preference (Dyck & Morse, 2011, Bernstein, Lerner, and Schoar, 2009). Despite the presence of liabilities in cross-border investments and hostilities from the host government; the increasing size and number of SWFs, as well as the large number of investments indicate that SWFs have become vital players in the global financial market (Beck & Fidora, 2008).

Existing studies regarding SWFs investments have been conducted on several domains. From the political perspective, there are three kinds of studies of SWFs: first, the impact of politics, how political bilateral relations influence SWFs investment

(6)

strategies (Knill et al., 2011); second, the relationship between SWF investments and the return-to-risk performance of target firms (Knill, Lee, and Mauck, 2012; Bortolotti, Fotak, and Megginson, 2009); third, the host country’s regulations toward SWFs (Rose, 2008). From the strategic perspective, previous studies have illustrated several determinants of investment decisions of SWFs, such as the determinants of SWFs investing in private organizations (Johan et al., 2013). When considering the investment decisions, some authors argue that SWFs tend to invest in public organizations (Johan et al., 2013), and countries with fewer protection (Johan et al., 2013; Knill et al., 2011); there are also others who argue that SWFs have a tendency to invest in private organizations (Bortolotti et al., 2015).

Traditional cross-border investments of multinational enterprises (MNEs) have two major entry modes choices: (1) non-equity modes, and (2) equity modes (Pan and David, 2000). Furthermore to that, there are two subcategories under each major entry mode, namely export, contractual agreements; equity joint ventures, and wholly owned subsidiary. All these entry modes are directly invested by the acquirers. However, that does not mean that every investor is able to invest cross-border directly, there are times when investors need to involve other companies into the investments, namely investing through vehicles or intermediaries, which might lead to a better performance. Investing directly and through vehicles could also be considered as alternative strategies (Murtinu & Scalera, 2016). Compared with the traditional entry modes, the indirect entry mode helps the firm to avoid foreign liabilities and

(7)

information asymmetries. For SWFs, vehicles are used to avoid exposures and hostilities from host countries (Murtinu & Scalera, 2016; Aguilera et al., 2016).

Murtinu and Scalera (2016) investigated the cross-border investment strategies of SWFs, specifically, the use of investment vehicle. However, the existing scholar has not presented a study concerning the effects of investment vehicle on SWFs’ acquisition. Under this circumstance, this thesis aims to study whether the use of an investment vehicle is able to affect the performance of target firm. Additionally, various studies have investigated how bilateral relations between home and host countries, and target firms’ industries affect the investment choice and performance of target firms (e.g. Knill et al., 2009). This thesis combines the bilateral agreements between home and host countries and target firm’s industry to investigate how they affect the performance of target firm when SWFs invest cross-border through vehicles. The investment vehicles have been widely studied in financial literature but not in International Business (IB) literature. In order to study how investment vehicle affects the performance, this thesis analogizes the investment vehicle to traditional entry modes, through analyzing the information asymmetry reduction ability of vehicles, and using the transaction cost theory to predict whether the target firm performs better when SWFs invest cross-border through vehicles. Is the performance positively or negatively affected under different circumstances? Specifically, the presence of bilateral agreements between home and host countries, and the different industries of target firms. Previous studies have used multiple theories, like transaction cost theory, resource-based view, to illustrate how firms choose the

(8)

boundary of ownership, location, and entry modes (e.g. Hennart, 1989; Balakrishnan & Koza, 1993). Among all these theories, the transaction cost framework (Williamson, 1975, 1985) is most widely used in the entry modes choice studies. K. D. Brouthers (2002) found firms which are using modes predicted by the transaction cost theory perform better than firms which are using modes predicted by other theories.

This thesis is developed as the following. In the first section, the aim and purpose of this thesis are introduced. In section 2, the previous studies are reviewed about SWFs’ definition, current situation, strategies, investments’ performance, and the introduction of transaction cost theory (TCT) which is used to predict the performance of SWFs investment. Section 3 illustrates the research gap and develops the hypothesis. To identify the research question of whether the target firms perform better when SWFs invest through vehicle, this thesis uses transaction cost theory to predict the performance, and to test whether the bilateral agreement between home and host countries, and target firm’s industry is able to affect the performance. Section 4 illustrates data and methodology where the research model and variable of the thesis are introduced. In section 5, on the basis of Fernandes’ (2014) work, several linear regressions are used to test the model. Section 6 discusses the result, and section 7 states the limitation and suggests the direction of future research.

(9)

2. Literature review:

2.1 Definition of SWFs

Sovereign Wealth Funds (SWFs) are defined as capitalism pools which are owned or/and controlled by governments to invest in stocks, bonds, real estate and other financial instruments, in order to maximize the long-term risk-adjusted returns (Aguilera et al., 2016; Johan et al., 2013). Beck & Fidora (2008) define three common elements of SWFs: “First, SWFs are state-owned. Second, SWFs have no or only very limited explicit liabilities and, third, SWFs are managed separately from official foreign exchange reserves.” (p.6) The funding sources of other pension funds originate from the investment of public pension funds (Johan et al., 2013), while the core funding source for SWFs originates from international reserves. Thus, one of the distinctions from other state-owned organizations is that SWFs have no explicit pension liabilities (Aguilera et al., 2016).

Generally, SWFs are classified into commodity funds and non-commodity funds. The main difference between these two funds is the funding source. Commodity funds (e.g. Government Pension Fund-Global) are funded through natural resources, mostly oil and gas; non-commodity funds (e.g. China Investment Corporation) are funded through foreign exchange reserves (Murtinu & Scalera, 2016). In addition to the variance of the funding sources, SWFs is a kind of heterogeneous groups because of its objectives (Allen & Caruana, 2008). Sovereign Wealth Funds--A Work Agenda 5,

(10)

main objectives: (1) stabilization funds, to insulate the budget and the economy against commodity price swings; (2) saving funds, to conserve nonrenewable resource for the future generation; (3) reserve investment corporations, whose assets are still considered as reserve assets, are established to increase the return on reserves; (4)

development funds, to help projects or industrial policies that might increase a

country’s output growth, and (5) contingent pension reserve funds, to provide for contingent unspecified pension liabilities on the government’s balance sheet (Allen and Caruana, 2008).

2.2 History and current status

The first SWF was established in 1953 by the Kuwait government. It operated on the oil revenue surplus of Kuwait through a foreign office located in London. In 1983, the organization officially changed into a public government entity called the Kuwait Investment Authority (KIA) (Overview of KIA, 2013). Rozanov (2005) is the first one who termed “sovereign wealth funds” to distinguish the structure and the objectives of this type of investors from other sovereign-investors (Bortolotti et al., 2015).

Although the first SWF has been established decades ago, SWFs have only started playing important roles in the global market in the recent years (Beck & Fidora, 2008). This could be seen from the rapid growth of SWFs since 2000 (Aguilera et al., 2016; Sovereign Wealth Funds Ranking, 2016). Ever since the establishment of the first SWF - KIA in 1953, there have been 74 SWFs in the current markets, and the ten

(11)

biggest SWFs hold $5.57 trillion. The assets held by SWFs accounts for around $7.4 trillion in total. Figure 1 in shows the year of establishment of SWFs - 67.10% of SWFs were established after 2000. Figure 2 shows that in 2015, 56.60% of the SWFs is oil and gas related, and 43.40% is other related (Sovereign Wealth Funds Ranking, 2016).

(12)

Figure 1: Year of Establishment

(13)

An impressive development speed does not equate to a matured and well-accepted SWF environment. There have been enormous outcries demanding for stricter SWFs-regulation (Alhashel, 2015). Regulations bring disclosure obligations onto SWFs, to avoid triggering the definition of control under regulations so that SWFs are required to publish their information. When investing in foreign markets such as the United States market (especially in financial market), SWFs tend to acquire stakes less than 10 percent (Rose, 2008). Most studies support further regulation on SWFs, which is in opposition to Rose’s arguments (2008), which argue the United States does not need more regulations (Alhashel, 2015). One reason for the researchers calling for more regulations is their worry about SWFs’ using economic clout for political pursuit (Epstein & Rose, 2009). The political pursuit leads to worries about the national ownership and control, as well as national security (Mattoo & Subramanian, 2009). Different from other studies, Mattoo, and Subramanian (2009) stated regulations, that constrain SWFs, should be issued by the World Trade Organization (WTO), not the host country. However, this kind of supervision might be good for SWFs, but not for the host country (Alhashel, 2015). Another reason for these researchers requiring regulations in their perceptions that transparency can bring SWFs benefits originating from protectionism. In line with the calling, Kotter and Lel (2011) found that the abnormal return of target firm is positively associated with the transparency.

(14)

2.3 Studies of SWFs strategies

The opacity of SWFs hinders researchers from studying their motivations and investment strategies. Previous studies have attempted to analyze SWFs with the information they could reach (Alhashel, 2015). Aguilera, Capapé, and Santiso (2016) categorized the main studies of SWFs into six disciplines: finance, strategy, political economy, economics, international law, and organizational theory. The research on finance, strategy and political economy are mainly related to SWFs’ cross-border investment strategies.

There are two categories of research about the investments of SWFs focusing on the finance discipline. The first category is the short-term and long-term impact on target firms (e.g. Bortolotti et al., 2013; Deweter et al., 2010), and this category is discussed in the sector 2.5. The second category is about the investment strategy, and the preference of target firms (Aguilera et al., 2016). Dyck and Morse (2011) found that SWFs choose target firms along with the national industry objectives. However, this argument is refuted by Avendaño and Santiso (2011), who argued that the political shareholdings have no influence on the SWFs’ strategies. All of their studies on the SWFs shareholdings were conducted at one specific time point, not in a continuous timeline. Miceli (2011) argued that SWFs do not have a systematic preference for any particular industries, they do not invest cluster buy or sell in one particular industry than others, and they tend to follow similar strategies within a period. He also indicated that SWFs are rational investors in a long-turn, which is in

(15)

contrary to Knill et al.’s (2012a) result, Knill et al. (2012a) regarded SWFs as irrational investors because they do not aim to maximize profits. The strategy discipline is mainly about the widespread investments in domestic private firms, the allocation of SWF portfolio, and how host countries and target firms benefit from the investments (Aguilera et al., 2016). Rose (2009) studied how external forces affect SWFs and target firms, he stated that transaction costs have a great impact on target firms, and SWFs change their investment structures to adapt the host country’s culture and regulations to reduce transaction costs, not only for their subsequent investments by the same SWF, but also for other SWFs.

Every research of SWFs cannot avoid the possibility of political pursuit, and the presence of politicians (e.g.Bertoni and Lugo, 2015). Some of the studies even focus on the impact of politics (e.g. Knill et al., 2011). Due to the global wave of privatization and liberalization, the number of state-owned organizations has rapidly decreased (Rodríguez, Espejo, and Cabrera, 2007). Although SWFs are different from other state-owned organizations, they are willing to share the ownership with non-government organizations, or provide technologies and skills to private organizations through various channels (Aguilera et al., 2016). Dyck and Morse (2011) studied how politics impact investment strategies of SWFs. They have found out that SWFs tend to invest half of their risky portfolio in private organizations. Among existing studies about investing in private organizations, part of the studies focuses on the determinants of SWFs investing in private organizations (Bernstein et al., 2009; Johan

(16)

et al. 2013), and the other part of the studies (e.g. Knill et al., 2012b) focuses on the effect of politics on SWFs investments.

Apart from the studies on the politics perspective, researchers have drawn the perspective to the firm level. They explored the SWFs’ strategic governance, and have developed four strategic governance approaches, which are: (1) shareholder activism, to deal with the principle-agency conflict; (2) in-house capabilities, to reduce the information asymmetries when SWFs invest in private organizations; (3) decoupling

and legitimacy, to be developed as a governmental tool, and SWFs could be

legitimated by investing in publicly traded firms; and (4) long-term learning, to acquire the knowledge and know-how. They claimed that SWFs are not static organizations, but dynamic organizations. Therefore, SWFs could shift their pursuits from economic to political. If SWFs are from the same country with target firms, the pursuits might shift from one strategic approach to another.

Johan, Knill, and Mauck (2013) examined the determinants of the SWFs’ investments when directly investing in private organizations. They have also found out that when investing in foreign organizations, SWFs are less likely to invest in private organizations. Compared with the markets that provide protections for investors, SWFs tend to seek for the markets where SWFs have potential to “take advantages of lax investor protection to fulfill political goals if required to do so” (p.23). Moreover, politics have a great influence on the likelihood of direct investment on a private organization (Johan et al., 2013). Their findings present the tendency of SWFs to invest in private organizations which locate in less protected

(17)

foreign markets. Therefore, the SWFs are able to avoid the inspections from the financial regulatory organizations or the surveillance of host countries. However, researchers are neither able to prove the SWFs investments involvement with politics nor to reject the involvement.

Knill et al. (2011) studied the relations between investment decisions and the bilateral political relations. They noted that investment decisions should be considered as a two-stage decision: (1) where to invest, (2) how much to invest. They investigated the role that politics plays in SWFs’ decision making. Comparing their results with FDI literature, SWFs are in favors of investing in countries with weaker political relations. Also they found out that bilateral political relations have a negative influence on SWFs decision makings.

2.4 Vehicles of SWFs

Similar to MNEs, SWFs also use these two kinds of equity entry modes (acquisitions and JVs) when entering a new foreign market. However, due to the nature of SWFs, the direct investment used by MNEs may not fit for SWFs. Because directly investing in target firms would maximize the exposure of SWFs to the host country, and consequently attract more hostilities from the host governments (Murtinu & Scalera, 2016). When it comes to investing in foreign markets, SWFs could invest through vehicles to avoid the hostilities and liabilities, because vehicles locate in the host country or a third country are more likely to be accepted (Murtinu & Scalera,2016). Previous literature has studied theories, such as resource-based view,

(18)

and transaction cost theory, to predict why, how, and where MNEs to invest, and predict the performance of investments (e.g. Hennart, 2007; Dyck & Morse, 2011). Despite the direct investment, in this SWF case, indirect investment (invest through vehicles) could be considered as an alternative, which also implicates control, risk, resource, and performance (Murtinu & Scalera, 2016).

Murtinu and Scalera (2016) shed light on investigating under what circumstances that SWFs tend to use vehicles. They found out that opaque SWFs and politicized SWFs are more likely to invest cross-border through an investment vehicle. Moreover, SWFs are also likely to invest through an investment vehicle when they acquire major equity stakes in cross-border firms or firms in a strategic industry. Furthermore, they identify three kinds of vehicles: (1) non-SWF majority-owned financial vehicles, (2)

non-SWF majority-owned corporate vehicles, and (3) other SWF investment vehicles.

2.5 Performance of the target firm

Bortolotti, Fotak, and Megginson (2009) stated that SWFs investments have positive effects on the target firm value, but the performance is not linear related to the investments. The positive effects on target firm value only last for a short term, but in the long-term (from 2 to 3 years), the performance of target firms shows a negative relation with the investments, and the extent of “negative relation” along with the size of acquired stakes. The underperformance is more severe when the SWF acquires a large stakes or acquires a seat on the broad, the negative impact associated with the larger stakes points out that SWFs do not create value through monitoring,

(19)

and might even set a conflict with minority shareholders due to the inefficiency of government ownership.

Fernandes (2014) noted that many studies indicate public enterprises have a relative inefficient performance because the SWFs pursue objectives beyond maximizing the financial returns, a public enterprise might enter a foreign market to access special resources for the home country government’s ambition. Fernandes (2014), Dewenter, Han, and Malatesta (2010) had vary results from Bortolotti et al. (2009). They argued that the target firms’ value is positively affected by the holdings of SWFs. Also this positive effect could be traced directly from the SWFs’ characteristics – long-term investment view, enormous available capital, and political connections. Even for the long-term (over 3- and 5- years), the target firm still has a significantly positive stock price returns. Meanwhile, divestment announcement has a negative impact on the target firm value (Dewenter et al., 2010). The empirical result of Dewenter et al. (2010) indicated that (1) the effects of direct investments on the firm value are greater than indirect investments. Their result is contrasting to Bortolotti et al. (2015) result, Dewenter et al. (2010) argued that the performance of direct investment is weaker than the performance when SWFs investing through investment vehicles or subsidiaries. (2) Reaction of the first investment is greater than subsequent investments. However, Bortolotti et al. (2015) studied the performance again, and they insisted that the positive reaction only last for a short period. In addition to that, their data shows that the performance of target firms, which is invested by SWFs, is weaker than firms invested by private organizations due to

(20)

“SWF discounts”. SWFs investments worsen the long-term performance in comparison to private-sector investments.

The existing evidence presents that politicians care more about the political pursuits, therefore tend to be bad owners, who have a negative impact on firms’ value (Bortolotti et al., 2015). However, the defenders of SWFs argue that the managers of SWFs tend to make the effects of SWFs’ investments as similar as the effects of other portfolio firms’ investments (Fernandes, 2014). In agreement with Dewenter et al. (2010), Fernandes (2014) found out that SWFs’ ownership positively impact on firm value. However, Bortolotti et al. (2015) showed the impact of SWFs investments is weaker than the impact of private-sector investments due to the involvement of politics.

When SWFs invest in firms, the acquisition size mainly has three effects on firm value in previous literature. (1) The firm value is positively associated with the size of acquired stakes. The larger the size of acquired stakes, the higher the firm value is (Fernandes, 2014). (2) Firm value does not significantly associate with the size of acquired stakes, Bortolotti et al. (2015) argued that different investors’ (SWFs) acquisitions might lead to different market reactions. For example, the market reaction is positive when the investor is Norwegian SWFs, but the market reaction is negative when the investor is a “political” fund (Bortolotti et al., 2015). It is possible that market reaction is not related to the size of acquired stakes, but determined by whether the SWF is a “political” fund. (3) The firm value is nonlinear associated with the size of acquired stakes, Dewenter et al. (2010) used the abnormal returns to check

(21)

the effects of SWFs investments on firm value. They stated that the abnormal returns increase along with the size of acquired stakes. At the maximum point, the abnormal returns start to decrease. The trend like an invert U-shape.

For investments in domestic and foreign firms, different authors have different results about the effects of SWFs investment on firm value. Dewenter et al. (2010) stated that due to the association with politicians, SWFs have the possibility to access the information about government actions prior to other investors, this might affect SWFs investment strategies and performance, especially when the target firm locates in the same country with SWFs, or in a highly regulated industry. The firm value is higher when SWFs invest in domestic firms than in foreign firms, and the value is higher when SWFs invest in a familiar industry than in an unfamiliar industry. Fernandes (2014), however had a different result, it illustrated that most higher firm values are associated with foreign SWFs investments, while the domestic impact has very limited evidence.

2.6 Transaction cost theory

Many theories have been used in related studies, namely transaction cost theory (Hennart, 1991), resource-based view (Erramilli et al., 2002), institutional theory (Brouthers & Hennart, 1998), and Dunning’s elective framework (Agarwal & Ramaswami, 1992) to analyze what affect the choice of entry modes. From the theory perspective of research, transaction cost theory (TCT) is the most widely used theory. The transaction cost is raised from the process of getting information, because firms

(22)

encounter information asymmetries and imperfect intermediate when investing in the foreign markets (Coase,1937). Dahlman (1979) improved the transaction cost theory, and identified the transaction cost into three categories: search and information costs, bargaining costs, and policing and enforcement costs. Even though, scholars argue transaction cost theory needs to be modified for the service industry (Erramilli and Rao, 1993), it is still the most applicable theory to all industries compared with other theories (Zhao et al., 2004). Brouthers et al. (2003) found firms choose entry mode predicted by transaction cost theory perform significantly better than firms do when they choose other entry modes. Meanwhile, one meta-analysis of TCT-based entry mode study has proven that TCT has done a good job when explaining the decision of entry modes (Zhao, Luo, & Suh, 2004), and North (1992) stated that firms perform better if they facilitate a low transaction cost.

This thesis considers invest through vehicles as an alternative entry mode compared with the direct entry modes. Therefore, it becomes the intention of this thesis to analyze whether the investment vehicle reduces the transaction costs, and subsequently, affect the performance of target firms.

3. Research gap

As the SWFs have gained more importance in the global financial market (Aguilera et al., 2016), more existing studies about SWFs have focused on how politics impact the investment strategies of SWFs (Knill et al., 2012b; Murtinu & Scalera, 2016), or have focused on the comparison of the performance between SWFs

(23)

investments and private investments (Bortolotti et al., 2015), or the determinants of SWFs investments (Johan et al., 2013; Bertoni & Lugo, 2015). Financial literature is widely studied about the use of investment vehicles. However, in international business literature, most studies still focus on direct investments. Murtinu and Scalera (2016) investigate the circumstances SWFs tend to use investment vehicles. However, there are only a few studies on how SWFs invest cross-border through vehicles is able to affect the performance of target firms.

As Murtinu and Scalera (2016) noted that future research should focus on the effect of investment vehicles on the performance of target firms. This thesis aims to study whether using the investment vehicles is able to affect the performance of target firms, and whether bilateral agreements and strategic industries is able to affect the performance when SWFs invest through vehicles.

The main research question is:

Does the target firm perform better when SWF uses investment vehicles?

3.1 Hypothesis development:

Previous studies have used the transaction cost theory to predict the best entry mode. The transaction costs arise from information asymmetry, extent studies showed the greater information asymmetry is, the greater transaction cost is. Researchers have tested the most applicable entry mode to minimize the information asymmetries, and thus to minimize the transaction costs (Hennart, 2007). For SWFs investments, similar to the MNEs investments, the performance of target firms is highly related to the

(24)

transaction costs (Rose, 2009). However, Existing studies of SWFs have focused on the direct cross-border investments, location or ownership boundary choice. there is little academic attention on SWFs cross-border investment through vehicles. This paper attempts to fill the gap, to investigate the information reduction ability of vehicles by analogizing the investment vehicles to the traditional entry modes. By comparing the information asymmetries SWFs potentially encounter when they use investment vehicles and when SWFs do not use vehicles.

The investment vehicles are classified into three categories: financial vehicle, corporate vehicle, and other SWF investment vehicle (Murtinu & Scalera, 2016). Firms or SWFs use different vehicle in different industries or for various purposes. As an example, SWFs use financial vehicles for financial objectives, and corporate vehicles to access sources, or use SWF investment vehicles for their political goals. For each vehicle the SWFs use, it is related to the industry of the target firm or for a particular purpose which other types vehicles or traditional entry modes cannot achieve. Thus, the first reason why I consider the investment vehicle can reduce information asymmetry is that the vehicles are related to the target firms or goals. The vehicles generally have more knowledge over the industry or the market. With such superiority, it is thereby advantageous for the cost reduction in integrating or accessing intangible knowledge. The second reason is that a vehicle can reduce foreign liabilities and avoid hostilities from the host country government (Murtinu & Scalera, 2016). As SWFs face hostilities from host countries that originate from the fears of political pursuits - the host country might issue laws or policies to constrain

(25)

the direct investments of SWFs from certain countries. In this case, SWFs invest through a vehicle from the host country or a third country which has better relations with the host country, could bypass the liabilities or constraints set by the host country. Therefore, using the investment vehicles could reduce the costs of overcome liabilities. With such understanding, it is assumed that greater information asymmetry leads to greater transaction cost, subsequently leads to worse performance of the target firms. In other words, if using the investment vehicles can reduce the transaction costs arising from information asymmetries and liabilities, then using the investment vehicles is able to improves the performance of target firms.

Hypothesis 1: The target firms perform better when SWFs invest cross-border

through vehicles.

Bilateral relations between home and host countries have two effects on the cross-border investments in previous studies: political, and economical effects. From the political perspective, relations between home and host countries affect the policies for foreign business and investors (Li and Vashchilko, 2010), existing studies show that the political instability is negatively related to FDI flows, and stable politics has a significant and positive relation with the FDI flows (Schneider and Frey, 1985; Loree and Guisinger, 1995). Political instabilities could be military conflicts, security alliance. One of the most recent example for political risks raised from military conflicts and security alliance is the Thaad System event. The South Korean government claimed that the purpose of installing the Thaad System is to prevent missiles from North Korea (The New York Times, 2017). On the opposite shore, the

(26)

Thaad System is considered as a huge threat by Chinese and Russian governments. Even though, the Chinese government has never officially issued any law or policy to constrain the South Korean business in the mainland, China. The South Korean business is still strongly impacted by this event. This has been observed especially for the retailing supermarket Lotte, which handed over its land to the South Korean government to install the Thaad System. The company has encountered a devastating fall in turnover in Chinese market. This was a result of Lotte’s boycotted by Chinese government as well as the Chinese customers. Since then, at least 23 Lotte stores in China have been shut down (CNN, 2017). Besides that, the Korean-pop industry, tourism industry, and the main pillars of South Korean economy - the media and entertainment industries were strongly impacted in Chinese market (CNN, 2017). The Thaad System event is one typical political risks raised from military conflicts without any official constraints. From the economic perspective, political risks could arise from trade embargoes, usually accompanied with laws or bills issued by the governments, the laws and bills could be issued by both sides governments related to the trade embargoes. One of the typical examples is the economic sanctions. The economic sanction might hurt firms from exporting country, such as in 2002 United States steel tariff forbade foreign steel producers from selling steels in U.S, the bill harmed the Chinese and Russian steel producers’ revenue seriously. Or the laws and bills might harm companies from countries that issue the bill, such as the Sanctions against China from 1989 to early 1990s. Companies from Europe and United States were forced to divest from mainland, China. In general, the political conflicts with or

(27)

without official laws, issued by home or host country affect the bilateral trade. SWFs’ nature determines their investments are never able to avoid the politics. In the SWFs case, Knill et al. (2009) showed that politics affect the SWFs investment strategy, the two-stage Cragg Model approach is used to prove that political relations affect where SWFs invest, and the political relations have a significant and negative relation with the location choice, but have no relation with the amount of investments. However, due to the opacity and involvement of politics, skepticism of SWFs investments would be amplified (Murtinu & Scalera, 2016). Some political leaders concerning the motives of SWFs might go beyond the financial pursuits, and it is hard to distinguish political pursuits from financial pursuits. Thus, the political pressure from the host country increases. The mighty political pursuits lead to fears of the unclear motives of SWFs. The host country government might hinder the SWFs investments from a certain industry, or issue stricter laws to restrict SWFs investments, which is done by the United States (Rose, 2008).

Constrains and military conflicts have a negative effect on the bilateral business between home and host countries. Contrary to countries with weaker relations, investors can anticipate either less confines or sometimes even favorable if they invest in a foreign country which have a better bilateral relation with their home countries. The bilateral agreement is regarded as a direct reflection of a stable bilateral political relation in this thesis due to the following frequent business movements. The stable relations then lead to a better investment and economic environments, furthermore, fewer conflicts between two countries are able to create more investment

(28)

opportunities (Li & Vashchilko, 2010). Meanwhile, the political stability also leads to a rapid economic growth, which attracts more foreign investments and cross-border trades (Schneider &Frey, 1985). Despite the stability, the other effect of bilateral relations – economic effect, also means liberal economic policies. By joining the international organizations such as World Trade Organization (WTO), the members not only have a lower tariff, but also have a more open market before they join. As the treaties require member countries to open up their markets, and lower their tariffs. Every member country has to apply the free trade commitment to all other members. This commitment also means less information asymmetry among the member countries. Similar to WTO, other trade agreements such as Preferential Trade Agreements (PTA) and Regional Trade Agreements (RTA) require their member countries to have some extent of the liberalization (Büthe and Milner, 2008). All these international institutions require their members to provide more information in several ways. Most international organizations (IOs) require their members to be more “visible” or transparent, part of them are even established for this purpose. Some IOs even gather and disseminate their members’ political information (Morrow, 1994). For example, WTO scrutinizes members’ economic policies (Büthe and Milner, 2008). PTAs are even more intensive than WTO. All these bilateral or multilateral institutions and agreements aim at reducing trade obstacles, information asymmetries, tariffs, and constraints among the member countries. To sum up, a bilateral agreement between home and host countries is able to positively moderate the performance of target firms when SWFs investing through vehicles. Therefore, this thesis expects:

(29)

Hypothesis 2: The performance of target firms is positively affected by the

bilateral agreements when SWFs investing through vehicles.

The host countries governments are often concerning about the SWFs investments due to the possibility of political pursuits. The fear of host country government increases as the SWFs invest in strategic industries. Consequently, the host country government might choose to hinder the SWFs from investing in strategic industry (Knill et al., 2012), because the strategic industries, such as financial institutions, energy, metals and metal products, post and telecommunications, usually relate to the country’s security and development policies. Since the government of host country concerns SWFs investment patterns are possibly stem from their sovereign governments political strategies, and due to the opacity and the size of SWFs, host countries might issue laws and constraints to prevent SWFs from acquiring firms operate in strategic industries. In the United States, the industry-specific laws are issued on SWFs for preventive measures (Keller, 2008). Constraints from the host countries cause greater foreign liabilities if SWFs directly invest in the target firm. If the target firms operate in the strategic industries, the greater liabilities are, the harder the direct investment to be completed. Even the direct investment is completed, SWFs have more costs to overcoming the liabilities and the constraints. Also, Murtinu & Scalera (2016) have found out that SWFs tend to use investment vehicles when target firms operate in the strategic industries, this result shows that the SWFs need vehicles to reduce either information asymmetry and/or liabilities in this circumstance. Thus, this thesis expects:

(30)

Hypothesis 3: The performance of target firms is positively affected by the

strategic industry when SWFs investing through vehicles.

Table 1 in shows the overall of hypotheses.

H1 The target firms perform better when SWFs invest cross-border through vehicles

H2 The performance of target firms is positively affected by bilateral agreements when SWFs investing through vehicles.

H3 The performance of target firms is positively affected by strategic industry when SWFs investing through vehicles.

4. Data and methodology

4.1 Data and sample

This thesis uses data from Murtinu and Scalera’s (2016) paper. And I select the data from their dataset by five steps: first, I select the “investments deal status” as “completed”. Second, as this thesis only focuses on the effect of using the investment vehicles, I consider the repeated investments under same SWFs as one investment, repetitive investments, namely subsequent investment from one SWF to the same target firms are deleted. The first investment strategies as the criterion. For a few investments, the acquiring companies are controlled by more than one parent SWFs, in this case, the parent SWF which locates in the same region with the acquirer company is selected. Third, I select categories of “majority”, “cross-border”, “vehicles”, “parent SWFs”, “target firms”, “acquiring years”, “target firm’s industries”, “SWF country code”, “target firm country code”, “target major sector”. Forth, Fernandes (2014) used two different profit measures, namely return on equity

(31)

(ROE) and return on assets (ROA), to analyze the impaction of SWFs investment on target firm’s performance. Thus I follow Fernandes’ method (2014) to choose ROE and ROA of the target firms of the first year after acquisition. Two missing values of ROE and ROA are deleted; this means that the two investments are ruled out when I test the hypotheses with ROE or ROA as the dependent variables. The performance of target firms, geographical distance, use of vehicles, size of acquired stakes, year of acquisition, foreign assets/total assets of SWFs are provided by Murtinu & Scalera (2016). Furthermore, I check bilateral agreements follow Murtinu & Scalera’s (2016) criteria, the data is obtained from the World Trade Organization website. There are 31 PTAs, and 291 RTA included (WTO, 2017). The cultural distance is obtained from Geert-Hofstede (Geert-hofstede.com, 2017), the corruption perceptive index is obtained from Transparency International (Transparency International, 2017), and the GDP data is obtained from the World Bank (World Bank, 2017). The final sample is composed of 140 investments, from 24 parent SWFs, and 16 countries, Targeting at 40 countries, 18 industries.

4.2 Variables and measures

4.2.1 Dependent variable

The performance data the thesis uses is collected by Murtinu & Scalera (2016), the performance of the year after acquisition is selected, and this thesis chooses return

(32)

on assets (ROA), return on equity (ROE) following Fernandes (2014) as the dependent variables.

4.2.2 Independent variables

The independent variable of this thesis is the use of investment vehicles. This is a dummy variable that equals to one if the SWFs invest cross-border through vehicles, and equals to zero if SWFs do not.

4.2.3 Moderators

- Bilateral agreement

The World Trade Organization (WTO) has 164 member countries till 29 July 2016 (WTO, 2017), almost all SWFs and target firms come from these member countries, thus I exclude the WTO as a bilateral agreement. Home and host countries are considered have bilateral agreements if they satisfy following conditions: (1) home and host countries have either the preference trade agreements (PTA) or regional trade agreements (RTA) which can be found on the WTO website (Murtinu & Scalera, 2016); (2) the European Union (EU) members, due to the EU treaties (eropa.eu, 2017), the EU members are considered have reciprocal bilateral agreements. The bilateral agreement is a dummy variable, which equals to one when the SWF home country and target firm’s country have either a PTA, or RTA, or the two countries are EU members.

(33)

This thesis follows Murtinu & Scalera (2016) to use a criterion which is suggested by Keller (2008) and Drezner (2008), and combine the final investments sample, to identify five industries as strategic industries, namely financial institutions, energy, metals and metal products, post and telecommunications, mining and transportation. The strategy industry is a dummy variable, which equals to one if the target firm operates in a strategic industry, and equals to zero if the target firm does not operate in a strategic industry.

4.2.4 Control variables

In this thesis, seven variables are controlled. I select the three control variables according to Murtinu & Scalera (2016). First, the SWFs thrived and have become the important global players during the financial crisis 2008. The financial crisis had a huge impact on the Western market, I follow Murtinu & Scalera (2016) and Bortolotti et al. (2015), added the Year Trend as a control variable which is a dummy variable that equals to one if the acquisition happens between 2008-2013, and equals to zero if the acquisition happens before 2008 or after 2013. Second, the Foreign assets/Total

assets, I select this variable as a control variable because it shows the degree of the

fund internationalization, and a higher proportion of the foreign assets means more international business experience, previous studies showed that cross-border investment experience often influence on the strategy making (Shimizu, Hitt, and Vaidyanath, 2004), and also leads to a different performance. However, it is unclear to what extent the multinational experience is able to affect the performance of target

(34)

firms. Third, because the SWFs are owned and/or controlled by the sovereign governments, the corruption of SWFs home countries has a vital effect on the investments. As “the corruption shapes the development of political and financial institutions” (Murtinu & Scalera, 2016, p.8). I control the corruption of SWFs home countries, more specifically, the corruption perceptive index (CPI) of the year after acquisition are controlled (Transparency international, 2017). Fourth, the results of Bortolotti et al. (2009) and Fernandes (2014) are vary from each other on how the size of acquired stakes affects the performance of target firms, Bortolotti et al.’s (2009) result showed that the size of acquired stakes is negatively associated with the performance of target firms. However, Fernandes’ (2014) result stated that the size of acquired stakes is positively associated with the performance of target firms. I consider the size of acquired stakes is able to affect the performance of target firms, but the relation between the acquiring size and performance has not reach an agreement yet. Thus, I select the size of acquired stakes as the fourth control variable. Fifth, I follow Bortolotti et al. (2015) to control the Gross Domestic Product (GDP) of both SWFs’ countries and target firms’ countries on the year after acquisitions, because the GDP reflects the financial situation of SWF home country, the data is collected from the World Bank. Finally, as many cross-border investments always consider the distances (Hofstede, 1980), I select both cultural distance and

geographical distance as control variables. The data of cultural distance is collected

from Geert-Hofstede (Geert-Hofstede, 2017). The geographical distance is identified as the world regions which SWFs and target firms locate in. This thesis identifies the

(35)

geographical distance as a dummy variable, which equals to one if the SWF home country and target firm country are located in different world regions; and equals to zero if the SWF home country and target firm country locate in the same region. Following Murtinu & Scalera (2016), the world regions are classified into six categories: “MENA”, “Europe”, “Asia-Pacific”, “Latin America”, “North America”, and “South Africa”. Table 2 shows the overall of variables.

(36)

Table 2

Variable Definition Source

Dependent variable

Performance of target firm I measure the performance of target firms with Return on Assets (ROA), Return on Equity (ROE) at the year after acquisitions.

Dataset from Murtinu & Scalera (2016) Independent variable

Use of vehicle Dummy that equals to 1 if SWFs invest through vehicles, and 0 if SWFs invest without vehicles

Murtinu & Scalera (2016)

Moderator

Industry Dummy that equals to 1 if the target firm operates in a strategic industry, and 0 if target firm operates in other industries

Keller (2008), Drezner (2008) Bilateral agreement Dummy that equals to 1 that the SWF country and the target

country have PTA, or RTA, or the two countries are EU members; and 0 if they does not have any bilateral agreements

World Trade Organization

Control variables

Year trends Dummy that equals to 1 if the acquisition happens between 2008-2013, and 0 if the acquisition happens before 2008 or later than 2013

Foreign assets/Total Assets Fund-level ratio between total assets and foreign assets. Murtinu & Scalera (2016)

Size of acquired stakes Dummy that equals to 1 if the SWF acquires majority stakes, and to 0 if the SWF acquires minority stakes

GDP (SWFs home countries) The Gross Domestic Product (current US $ billion) The World Bank GDP (target firms countries) The Gross Domestic Product (current US $ billion) The World Bank Cultural distance The data cultural distance is collected from Geert Hofstede Geert Hofstede Geographical distance Dummy that equals to 1 if the SWF and the target firms come

from different world regions, and 0 if the two countries come from same region

Corruption perceptive index The data is collected from Transparency international Transparency international

(37)

4.3 Methodology

This thesis uses quantitative research design to test the research question “Does the target firms perform better when SWFs invest cross-border through vehicles?”, and the hypotheses presented in the section 3.1, how the bilateral agreements and strategic industries moderate the performance of target firms. These analyses use the IBM SPSS Statistic Program, and I follow Fernandes (2014) to perform several general linear regressions. First, the relation between the control variables, moderators and the dependent variables are tested, more specifically, I regress the year trends, foreign assets/total assets, size of acquired stakes, GDP of SWFs home countries, GDP of target firms’ countries, cultural distance, geographical distance, and corruption perceptions (CONTROLS), bilateral agreements, and strategic industry (MODERATORS) with ROE or ROA (DEPENDENT VARIABLE). Second, I add the dummy variable - vehicle (INDEPENDENT VARIABLE) in model 2, to test the effect of independent variable on the dependent variables, which is the hypothesis 1. Third, to test the effect of moderators, I create two interactions (MODERATORS * INDEPENDENT VARIABLE). Two interactions are (1) vehicle * bilateral agreement (model 3 for hypothesis 2) and (2) vehicle * strategic industry (model 4 for hypothesis 3). The linear model is specified as follows:

Performance (ROA or ROE) =

𝛼₀+ 𝛼₁𝑉𝑒ℎ𝑖𝑐𝑙𝑒 + 𝛼₂𝑉𝑒ℎ𝑖𝑐𝑙𝑒 ∗ 𝐵𝑖𝑙𝑎𝑡𝑒𝑟𝑎𝑙 𝐴𝑔𝑟𝑒𝑒𝑚𝑒𝑛𝑡 (𝑜𝑟 𝛼₃𝑉𝑒ℎ𝑖𝑐𝑙𝑒 ∗ 𝑆𝑡𝑟𝑎𝑡𝑒𝑔𝑖𝑐 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦) + 𝛽₀𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒

(38)

5. Results

5.1 Descriptive statistics

I provide the descriptive statistics in tables 3. As shown in table 3, the frequencies of ROA and ROE are both 140. In the final sample, 18.6% of the investments use investment vehicles; 49.3% of the target firms are operated in the strategic industries; 43.6% of the investments have bilateral agreements; 52.9% of the investments cross the world regions; only 2.9% of investments acquire major stakes; 39.3% of investments happened between 2008 - 2013.

The cultural distance between SWF country and target firm’s country ranges from 0.107 to 7.977, with an average at 2.034; the Foreign assets/Total assets ranges from 10% to 100%, with an average at 83.68%; the corruption perceptive index of SWFs home countries ranges from 21 to 87, with an average at 77.76. The GDP differences of SWFs countries and target firms’ countries are huge. The standard deviations of GDP of SWFs’ countries and target firms’ countries are 1952.062 and 3567.51 separately. The smallest GDP of SWF’s and target firm’s country is Libya and Cayman Island separately; the biggest GDP of SWF’s and target firm’s countries is the United States.

(39)

Descriptive Statistics

N Minimum Maximum Mean Std. Deviation

ROE 140 -213.070 72.110 -.134 38.255 ROA 140 -55.020 25.960 1.771 11.798 Vehicle 140 .0 1.0 .186 .390 Strategic industry 140 .0 1.0 .493 .502 Bilateral agreements 140 .0 1.0 .436 .498 region 140 .0 1.0 .529 .501 Cultural distance 140 .107 7.997 2.024 1.378 Majority 140 .0 1.0 .029 .167 Acquiring year 140 .0 1.0 .393 .490 Foreign/Total 140 10.00% 100.00% 83.684% 22.088% corruption 140 21.0 87.0 77.757 14.925 GDP (SWF) 140 63.028 14477.635 791.992 1952.062 GDP (target) 140 3.207 16691.517 2658.644 3567.51 Valid N (listwise) 140 5.2 Correlation

I provide an overview of the Pearson correlation in table 4. The correlation matrix includes the dependent variables, the independent variable, the moderators, and the control variables for these regression analyses. The matrix shows that two dependent variables (ROA, ROE) are highly correlated, because ROA and ROE are different measurements of the target firms’ performance, the high correlation between ROA and ROE is comprehensible (Kabajeh et al., 2012). And the two dependent variables are tested separately, the high correlation does not influence the whole analysis. Besides the two dependent variables, none of the variables and moderators are highly correlated (>0.7). The independent variable - vehicle has a significant and negative correlation with one of the dependent variables (ROA). The data is sufficient

(40)

for SPSS linear analysis. Then I ran the Variance inflation factor (VIF), the result confirms the absence of multicollinearity concerns: the mean VIF is 1.47, and the biggest VIF is 2.068 (when dependent variable is ROE); the mean VIF is 1.47, and the biggest VIF is 2.079 (when dependent variable is ROA).

(41)

Table 4 Correlation metrix Correlations 1 2 3 4 5 6 7 8 9 10 11 12 13 1. ROE 1 2. ROA .821** 1 3. Vehicle -.110 -.222** 1 4. Strategic industry -.050 -.120 -.030 1 5. Bilateral agreement -.041 .048 -.012 .113 1 6. Geographical distance -.148 -.218** .157 -.071 -.584** 1 7. Cultural distance -.182* -.212* .089 -.132 -.299** .374** 1 8. Majority .081 .034 .249** .088 .109 -.096 -.003 1 9. Time trends -.041 -.100 .067 -.003 -.205* .144 .053 -.050 1

10. Foreign asset/Total asset .157 .146 .048 .076 .011 -.017 -.057 -.159 -.096 1

11. Corruption perception index .043 .107 -.138 .159 .393** -.392** -.044 .098 -.187* .363** 1

12. GDP (SWF) -.068 -.017 -.130 -.113 -.061 .128 .308** -.020 .159 -.509** -.420** 1

13. GDP (target firm) -.255** -.185* .063 .063 .027 .130 -.018 -.055 .033 -.119 -.130 .048 1

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

(42)

5.3 Regression analysis

As it has been done by Fernandes (2014), several linear regression analyses are carried out for this analysis. In table 5, I provide the coefficient matrix of dependent variables (ROE), independent variable, control variables, moderators, and interactions. Table 6 presents an overview of the result of four models. In model 1, the control variables (geographical distance, cultural distance, majority, time trends, foreign asset/total asset, corruption perception index, GDP of SWF country, GDP of target firm’s country) and the moderators (bilateral agreement and strategic industry) are entered. This model solely including control variables and moderators is statistically significant with dependent variable (ROE), F (2.475) =0.010, p<0.05. Three control variables and one moderator contribute to the model. In model 2, I add the independent variable (vehicle) to test hypothesis 1. The model 2 is significant, F (2.365) = 0.011, p<0.05. However, the coefficient matrix shows that the independent variable has a negative but insignificant effect on the dependent variable (ROE), thus, hypothesis 1 is rejected. Model 3 tests the hypothesis 2, I add the interaction of vehicle and moderator (bilateral agreement), the result shows that the model 3 is statistically significant, F (2.175) = 0.017, p<0.05. However, the table 5 shows the interaction has a positive and statistically insignificant effect on the dependent variable (ROE), thus, the hypothesis 2 is rejected. Model 4 tests the hypothesis 3, I add the interaction of independent variable and moderator (strategic industry), table 6 shows the model 4 is statistically significant, F (2.465) =0.006, p<0.01, and table 5

(43)

shows that the interaction has a statistically significant and positive effect on the dependent variable (ROE), hypothesis 3 is thus supported.

Following Fernandes’s methodology (2014), I use ROA as the second dependent variable to re-check the result of regression. In table 8, I provide the coefficient matrix of dependent variables (ROA), independent variable, control variables, moderators, and interactions. Table 8 presents an overview of the result of four models. In model 1, the control variables and the moderators are entered. This model solely includes control variables and moderators which are statistically significant with dependent variable (ROA), F (2.777) =0.004, p<0.01. Four control variables and one moderator contribute to the model. In model 2, I add the independent variable (vehicle) to test hypothesis 1. The model is significant, F (2.978) = 0.001, p<0.01. The table 7 shows a different result from the first regression, it shows that the independent variable has a statistically significant and negative effect on the dependent variable, thus, the hypothesis 1 is rejected. Model 3 tests the hypothesis 2, I add the interaction of vehicle and moderator (bilateral agreement), the result shows the model 3 is statistically significant, F (2.805) = 0.002, p<0.01. But the table 7 shows the interaction has a positive and statistically insignificant effect on the dependent variable (ROA), the hypothesis 2 is rejected. Model 4 tests the hypothesis 3, I add the interaction of independent variable and moderator (strategic industry), table 8 shows the model 4 is statistically significant, F (3.387) =0.000, p<0.01, and table 7 shows the interaction has a statistically significant had positive effect on the dependent

(44)

variable (ROA), the hypothesis 3 is supported. Table 9 presents the overview the

(45)

Table 5

Dependent variable: ROE Model 1: controls, moderators Model 2: hypothesis 1 Model 3: hypothesis 2 Model 4:hypothesis 3

Control variable Beta t Sig. Beta t Sig. Beta t Sig. Beta t Sig.

Geographical distance -0.159 -1.463 0.146 -0.145 -1.327 0.187 -0.142 -1.293 0.198 -0.142 -1.313 0.191

Cultural distance -0.217 -2.238 0.027** -0.199 -2.030 0.044** -0.201 -2.042 0.043** -0.158 -1.577 0.117

Majority 0.115 1.378 0.171 0.144 1.644 0.103 0.128 1.372 0.172 0.125 1.426 0.156

Time trends -0.035 -0.422 0.673 -0.026 -0.312 0.755 -0.035 -0.408 0.684 -0.030 -0.358 0.721

Foreign asset/total asset 0.203 2.040 0.043** 0.212 2.122 0.036** 0.211 2.110 0.037** 0.186 1.861 0.065* Corruption perception index -0.010 -0.093 0.926 -0.041 -0.371 0.711 -0.052 -0.455 0.650 -0.025 -0.227 0.821

GDP (SWF) 0.116 1.083 0.281 0.087 0.789 0.423 0.083 0.748 0.456 0.060 0.541 0.590 GDP (target firm) -0.204 -2.450 0.016** -0.200 -2.401 0.018** -0.205 -2.434 0.016** -0.181 -2.164 0.032** Moderator -Bilateral agreement -0.196 -1.833 0.069* -0.175 -1.605 0.111 -0.170 -1.550 0.124 -0.170 -1.575 0.118 Strategic industry -0.069 -0.831 0.407 -0.073 -0.876 0.382 -0.071 -0.853 0.395 -0.071 -0.866 0.388 Independent variable vehicle -0.100 -1.106 0.271 -0.096 -1.059 0.292 -0.093 -1.037 0.302 Interaction

Vehicle * bilateral agreement 0.045 0.493 0.623

Vehicle * strategic industry 0.151 1.770 0.079*

*Significant at P-value < 0.1 ** Significant at P-value < 0.05 *** Significant at P-value < 0.01

(46)

Table 6

Model F-test Sig.

1 2.475 0.010** 2 – H1 2.365 0.011** 3 – H2 2.175 0.017** 4 – H3 2.465 0.006*** *Significant at P-value < 0.1 ** Significant at P-value < 0.05 *** Significant at P-value < 0.01

(47)

Table 7

Dependent variable: ROA Model 1: controls, moderators Model 2: hypothesis 1 Model 3: hypothesis 2 Model 4:hypothesis 3

Control variable Beta t Sig. Beta t Sig. Beta t Sig. Beta t Sig.

Geographical distance -0.179 -1.669 0.098* -0.154 -1.441 0.152 -0.148 -1.383 0.169 -0.150 -1.425 0.154

Cultural distance -0.268 -3.786 0.006** -0.235 -2.447 0.016* -0.239 -2.482 0.014** -0.178 -1.835 0.069*

Majority 0.065 0.786 0.433 0.118 1.373 0.172 0.088 0.961 0.338 0.091 1.076 284

Time trends -.086 -1.039 0.301 -0.069 -0.844 0.400 -0.086 -1.028 0.306 -0.074 -0.924 0.357

Foreign asset/total asset 0.211 2.138 0.034** 0.226 2.319 0.022** 0.225 2.308 0.023** 0.191 1.975 0.050*

Corruption perception index 0.097 0.908 0.366 -0.040 -0.363 0.718 -0.020 0.179 0.858 0.062 0.579 0.563

GDP (SWF) 0.230 2.167 0.032** 0.177 1.636 0.104 0.169 1.560 0.121 0.139 1.300 0.196 GDP (target firm) -0.122 -1.475 0.143 -0.114 -1.401 0.164 -0.123 -1.499 0.136 -0.087 -1.080 0.282 Moderator Bilateral agreement -0.167 -1.574 0.118 -0.127 -1.194 0.235 -0.118 -1.104 0.272 -0.121 -1.158 0.249 Strategic industry -0.155 -1.886 0.062* -0.162 -1.993 0.048** -0.159 -1.952 0.053* -0.160 -2.009 0.047** Independent variable vehicle -0.183 -2.068 0.041** -0.176 -1.982 0.050* -0.173 -1.998 0.048** Interaction

Vehicle * bilateral agreement 0.085 0.960 0.339

Vehicle * strategic industry 0.210 2.547 0.012**

*Significant at P-value < 0.1 ** Significant at P-value < 0.05 *** Significant at P-value < 0.01

(48)

Table 8

Model F – test Sig.

1 2.777 0.004*** 2 – H1 2.978 0.001*** 3 – H2 2.805 0.002*** 4 – H3 3.387 0.000*** *Significant at P-value < 0.1 ** Significant at P-value < 0.05 *** Significant at P-value < 0.01 Table 9

H1 The target firms perform better when SWFs invest cross-border through vehicles. Rejected H2 The performance of target firms is positively affected by bilateral agreements

when SWFs investing through vehicles

Rejected

H3 The performance of target firms is positively affected by the strategic industry when SWFs investing through vehicles.

(49)

6. Discussion and conclusion

The SWFs have been important players in the global financial market for decades. Due to their tremendous monetary value, and political ownership, SWFs have attracted both political and academic attentions (Aguilera et al., 2016). Previous studies have focused on the strategies of SWFs (e.g. Alhashel, 2015; Aguilera et al., 2016), the performance of the target firms (e.g. Fernandes, 2014; Bortolotti et al., 2009, 2015). But same as the traditional IB literature, there are a few studies that investigate SWFs investing through vehicles. Murtinu & Scalera (2016) started the topic of SWFs invest cross-border through vehicles. This thesis follows Murtinu & Scalera (2016) to investigate the effect of SWFs invest through vehicle on the performance of target firms. I combine the transaction cost theory, the previous studies on the effect of political risks (Rose, 2009; Li & Vashchilko, 2010), and studies of the effect of strategic industry (Knill et al., 2012) on the performance of target firms. To investigate when SWFs invest cross-border through vehicles, how do political risk, namely the bilateral agreement, and industry which the target firms operate in, affect the performance of target firms. Due to the opacity and political involvement of SWFs, their cross-border investments always face the hostilities and constraints from the host country, which lead to the information asymmetry. Those obstacles from the host country worsen the performance of target firms. This thesis assumes that investing cross-border through vehicles reduces the transaction costs, therefore increasing the performance of target firms. Meanwhile, this thesis consider

Sovereign wealth funds : the effect of investment vehicles on the performance of target firm