Are Sovereign Wealth Funds political playgrounds? : a research into the impact of political influence on the performance of Sovereign Wealth Funds

1

Are Sovereign Wealth Funds political playgrounds?

A research into the impact of political influence on the performance of Sovereign Wealth Funds

Name: Tyrone Pater

Student Number: 11421053

MSc. In Business Administration – International Management Master Thesis

Supervisor: Dr. V.G. Scalera

2

Statement of Originality

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

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

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

3

Inhoud

2.1 Sovereign wealth funds ... 10

2.1.1 Size, growth and Identification of sovereign wealth funds ... 10

2.1.2 Reasons to found a sovereign wealth fund ... 13

2.2 Sovereign Wealth Fund Management and Opacity... 14

2.2.1 Management of SWFs ... 14

2.2.2 SWF opacity ... 15

2.3 Political influence, strategy and performance of sovereign wealth funds... 16

2.3.1 Political connections and SWFs ... 16

2.3.2 SWF Strategy ... 18

2.3.3 SWF Strategic Industries ... 19

2.3.4 SWF performance ... 20

2.4 SWF Location Choices... 22

2.5 Research Gap... 24

3 Conceptual Framework and Hypotheses Development ... 25

3.1 Hypotheses Development ... 25

3.1.1 Political influence and performance ... 25

3.1.2 Strategic industry ... 27

3.1.3 Location Choice ... 28

3.1.4 Transparency ... 29

3.2 Conceptual Model ... 31

4. Methodology ... 32

4.1 Sample & Data ... 32

4.2 Variables... 35

4.2.1 Dependent Variable ... 35

4.2.2 Independent Variable ... 35

4.2.3 Moderating Variables ... 36

4

4.4 Procedure ... 41

5. Results ... 42

5.1 Normality and Assumptions Check ... 42

5.2 Descriptive Statistics ... 43

5.3 Correlations ... 44

5.4 Regression Analysis ... 47

5.4.1 Regression analysis dependent variable performance ROA ... 47

5.4.2. Regression analysis dependent variable performance NSR ... 51

6. Discussion ... 55

6.1 Hypotheses and Results Discussion ... 55

6.2 Theoretical and Practical Implications ... 60

6.3 Limitation and Suggestions for Further Research ... 61

7. Conclusion ... 63

Acknowledgements ... 64

8. References ... 65

9. Appendix ... 69

Raw Data Example ... 69

Dataset Codebook ... 70

List of Figures

List of Tables

Table 2: SWFs and number of investments………...34

Table 3: Overview of variables………..40

Table 4: Descriptive statistics………44

Table 5: Correlation Matrix………...46

Table 6: Regression results………50

Table 7: Regression results………54

5

Abstract

Sovereign wealth funds (SWFs) have become major players, controlling assets of around six trillion US dollars. Several authors have written about SWF issues regarding political influence, transparency, location choice and strategy. This work aims at answering the following questions: What is the influence of politicised SWFs on target firm performance? And is this relation moderated by strategic industry, location choice and transparency? This research aims to answer these questions by formulating several hypotheses and tests these by submitting a sample to quantitative analysis in SPSS. The sample consists of 139 investments made by nineteen different SWFs in 34 different countries between 1997 and 2013. Most of the collected data is secondary data. The results conclude that there is a negative influence on performance when an SWF is politicised. There were no significant moderation effects found. Therefore, it is safe to conclude that the main contribution of this thesis is to confirm that politicised SWFs tend to perform worse than non-politicised SWFs, as was already suspected by several authors.

Additionally, the results in this thesis show that more research is necessary to be able to state if there are moderation effects between the mentioned variables or not.

Key Words: Sovereign wealth funds, firm performance, cross border acquisitions, strategic

6

1. Introduction

Over the past years, sovereign wealth funds (SWFs) have become major players in an increasingly integrated world. SWFs control around six trillion US dollars in assets and are the fastest growing players in the market of global investors, it is therefore not preposterous to assume that SWFs will be relatively influential in the international investing world for years to come (Aguilera et al. 2016; Megginson et al. 2015; Bernstein et al. 2013).

SWFs owe their naming to Andrew Rozanov, who first conned the term in 2005 (Rozanov, 2005). These funds are always government owned and are founded for different reasons. One of the main reasons for countries to set up an SWF is after the discovery of natural resources, often oil. Extensive spending of the income from these resources could help the economy in the short-term, but be harmful in the long run. The most written about example of mismanagement of these resources is the so called ‘Dutch disease’, about which later more. SWFs help countries mitigate the ‘Dutch disease’ by investing proceeds from commodities in domestic or foreign firms, in that way diversifying the portfolio and spreading risks (Aguilera et al. 2016; Johan et al. 2013; Murtinu & Scalera, 2016;).

As mentioned before, SWFs are government owned and controlled, this makes these funds interesting to analyse for several reasons. Recently, scholars have written about the active influence of politicians in these funds and on the investment tendencies of SWFs, often

concluding that politicians have a tendency to use SWFs to fuel their own short-term agenda over more sensible long-term investments (Bernstein et al. 2013). Additionally, questions have also been raised about what the consequences of SWFs investing in foreign assets could mean for relations between countries (Megginson et al. 2016). Another relevant topic of study

7 transparency (Bernstein et al. 2013; Megginson et al. 2015; Murtinu & Scalera, 2016). Overall, many of the questions and critiques surrounding SWFs seem to be politically oriented

(Megginson et al. 2015).

Internal governance and leadership structure of SWFs is interesting for several reasons. First, Megginson et al. (2015) note that most of the large SWFs are controlled by very small staffs, while controlling very large amounts of assets. Additionally, throughout the literature it is suggested that when politicians are influential within SWFs, the foreign investments made target key industries or countries. This, in turn, might generate hostility in the target country (Murtinu & Scalera, 2016). Lastly, Bernstein et al. (2013) find that politician-led SWFs tend to invest domestically and short-term oriented, instead of looking for long-term profit maximisation, which is often the case when SWFs are led by external managers.

Considering the increased role of SWFs in the investing world, and the increasing amount of questions considering political influences within and on the policies of these funds, more research on this topic is warranted. This paper will research what the influence on performance of SWFs is when SWFs are led by politicians, as opposed to SWFs led by external managers. Additionally, this work aims to research whether investing in strategic industries, location choices of investments or transparency issues moderate the main relationship between political influence and performance. Formulated as questions; What is the influence of politicised SWFs on target firm performance? And is this relation moderated by strategic industry, location choice and transparency? As already described by Bernstein et al. (2013), funds heavily influenced by politicians will most likely have different, more strategic goals and might possibly perform worse. These funds tend to invest more domestically, in industries that have a high price-to-earnings ratio but experience a drop in these ratios immediately after. Furthermore, Bernstein et

8 al. (2013) seem to hint that politicians face political pressure that causes them to strategically invest in, often domestic, industries that are underperforming.

By researching the impact on performance when SWFs are led by politicians a gap in the existing literature is bridged. Bernstein et al. (2013) already did some expeditionary work into this topic, but further analysis is required. Bernstein et al. (2013) did not include data from the largest SWF, Norway’s Government Pension Fund – Global. Additionally, their research only uses data up until 2007. By using newer datasets and including more SWFs, new conclusions can be drawn when considering performance and focus of SWFs. Furthermore, several scholars have researched SWF investment behaviour, specifically on the why and how SWFs make the

investments they do (Bortolotti et al. 2010; Bortolotti et al. 2015; Johan et al. 2013; Keller 2008). Other scholars have reported about the political and strategical issues that surround SWFs

(Bagnall & Truman, 2013; Domadenik, 2016; Knill et al. 2011; Solji & Tham, 2017). This research contributes to both of these streams of SWF research by providing results of SWF performance, which could help strengthen statements or findings made by other researchers. The practical relevance of knowing what impact influence of politicians on SWF performance is could be very relevant for managers and governments alike. If it is shown that investments championed by politicians have significantly lower returns and are not focused on profit maximisation, this could mean many millions of US dollars in lost assets. Therefore, answering this question could help governments and policy-makers manage their SWFs better. Another practical implication may be the creation of awareness. If it is proven that SWFs controlled by politicians perform worse, it may be that these politicians take a step back in decision-making. It could also create more hostility in target firm top management teams, knowing that there is a larger chance of worse performance for the company when acquired by

9 an SWF.

This paper will contain the following structure. First, an extensive literature review will be conducted in which the key concepts and articles will be explained. Furthermore, this section will contain the research question and describe the research gap in more detail. Second, a

theoretical framework will be developed and several hypotheses will be proposed and explained. Afterwards, a section will be dedicated to the method this paper will follow, justifying the research design and methodological choices. The fourth section will be the results section, in which the results will be presented. After the results section, a discussion chapter will be added to discuss the results in further detail, describe theoretical and practical relevance, and describe limitations and recommendations for further research. Lastly, a conclusion will be included to summarize the most important conclusions of this research.

10

2. Literature Review

2.1 Sovereign wealth funds

Among scholars and throughout the literature available on SWFs there is no conclusive and coherent definition of what an SWF exactly constitutes. Perhaps the most straightforward definition is provided by Bertoni and Lugo (2014), who use the following definition:

‘’Government-owned investment vehicles that manage portfolios including foreign financial assets’’. Most definitions of SWFs have in common that they suggest that these funds are

state-owned investment funds that make investment, both domestically and internationally as well as short-term and long-term, in search of commercial returns (Megginson et al. 2015).

There is also much debate over which funds should be included as SWFs, this is in part because not many funds have disclosed key organizational information and are thus not very transparent, and in part because definitions of SWFs vary (Bernstein et al. 2013; Epstein & Rose, 2009; Megginson et al. 2015). An example of this debate would be sovereign pension reserve funds (SPRFs), these funds are set up by the government and receive their cash inflow directly from the government (Murtinu & Scalera, 2016). Therefore, some scholars would include these funds as SWFs, while others would not. In this paper SPRFs will be included as SWFs.

2.1.1 Size, growth and Identification of sovereign wealth funds

There are roughly five types of SWFs that are identified. The first type is the Stabilisation

fund, this type of SWF has as primary objective to protect the budget and the economy in the

case of severe price swings in commodities. Saving funds for future generations are the second type of SWF that is identified, these funds usually appear when countries discover a source of oil or natural gas which can be exploited for significant gains in the short term. The fund is tasked with diversifying the portfolio of non-renewable assets to prevent potentially harmful effects of

11 spending the assets too fast which leads to long-term problems. The third type of SWF is the

reserve investment corporation, this type of SWF manages reserve assets and are typically

established to increase returns on said reserves. Development funds are the fourth type of SWF, these funds are typically tasked with funding projects that improve the economic growth of a country. Finally, the fifth type of SWF is the contingent pension reserve fund, which provides a cover for any deficits regarding pensions on the governments’ balance sheet (Allen & Caruana, 2008).

Concerning the origin of SWFs, most researchers distinguish two specific types of wealth SWFs are founded upon. For most of the large SWFs oil is the main source of wealth, some smaller SWFs find their wealth in other commodities like diamonds. The other category is referred to as non-commodity based wealth, these funds find their wealth in, for example, trade surpluses or the sale of state-owned enterprises (Bernstein et al. 2013). The dominance of oil related SWFs is something worth mentioning, of the 33 funds that meet the classifications of a SWF as put forward by the Sovereign Investment Laboratory, 21 are oil related (Megginson et al. 2015).

Today, SWFs control about six trillion US dollars in assets (Aguilera et al. 2016). The industry is seeing a rapid growth, considering that the total size of SWFs was five trillion US dollars in 2013, 2.5 trillion US dollars in 2009 (Epstein & Rose, 2009), and only 500 billion US dollars in 1990 (Bernstein et al. 2013). The top ten largest SWFs control almost seventy per cent of the total industry size. The largest SWF is Norway’s Government Pension Fund, which

controls well over 850 billion US dollars in assets. Table 1 (below) gives an overview of the ten largest SWFs in the world and where they are located, as well as the origin of their wealth.

12 Table 1: (Megginson et al. 2015) An overview of the ten largest SWFs in the world.

As mentioned, SWFs have been growing rapidly over the past few decades. There are several circumstances that caused the industry to grow as fast as it did. It is no secret that most of the large SWFs are based upon oil benefits. Therefore, the first factor that played a major role in SWF growth is the rise in the price of oil since 1998, which has since grown from ten US dollars a barrel to 148 dollars before stabilizing between ninety and 110 dollars during the years

between 2010 and 2014. Although the price of oil has decreased to around 55 US dollars a barrel in recent years the sector still seems to be growing exponentially (Bernstein et al. 2013;

Megginson et al. 2015). The other important factor that has contributed to growth of assets in the SWF sector is the accumulation of official reserves, often foreign and US dollar denominated in nature, by central banks as a response to the East-Asian financial crisis in 1997-1998.

Governments have saved up increasingly immense amounts of foreign exchange reserve

holdings over the past twenty years. This has led many governments to transfer wealth to SWFs to search for an increase in revenue while still dealing in US dollars (Megginson et al. 2015).

13

2.1.2 Reasons to found a sovereign wealth fund

We have already identified five different types of SWFs: Stabilisation funds, Saving funds for future generations, reserve investment corporation, Development funds and contingent pension reserve funds (Allen & Caruana, 2008). Most SWFs are the evolved form of stabilisation funds, which were the dominant type of established fund during the 1980s and 1990s, because of the success of the Chilean Social and Economic Stabilisation Fund. This fund displayed a

significant amount of the traits modern SWFs possess and benefited greatly from an independent board that was committed to setting and achieving targets as well as minimizing political

influence and thus public spending, which had plagued funds before the 1980s (Megginson et al. 2015). Due to the success of the Chilean fund the World bank advised other countries to follow this example, which caused the fact that most SWFs started out as stabilisation funds. The primary goal of the 1980s and 1990s stabilisation funds was to protect the budget and the economy in the case of severe price swings in commodities, in the last fifteen years or so SWFs have also focused on investing abroad to diversify the portfolio and attain financial returns (balding, 2012; Megginson et al. 2015). According to Megginson et al. (2015), most SWFs that were established in the 2000s were founded with the sole purpose of being an SWF, instead of the old idea of being a stabilisation fund first. They do, however, recognize that for many SWFs stabilisation remains one of the mandates.

Another reason that a significant number of scholars mention for countries to set up SWFs is to prevent the so called ‘’Dutch disease’’. In the 1960s The Netherlands unearthed a significant gas deposit, the subsequent mismanagement of the revenues that came out of this deposit caused long-term damage to other, non-commodity, sectors of the Dutch economy. This is because the Dutch government chose to invest large amounts of the revenues in the own

14 economy, in this way stimulating it short-term, but hurting the economy long-term. SWFs

mitigate this effect by having a mostly foreign focus, which causes SWFs to diversify portfolios and invest commodity based revenues in more tangible assets abroad. This in turn ensures that the local currency does not appreciate due to the commodity based revenues (Bernstein et al. 2013; Epstein & Rose, 2009 Megginson et al. 2015; Murtinu & Scalera, 2016).

2.2 Sovereign Wealth Fund Management and Opacity

2.2.1 Management of SWFs

How SWFs are managed is an interesting topic of discussion, as well as one that is hard to exactly pinpoint because of the transparency issues that surround many SWFs. One of the most important facts about SWF management is how small the staff of said funds tend to be. For example, despite managing assets worth over 2.28 trillion US dollars, Norway’s GPFG, China’s CIC, and Abu Dhabi’s ADIA collectively have fewer than 3000 employees. When that number

of employees is compared with a privately-owned investment agency of comparable size, those agencies tend to have anywhere around 40,000 employees (Megginson et al. 2015).

The relatively small number of employees SWFs employ have two major implications when considering SWF operations and management. First, many SWFs outsource a significant amount of investing and overseeing work to external managers. The AIDA, for example, sources about two thirds of total operations out to external managers. As an opposite to this, the GPFG manages about 95 per cent of operations in-house (Al-Kharusi et al. 2014; Megginson et al. 2015;). Second, SWFs have limited capabilities of directly influencing the corporate governance of their target companies. At any point, many of the larger funds have active investments in several thousand different companies. Because of this many SWFs do not appoint employees to sit on boards of target firms or communicate intensively with investee firms (Megginson et al.

15 2015; Bortolotti et al. 2010). Furthermore, Bortolotti et al. (2010) found that even when funds do appoint members to sit on the board of investee firms, it often concerns domestic firms.

2.2.2 SWF opacity

As mentioned before in this paper, and throughout the literature on SWFs, one of the main points of criticism that surrounds SWFs is the issue of the lack of transparency that surrounds many funds. Megginson et al. (2015) note that transparency issues were one of the early concerns voiced by SWF critics and that this remains an issue today. The main reason for this lack of transparency is provided to us by Keller (2008), who notes; ‘’SWFs, unlike privately

owned and regulated funds, are not required to disclose information such as fund performance or investment strategy to stockholders’’ (Keller, 2008, p. 342). Because of this, many SWFs

choose not to disclose information concerning the size, intent and origins of investments (Bernstein et al. 2013; Murtinu & Scalera, 2016). Bernstein et al. (2013) add to this concern by arguing that SWFs with a distinct lack of transparency that are headed by politicians could face questionable decision-making abilities.

In recent years, the lack of transparency from SWFs has led to several works and measures that score SWFs on some different transparency related factors. These measures are often created by corporate governance experts who are fascinated by this new class of

international investors with massive capital bases. Megginson et al. (2015) identify two measures of SWF transparency that have become commonly accepted throughout the years. The first measure is the Linaburg-Maduell Transparency Index, which is used by the Sovereign Wealth Fund Institute. This index scores SWFs on ten different categories, and designates them either a one, when a fund is transparent in that given category, or a zero, when it is not. Therefore, the maximum score a fund can achieve is ten, the minimum score is zero. The Sovereign Wealth

16 Fund Institute recommends that a fund should score at least an eight to be sufficiently

transparent. In the latest available rankings, fourth quarter 2016, around half of the listed SWFs scored at least an eight. Thirteen funds scored a perfect ten, while four funds scored a one. In comparison, in December 2014 only ten funds scored a perfect ten and six funds scored a one (Megginson et al. 2015; Sovereign Wealth Fund Institute, 2017).

The second measure is developed by Edwin Truman, and is an SWF scoreboard. This scoreboard measures SWFs in four different categories; Structure of the fund, governance of the fund, accountability, and transparency of the fund in its investment strategy and behaviour of the fund in managing its portfolio and its risk management policies. The maximum score an SWF can achieve is one hundred, the minimum is zero. Norway’s GPFG topped this scoreboard with a score of 96, while other SWFs like the Qatar Investment Authority (QIA) scored only fifteen (Megginson et al. 2015; Truman, 2008; Truman, 2011).

2.3 Political influence, strategy and performance of sovereign wealth funds

2.3.1 Political connections and SWFs

One of the most distinguishing traits of an SWF is that it is owned and funded by the state. As is to be expected, this leaves these funds open and vulnerable to political influences and corruption when not managed correctly. Throughout the recent, and not so recent, literature the political influence on SWFs has been a recurring point of interest. Many scholars have pointed out that SWF investments can be used to further the political goals of the controlling government and politicians, be that in their own country or in a foreign country. Russia, for example, in 2007 used their SWFs to invest in European energy infrastructure and oil pipelines, which led to uncertainty and worries within the European Union and the USA (Aguilera et al. 2016; Bortolotti et al. 2010; Epstein & Rose, 2009; Keller, 2008; Megginson et al. 2015; Murtinu & Scalera,

17 2016)

Additionally, when politicised SWFs pursue strategies that might not directly be related to profit maximisation, but rather to pursue the geopolitical goals of the government this might lead to tensions and suspicions between nations. Keller (2008), for example, notes that currently nothing can stop nations using their SWFs to invest in important foreign industries like high tech companies, energy resources, defence firms or even launch an attack on a currency. Although most researchers do not go as far as Keller does in this instance, most do note that SWFs could technically be used in this way and acknowledge the potential threats this could pose (Bernstein et al. 2013; Bortolotti et al. 2010; Megginson et al. 2015; Murtinu & Scalera, 2016).

Several scholars have written about politically connected firms and the consequences of this. Calluzo et al. (2017) discuss the influence foreign states can have on the political system and elections in the USA. They note that the founding fathers of the USA banned any foreign contributions to political campaigns from kings, princes or states. However, SWFs technically do not identify under those conditions. The authors find that foreign SWFs have an attraction to US campaign finance firms and other politically active firms, in this way non-US SWFs can

theoretically influence the political process within the USA. Faccio (2006) sheds more light on this by researching why firms become politically connected. Faccio (2006) argues that the source of such value can take various forms, including preferential treatment by government owned enterprises such as banks or raw material producers, lighter taxation, preferential treatment in competition for government contracts, relaxed regulatory oversight of the company in question, or stiffer regulatory oversight of its rivals, and many other forms.

Most authors seem to agree that politicised SWFs are less focused on profit maximisation due to their focus on (geo)political goals, which might hurt their actual results (Bortolotti et al.

18 2010; Bernstein et al. 2013; Megginson et al. 2015). Some authors argue that political

connectedness can be beneficial to fund performance. Sojli & tham (2017) note; ‘’In general, we

find that firms with foreign political connections benefit from these relations, which is consistent with decreasing liability of foreignness. We find that the foreign corporate political strategy of having large and active politically connected shareholders is associated with increases in the number of granted foreign government contracts, foreign sales, and DOI.’’ (Sojli & Tham, 2017,

p. 245). Lastly, there are also some authors who mention that a set of best practices should be created for SWFs, which every SWF should then follow. Epstein & Rose (2009), for example, argue that every SWF should publicly declare a dedication to commercial gains, instead of political goals.

2.3.2 SWF Strategy

When discussing SWF strategy and governance Aguilera et al. (2016) propose an interesting framework. In this framework, they introduce four types of strategic governance SWFs generally employ. Aguilera et al. (2016) state that SWFs typically have one of two motives when considering investments; financial or strategic. Furthermore, investee ownership can be either private or public. The four types of strategic governance an SWF can follow are

shareholder activism, with financial investment motive and public ownership; in-house capabilities, also financial investment motives but private ownership; Legitimacy and decoupling, which is public owned but pursues a strategic investment motive; long-term learning, with a strategic investment motive and private ownership structure.

Especially relevant for this paper is the level of political influence within SWFs and how this affects SWF strategy. Regarding this topic several scholars have contributed to the debate. Bortolotti et al. (2015) argue that SWFs which are highly politicised see lower returns on their

19 investments in the stock market, this seems to imply that SWFs with a high degree of political influence negatively affects firm performance. Bernstein et al. (2013) hints at a similar

conclusion, they note that SWFs with a high degree of political influence tend to invest more domestically as opposed to SWFs controlled by external managers. The authors argue that this could be the effect of domestic political pressure, which pressures politicians to use SWF assets to help struggling local industries. Other authors also mention that political connections might influence the location and the industry an SWF invests in (Johan et al, 2013; Knill et al. 2011). Johan et al. note that the degree of political connectedness has an influence on whether SWFs invest in private equity. Knill et al. (2011) add that political influence does not seem to determine the amount of assets that is invested.

2.3.3 SWF Strategic Industries

Related to SWF strategy is the target industry SWFs choose to invest in. Due to the nature of SWFs, which are government-owned, several concerns might arise when funds aim to invest in strategic industries. Several scholars have written about the consequences and reactions of SWFs investing in strategic industries (Bortolotti et al. 2010; Drezner, 2008; Keller, 2008; Megginson et al. 2015: Murtinu & Scalera, 2016; Okhmatovskiy, 2010).

Most scholars seem to agree that SWF investment in foreign strategic industries is one of the more significant concerns surrounding SWFs at this point in time, adding to the earlier mentioned threats of low transparency and possible political influences. Megginson et al. (2015) note on this that ‘’the possibility that their (SWF) capital could be used to further political

purposes and to acquire stakes in strategic industries’’ (Megginson et al. 2015, p. 741). They

categorise this as one of the seven main issues surrounding SWFs.

20 Murtinu & Scalera (2016) classify the following industries as strategic: ‘’financial institutions

(banks, insurance companies), construction and infrastructures, energy (gas, water, electricity), metals and metal products, post and telecommunications, mining, and transportation as strategic industries’’ (Murtinu & Scalera, 2016, p. 8). They base this on the classifications made by both

Drezner (2008) and Keller (2008), both of whom also argue that SWFs may willingly or

unwillingly impact target country politics and decision making when acquiring key infrastructure or firms in strategic industries. An example of this would be the earlier mentioned case of Russia in 2007, who invested in European energy infrastructure and oil pipelines, which in turn led to uncertainty and worries within the European Union and the USA (Aguilera et al. 2016; Drezner, 2008; Keller, 2008).

Lastly, several scholars note that investing in strategic industries, or indeed for strategic purposes, may not only lead to resistance and hostility, but also to lower performance. To support this, Bortolotti et al. (2010) hypothesized that returns for investments made in strategic industries would most likely be lower than other investments. A sentiment also shared by Bernstein et al. (2013), who note that investments aimed at a country’s long-term key industry

development strategy often stand in the way of more practical profit maximisation strategies.

2.3.4 SWF performance

For many companies, performance is a key aspect of doing business. This is no different for SWFs, of which most are or should be focused on profit maximisation. Accordingly, in terms of SWF performance scholars often look at the target firm performance in the period prior to the investment, as well as a certain period after the investment is made. Indeed, a significant portion of existing research on SWFs implies or incorporates performance questions (Aguilera et al. 2016; Bernstein et al. 2013; Bortolotti et al. 2010; Bortolotti et al. 2015; Megginson et al. 2015;

21 Murtinu & Scalera, 2016). For this research, it is especially relevant to focus on those papers that have described SWF performance and have linked this to political presence. Additionally, it is also relevant to discuss papers that have focused on political influences within other types of companies besides SWFs.

Bernstein et al. (2013) research political influences within SWFs, one of the pillars of their research is that of the effect direct political influence has on target firm and SWF

performance. As mentioned earlier, they observe that SWFs that are influenced or dominated by politicians often tend to invest more domestically and in sectors that have a high price to

earnings ratio, that often drops off immediately afterwards. Bernstein et al. (2013) argue that this strategy leads to worse performance, noting that in terms of performance politically influenced SWFs are deviating from long-term profit maximisation. The authors do not go as far as bluntly stating that highly politicised SWFs perform worse, but they do strongly hint at that possibility. Bortolotti et al. (2010) also touch on this subject in their research. They argue that the imposition of political goals is not consistent with long-term profit maximisation. Therefore, they hypothesize that highly politicised SWFs have a negative effect on long-term target firm and SWF performance. Even though the authors never state if this hypothesis is confirmed or not they do note that Norway’s GPFG performs well in their dataset, and seeing this is a fund that is typically not encumbered by political influences they state that this could support their

hypothesis.

Additionally, other authors also hint that political influences and motivations may cause an SWF to pursue goals and thus investments that are not per definition the best option for profit maximisation. Bortolotti et al. (2015) go as far as stating that a political agenda most likely has a negative effect on target firm performance. Knill et al. (2011) are more nuanced, this is logical

22 seeing as they not directly research the performance of SWFs. They note that political

investment decisions might not always be the most practical, and could therefore possibly have a negative effect on performance.

Additionally, Bortolotti et al. (2015) argue that SWFs which are highly politicised see lower returns on their investments in the stock market, this seems to imply that SWFs with a high degree of political influence negatively affects firm performance. Johan et al. (2013) and Knill et al. (2011) add to this by stating that politicised SWFs make different choices when investing as compared to other SWFs. Bernstein et al. (2013) also mentions this, they also openly question the motives of highly politicised SWFs and how this affects performance.

Outside of SWF research on political influence and performance there is also debate on whether political influence or political connectedness has a negative or positive effect on firm performance. Several authors argue that political influence and connectedness may lead to better performance, mostly because this helps bridge institutional barriers like permits and other government regulations. It is also argued that it might help with landing government contracts (Faccio, 2006; Okhmatovskiy, 2010; Solji & Tham, 2017). Other authors argue that political influence and connectedness can also have, and mostly does have, a negative impact on firm performance. These authors often provide examples of corruption or political motives that behave contrary to profit maximisation (Domadenik et al. 2016).

2.4 SWF Location Choices

One of the more important and consequential decision SWFs face when seeking

investment opportunities is in the location choice. When discussing location choice this research distinguishes between investing domestically or cross-border, no further distinctions will be made.

23 The existing literature on SWF cross-border investments is rich, this is most likely due to the political and strategic implications cross-border investments by state-owned enterprises usually are accompanied by. Several authors point out that SWFs that choose to invest cross-border face political pressure from target country governments to remain passive investors (Bortolotti et al. 2010, Megginson et al. 2015). Megginson et al. (2015) note that it would be naive not to recognize that SWFs are state-owned entities that often make politicized capital allocations, however, we need to be mindful of the fact that no evidence exists, to date, of political interference in the behaviour of the foreign targets in which SWFs invest.

Nevertheless, the concerns regarding SWF cross-border investments remain, many scholars note that, despite the lack of evidence, SWFs still face hostility and protectionism measures. This is largely because SWFs tend to be vulnerable to political influences and are often not very transparent (Bernstein et al. 2013; Bortolotti et al. 2010; Bortolotti et al. 2015; Murtinu & Scalera, 2016).

In terms of performance, several authors have written and speculated about SWF investment decisions and location. Al Kharusi et al. (2014) note that institutional investors increasingly seek to invest in developing countries as to maximise their profit and to achieve greater diversification. Interestingly, Al Kharusi et al. (2014) also note that the funds that employ this strategy increasingly turn to external managers as to reduce the political influence within the organisation. What is also interesting to mention is that Bortolotti et al. (2010) conclude that SWF foreign investments systematically lead to negative results, when excluding Norway’s

GPFG. They also mention that SWFs are far less likely to install someone on the board of directors when a foreign investment is made as opposed to domestic investments.

24 et al. (2013) argue that highly politicised SWFs tend to invest more domestically due to political pressures from within the own country. They also suggest that these domestic investments could present a conflict with profit maximisation goals. Bernstein et al. (2013) also note that domestic investments by SWFs tend to be in industries that are underperforming, and these investments function to bail said industries out. Lastly, Johan et al. (2013) find that SWFs are less likely to invest in private equity over public equity when the investment is outside their domicile nation, but also that this has no particular effect on performance.

2.5 Research Gap

To sum up, many scholars researching SWFs have reported on SWF investment

behaviour, specifically on the why and how SWFs make the investments they do (Bortolotti et al. 2010; Bortolotti et al. 2015; Johan et al. 2013; Keller 2008). Other scholars have reported about the political and strategical issues that surround SWFs (Bagnall & Truman, 2013; Domadenik, 2016; Knill et al. 2011; Solji & Tham, 2017). Additionally, there is also some research into the governance structure and internal culture of SWFs (Aguilera et al. 2016; Megginson et al. 2015). Throughout these different types of SWF research many scholars state certain expectations, or

hint at possible outcomes in relation to SWF performance and what affects this performance. Several researchers have hinted that the role of politicians in SWFs might have a negative effect

on the performance of these funds (Bernstein et al. 2013; Bortolotti et al. 2010; Bortolotti et al. 2015; Johan et al. 2013; Knill et al. 2011; Megginson et al. 2015).

Of these scholars, Bernstein et al. (2013) and Bortolotti et al. (2015) have made the most concrete statements about SWF performance relating to political influence. Bernstein et al (2013) note on this that politicised SWFs tend to invest more domestically, with short term goals instead of profit maximisation. Bortolotti et al. (2015) find that highly politicised SWFs see lower

25 returns on their investments in the stock market, this seems to imply that SWFs with a high degree of political influence negatively affects firm performance. However, both Bernstein et al. (2013) and Bortolotti et al. (2015) do not go as far as directly stating that politicised SWFs perform worse than non-politicised SWFs.

This research aims to enrich SWF research by testing whether SWF performance is influenced by the degree to which these funds are politicised. This bridges a gap in the literature, seeing as most research into SWFs does mention expectations regarding performance.

Additionally, as mentioned throughout this literature review, several questions have been raised concerning SWF performance in relation to political influence. By testing the effect political influence has on SWF performance several of the statements made in past papers can be strengthened or, perhaps, dismantled.

3 Conceptual Framework and Hypotheses Development

3.1 Hypotheses Development

3.1.1 Political influence and performance

One of the main themes within this research is the link between politicians and SWF performance. As mentioned throughout the literature and in the literature review of this paper, there are several authors that discuss this topic in some detail. Most of these authors seem to agree that highly politicised SFWs focus too much on achieving (geo)political goals at the cost of investments aimed at profit maximisation. It is argued that this might lead to worse results when compared to funds that are not highly politicised but instead managed by external managers (Bortolotti et al. 2010; Bernstein et al. 2013; Megginson et al. 2015).

There are several mechanisms that cause negative performance implications for highly politicised SWFs, I want to focus on three of these mechanisms. The most described way in

26 which political connections and influences negatively affect SWF performance per scholars is that of profit maximisation. Several authors argue that investments made by SWFs that are controlled or decided by politicians tend to not have the goal of profit maximisation. Instead, these investments are more often associated with political goals or pressure. This leads to reduced SWF performance, as well as reduced target firm performance (Bernstein et al. 2013; Bortolotti et al. 2010; Bortolotti et al. 2015; Megginson et al. 2015).

Linked to the mechanism of profit maximisation is the issue of strategic investment that deserves more attention. Many authors have expressed their concern about SWF investments that do not seem to have any other function than establish political presence, pressure of legitimacy (Aguilera et al. 2016; Bernstein et al. 2013; Bortolotti et al. 2010; Keller, 2008; Megginson et al. 2015; Murtinu & Scalera, 2016). A striking example of an SWF investment like this is the investment made by the Libyan SWF in 2003 into the Italian football club Juventus. Which per Knill et al. 2011 was nothing more than an attempt by the Libyan Gaddafi family to gain respectability in the West and not a sensible performance enhancing SWF investment. This is also a good example of political influence over SWF investments.

The third mechanism that is mentioned in the literature is that of the expertise required to make sound investments as an SWF. Several authors, like Bortolotti et al. (2010) and Bernstein et al. (2013), argue that it is likely that external managers are probably more suited to make investment decisions that lead to better performance. In one part, this is because they often followed an education that provides them with the proper skillset to make these decisions based on data. Secondly, external managers are often impartial and not influenced by political motives or pressure. As example Norway’s GPFG is often used. This fund is relatively unencumbered by political influences and mostly run by external experts (Bernstein et al. 2013; Bortolotti et al.

27 2010; Bortolotti et al. 2015; Megginson et al. 2015).

Together, these mechanisms and their portrayal in the existing literature have led me to believe that a direct political presence within the decision-making process of SWFs has a negative effect on SWF and target firm performance. To test this assumption the following hypothesis has been developed:

Hypothesis 1: The active presence of politicians within SWFs has a negative effect on SWF

performance.

3.1.2 Strategic industry

Throughout the existing literature on SWF several scholars have written about the issue of SWFs investing in foreign strategic industries, and how this is one of the main problems surrounding SWFs and their public image. Often, investments in strategic industries, or the threat of an investment in a strategic industry is enough to be met with hostility or concerns by target country governments (Bortolotti et al. 2010; Drezner, 2008; Keller, 2008; Megginson et al. 2015: Murtinu & Scalera, 2016; Okhmatovskiy, 2010).

As many researchers note, SWF investment into foreign strategic industries are often not investments made from profit maximisation considerations. Instead, these investments pursue political purposes or legitimacy goals (Drezner, 2008; Megginson et al. 2015). When taking this into account, SWF investment in strategic industries becomes especially relevant for this

research. Bernstein et al. (2013) note on this that they classify an SWF as strategic when they invest in physical assets, strategic assets, or domestic development. They then argue that funds they classify as strategic are more often politically connected. Furthermore, they say: ‘’politically

connected managers are not purely making poor decisions when investing but that there is a strategic component.’’ (Bernstein et al. 2013, p. 232).

28 Other authors also express concerns about the performance of SWFs that invest in

strategic industries. Bortolotti et al. (2010) note that returns for investments made in strategic industries would most likely be lower than other investments. Drezner (2008) adds to this a survey dispersed to global financial institutions, including SWFs. They find that SWFs were more likely to seek strategic interests than to maximize their financial returns.

Together this has led me to formulate a hypothesis regarding the influence of investing in strategic industry in relation with the main themes in this research: political influences and performance. I expect to find that SWFs that invest more in strategic industries more often pursue political goals over profit maximisation goals and, therefore, perform worse that other funds.

Hypothesis 2: SWFs that invest more in strategic industries positively moderate the relationship

between highly politicised SWFs and performance.

3.1.3 Location Choice

As stated earlier, in this research with location choice is meant the distinction between cross-border or domestic investment decisions by SWFs. For this research, the impact on performance of these respective choices are particularly interesting. As mentioned in the literature review, several things have been said regarding this in the existing literature. It is noted by many authors that SWFs encounter hostility and sometimes even protectionism when investing cross-border due to the concerns surrounding transparency and political issues (Bernstein et al. 2013; Bortolotti et al. 2010; Bortolotti et al. 2015; Murtinu & Scalera, 2016). Several scholars have also written about the effect SWF cross-border investments have on target firm performance. Bortolotti et al. (2010) note on this that in the three years after the initial investment, target firms generally experience lower returns. In a later research,

29 however, Bortolotti et al. (2015) find that SWFs investments overall tend to lead to decreased performance.

As mentioned before, many authors suggest in one way or another that there is a link between political influence and SWF performance (Aguilera et al. 2016; Bernstein et al. 2013; Bortolotti et al. 2010; Bortolotti et al. 2015; Johan et al. 2013; Knill et al. 2011; Murtinu & Scalera, 2016). Several of these authors also link this to domestic investments made by SWFs. Bernstein et al. (2013) note on this that political SWF managers do not just make worse

decisions than external managers, but purposely make strategic investments, often domestically. Aguilera et al. (2016) argue that SWF investment strategies tend to have a substantial domestic bias when politicians are involved. Lastly, in a more nuanced statement, Megginson et al. (2015) note that SWFs often are too constrained by their governments to form a financial threat. This might hint at decreased performance because of a more domestic focus, which would be in accordance with the findings of other scholars.

Together, these arguments have led me to believe that location choice might have a moderating role in the relationship between politicised SWFs and performance. I expect to find that SWFs that invest more domestically perform worse than those that have a more balanced or foreign view. Additionally, these funds are most likely also more politicised. To test this

relationship the following hypothesis is formed:

Hypothesis 3: Domestically oriented SWFs positively moderate the relationship between highly

politicised SWFs and performance.

3.1.4 Transparency

When analysing the literature on SWFs it becomes clear that SWFs have historically been struggling with transparency. It is almost impossible to find an article about SWFs that does not

30 discuss the transparency issues surrounding these funds. SWFs, contrary to many privately-owned firms, do not have to disclose information regarding performance, strategy or structure (Keller, 2008). Therefore, many SWFs choose not to disclose information concerning the size, intent and origins of investments (Bernstein et al. 2013; Murtinu & Scalera, 2016). Additionally, Bernstein et al. (2013) note that SWFs with a distinct lack of transparency that are headed by politicians could face questionable decision-making abilities, adding to the concerns raised by hypotheses 1.

There are several ways in which low SWF transparency can influence the negative relationship that I expect to find between highly politicised SWFs and performance, and possibly even further enhance this. When examining SWF transparency using the generally accepted Linaburg-Maduell Transparency Index and the Truman Index, what stands out is that it is typically those SWFs from countries that are regarded as undemocratic or less democratic that underperform on transparency, and are, indeed, more often classified as politicised SWFs. Several of these funds with low transparency have been suspected of worse target firm

performance due to investments not necessarily being made with the goal of profit maximisation. On the other hand, one of the most transparent funds, Norway’s GPFG, has been known to perform relatively well (Megginson et al. 2015; Sovereign Wealth Fund Institute, 2017; Truman, 2008; Truman, 2011).

Additionally, funds that are both politicised and non-transparent are not encumbered by any kind of legislation or control (Keller, 2008). Therefore, there is possibly even more room for malign investments that make no sense from a profit maximisation stance point. In this case I expect to find that the lack of transparency contributes to the relationship I expect to find

31 between political influence and performance (hypotheses 1). To test this assumption the

following hypothesis was created:

Hypothesis 4: Low transparency positively moderates the relation between highly politicised

SWFs and performance.

3.2 Conceptual Model

The conceptual model provides an overview of the variables and how they position themselves in relation to one another in this research.

32

4. Methodology

This paper aims to guarantee validity, to ensure validity several cases of SWF investment are used. This way, drawn conclusions can be relied upon to be about what they state to be about. Additionally, by gathering data systematically from available databases as has been done by other researchers, this thesis ensures reliability. Furthermore, this paper follows a largely deductive approach, but also includes some inductive elements. The deductive approach is shown by a beginning point in theory, followed by data analysis. However, as this data is specific data about SWF deals, but aims to contribute to the wider spectrum of SWF research, it also shows inductivity (Saunder & Lewis, 2012). Additionally, this thesis relies for the largest part on secondary data gathered from databases and online sources.

4.1 Sample & Data

This thesis includes a final sample of nineteen SWFs and 139 completed investment into 34 different countries. Table 2 gives insight into the nineteen SWFs and how many investments each respective fund made. Following previous studies that used SWF investments (Bernstein et al. 2013; Murtinu & Scalera, 2016), this study will focus on SWF investments made within a certain period. The dataset used in this thesis will include investments made between 1997 and 2013.

SWF data concerning investments was mostly provided from Bureau van Dijk’s Zephyr

database, following Murtinu & Scalera’s (2016) approach. Zephyr is a comprehensive database containing detailed information about worldwide mergers and acquisitions (Bureau van Dijk, 2017). Identification and classification of SWFs was done in line with comparable research. Murtinu & Scalera’s (2016) approach was followed to classify what an SWF entails and which organisations count as SWF for this research. The list of SWFs as published by the Sovereign

33 Wealth Fund Institute (2017) and Truman’s (2009) paper where used for this task.

Information regarding country specific variables within the sample were collected through several methods. To gather specific information concerning country GDP the World Bank Open Data database was used. This database offers free information about countries in the entire world, including information regarding their development (The World Bank, 2017,). Data regarding measures of cultural distance between SWF home country and target country was calculated through the formula provided by Kogut & Singh (1988). Lastly, information regarding

firm total assets and total employees were also gathered from Bureau van Dijk’s Zephyr database, following Murtinu & Scalera’s (2016) approach.

34

SWF Parent Acquirer Number of Investements

1. Aabar Investments PJSC 1

2. Caisse De Depot et Placement du Quebec 16

3. Canada Pension Plan Investment Board 2

4. China Investment Corporation 5

5. Dubai International Capital LLC 4

6. The Government of Singapore Investment Corporation Pte Ltd 36

7. Government Pension Fund Global 2

8. Guardians of New Zealand Superannuation Fund 1

9. International Petroleum Investment Company 5

10. Investment Corporation of Dubai 3

11. Istithmar PJSC 3

12. Khazanah Nasional Bhd 8

13. Kuwait Investment Authority 2

14. Lybian Investment Authority 2

15. Mubadala Development Company Pjsc 2

16. National Pensions Reserve Fund, The 2

17. Qatar Investment Authority 5

18. Stichting Pensioenfonds ABP 21

19. Temasek Holdings Pte Ltd 19

Total 139

Table 2: SWFs and number of investments.

35 4.2 Variables

4.2.1 Dependent Variable

The dependent variable of this thesis is performance. This variable is measured by analysing two numerical indicators of target firm performance: Return on assets (ROA) and net sales/revenue (NSR). Both indicators are measured in the year leading up to the investment, as well as one year after the investment. This means that two different sets of regression will be run, and be reported on in the results section. These indicators for performance were selected because they accurately represent the size of a company, and therefore its ability to be economically relevant and create and maintain economies of scale. Additionally, ROA provides a view of how efficient a company is at generating revenues. NSR can also depict an up- or downward trend for firms. Additionally, these measures have been used in other research studying performance (Boardman et al. 1989; Bortolototti et al. 2010; Dominguez et al. 2012; Solji & Tham, 2017). The decision to analyse the measurements in the year leading up to the investment, and the year after investment is based on statements made by Bortolotti et al. (2010, 2015), who notes that the effect of an SWF investment can generally be seen relatively fast. Data about the performance measures was gathered from Datastream Worldscope (2017).

4.2.2 Independent Variable

This thesis’ independent variable is whether or not an SWF is politicised or not, and is a

dummy variable. To determine whether an SWF is politicised or not, this thesis follows the information provided by a J.P. Morgan report created in 2008, and described in Fernandez & Eschweiler (2008). As described by Bernstein et al. (2013), one of the key variables Fernandez & Eschweiler (2008) denote is the presence of politicians in the managing bodies of SWFs.

36 that are politicised with a 1, and those that are not with a 0. In following this method, this thesis follows other works that have used this approach (Bernstein et al. 2013; Murtinu & Scalera, 2016).

4.2.3 Moderating Variables

This thesis makes use of three moderating variables. These variables affect the strength of the relationship between the dependent and independent variables, impacting them in a way (Saunders & Lewis, 2012). The moderating variables in this thesis are: (i) strategic industry, (ii) location choice (domestic vs. cross-border) and (iii) transparency.

First, strategic industry is a dummy variable that controls if the target firms operate in a strategic industry, when the target firm is active in a strategic industry it is denoted by a 1, if it is not it is classified as 0. As argued by Drezner (2008), political motives and strategies may influence SWF decisions regarding investments made in strategically important sectors in foreign countries. This tactic might not always be in the best interest of profit maximisation efforts (Bernstein et al. 2013; Bortolotti et al. 2010; Megginson et al. 2015). This thesis follows Keller (2008) and Drezner (2008) in identifying which sectors are denoted as strategic industries. Therefore, financial institutions (banks, insurance companies), construction and infrastructures, energy (gas, water, electricity), metals and metal products, post and telecommunications, mining, and transportation are classified as strategic industries. This method has been previously used by Murtinu & Scalera (2016).

The second moderator in this research is location choice (domestic vs. cross-border). This variable is also a dummy variable that equals 1 when an investment is a cross-border investment, and equals 0 when an SWF invests in the home country itself. It is not uncommon in SWF research to include location choice as a variable, as it is theorised that location choice

37 might influence SWF performance (Bortolotti et al. 2010; Murtinu & Scalera, 2016). This

research follows Bernstein et al. (2013) in distinguishing between domestic and cross border investments. Bernstein et al. (2013) also imply that politicians tend to invest more in domestic markets, therefore it is logical to distinguish between domestic and cross border in this research seeing as it is exactly the type of relation that this study aims to research.

The last moderator variable is transparency. Transparency is represented by a dummy variable that equals 1 if the SWF is deemed transparent, and 0 if it is not. To measure

transparency Truman’s (2008) SWF scoreboard, as explained in paragraph 2.2.2, is used. An SWF is deemed transparent when it scores higher than 59, the average score of all SWFs. Bagnall & Truman’s (2013) SWF work is used as a source to determine transparency. In doing

this, this thesis follows the threshold for transparency set by Murtinu & Scalera’s (2016). Although this is a crude designation of transparency, it still does justice to some of the points raised by Bagnall & Truman (2013), arguing that transparency and accountability are very pressing issues faced by SWFs.

4.2.4 Control Variables

This thesis employs eight control variables, controlling on the level of the SWF, company and on the country level. The control variables are: whether the investment was made during an economic crisis, If the fund qualifies as a SPRF, the size of the SWF, if the purchased stake was minority or majority, target country GDP, cultural distance between countries, firm total assets and firm total employees. Together, these variables control for the case in which they may influence performance in one way or another.

First, crisis, controls for time trends and is a dummy variable that equals 1 if the

38 similar approach to control for time trends was used by Klapper & Love (2011), Magud et al. (2014) and Murtinu & Scalera (2016), this is done because a variance in performance could be caused by an event like an economic crisis. The second control variable, SPRF, is a dummy variable that equals 1 if the SWF is classified by Bagnall & Truman (2013) as an SPRF. Although Drezner (2008) argued and concluded that SPRFs and SWFs mostly have similar goals, this thesis follows Murtinu & Scalera’s (2016) method and thus controls for potential differences between SWFs and SPRFs. Third, SWF size, is the size (total assets) of an SWF in Billions of US dollar. This information is collected from the Sovereign Wealth Fund Institute (2017) and is accurate up until April 2017.

Furthermore, majority, is a dummy variable that controls for the stake of the company that is purchased by the SWF. More specifically, when an SWF purchases more than fifty per cent stake in a firm, this control variable equals 1. A majority stake could have more effect on target firm performance because the SWF would have the possibility to direct the firm in

whichever way it would like to (Megginson et al. 2015). Target country GDP is the fifth control variable and is a numerical variable. GDP gives an insight in the total size of the target country economy, and could shed a different light on investments made by SWFs, measures of GDP or GDP related measures have been used in SWF research before (Bortolotti et al. 2015; Johan et al. 2013; Knill et al. 2011). This data is collected from The World Bank (2017) Open Data Database and provides country GDPs that are updated up until and including 2015. Cultural distance, is a numerical variable that is included to show distance between respective countries. This distance might be part of the explanation when analysing disappointing performance (Hofstede, 1983; Shenkar, 2011). Data regarding measures of cultural distance between countries was gathered from existing databases, following the method used by Murtinu & Scalera (2016).

39

Firm total assets (in absolute number) is a numerical variable that is included to have a

control on the firm level. Total assets can be a useful tool in establishing what measure of impact an investment can have, and therefore could also be helpful in deducing the intent behind an investment. Lastly, firm total employees denotes how many people work at a company. Like firm total assets this is a control on company level. How many employees work at a company might establish a good sense of the size and importance of the investment. Additionally, larger companies might receive more publicity, which could also have an impact on performance. Table 3 provides a detailed overview and description of all the variables use in this thesis.

40

Variable Definition

Performance ROA DV - Performance indicator return on assets

Performance NSR DV - Performance indicator net sales or revenue

Political Influence IV - Dummy variable that equals 1 if an SWF is politicised

Strategic Industry MV - Dummy variable that equals 1 if the

investment is in a strategic industry

Location choice MV - Dummy variable that equals 1 if the

investment is cross border

Transparency MV - Dummy variable that equals 1 if an SWF is

transparent

Crisis CV - Dummy variable that equals 1 if the

investment was made during an economic crisis

SPRF CV - Dummy variable that equals 1 if an SWF is

classified as a Sovereign Pension Reserve Fund

SWF Size CV - SWF size in billion US dollars, SWF total

assets

Majority CV - Dummy variable that equals 1 if a majority

stake was purchased

Target Country GDP CV - Target country GDP in billion US dollars Cultural Distance CV - Value denoting the difference in culture

between target and host country

Firm Total Assets CV - Total firm assets in absolute numbers, US dollars

Firm Total Employees CV - Total number of firm employees

41 4.4 Procedure

To bring this thesis to a satisfying close, and provide a valid answer to the research question, this thesis employs a research strategy that is able to test hypotheses. The statistical testing program SPSS provided by IBM is used to ensure valid statistical testing and the using of various type of tests (Saunders & Lewis, 2012). The raw data that has been collected and

composed in a dataset, its sources described in section 4.3, is quantified and inserted into SPSS. Before running the regression analysis several checks were conducted to ensure normal distribution of the independent and moderator variables. Additionally, a check was conducted to ensure that there are no excessive correlations between the variables. After ensuring that the data can be responsibly tested with the use of SPSS, multiple regression analyses were conducted to test the conceptual model as a whole.

This study has a dependent variable that is measured on the continuous level.

Additionally, this study works with one independent variable measured on the nominal level, and three moderators that are also measured on a nominal level. As SPSS recognizes moderators as independent variables, enough information is available to select a type of regression analysis. In line with my dependent variable, and independent or moderator variables, an OLS regression analysis will be used in this study (Berry, 1993; Gelman & Hill, 2007).

42

5. Results

5.1 Normality and Assumptions Check

To check for normality, the dependent variable, independent variable and moderators of this research were measured by skewness and kurtosis. It is important to conduct this step before running descriptive statistics, checking for correlations and the regression analysis itself. Both skewness and kurtosis values should be, preferably, close to zero. However, it is generally considered acceptable to approve and assume normality when these values are between -2 and 2 (Gelman & Hill, 2007; Groeneveld & Meeden, 1984).

After testing for skewness and kurtosis it showed that the independent variable, political influence, as well as all the moderators, strategic industry, location choice and transparency, all are normally distributed with reported values of skewness and kurtosis between -2 and 2 and within the range of the lower and upper confidence interval of 95%. It was also shown that both measure for the dependent variable, ROA and net sales (or revenue), were not normally

distributed due to showing values outside of the boundaries set. ROA’s skewness is acceptable at 1.076, Kurtosis however shows an unacceptable value of 4.590. Net sales (or revenue) has a skewness value of 2.696 and a kurtosis value of 15.585, both of which are unacceptable. As noted by Gelman & Hill (2007) these unacceptable values for skewness and kurtosis can be caused by outliers in the dataset. Therefore, an outlier analysis was conducted to check for possible errors in data entry or processing (Gelman & Hill, 2007). It was shown that there were no errors, thus several outliers were removed from the dataset. After removal of outliers both ROA (skewness: 0.010, kurtosis: 1.995) and net sales (or revenue) (skewness: 0.092, kurtosis: 1.021) returned to acceptable levels.

43 There was linearity as assessed by partial regression plots and a plot of studentized residuals against the predicted values. There was independence of residuals, as assessed by a Durbin-Watson statistics of 1.732 (ROA) and 1.847 (net sales or revenue). There was homoscedasticity, as assessed by visual inspection of a plot of studentized residuals versus unstandardized

predicted values. There was no evidence of multicollinearity, as assessed by tolerance values greater than 0.1. There were no studentized deleted residuals greater than ±3 standard deviations, no leverage values greater than 0.2, and values for Cook's distance above 1 (Laerd Statistics, 2015).

5.2 Descriptive Statistics

As becomes clear when looking at table 4, out of the original 139 investments made by SWFs, 111 are valid after the exclusion of outliers and missing values in the previous section. It is also interesting to note that when looking at ROA as a performance measure the mean change measured seems to be more noticeable at 22.09 per cent, than when looking at net sales (or revenue) as a performance measure, which has a mean of 5.86 per cent. Additionally, it seems that 45 per cent of the investments studied were made in strategic industries, 71 per cent of investments were cross border and in 59 per cent of the cases the involved SWF was considered to be transparent.

Furthermore, 68 per cent of investigated investments was made during crisis years, and twenty per cent of the investments involved an SWF that is considered to be a SPRF. The smallest SWF in terms of total assets controlled funds worth twelve billion US dollars, the largest controls over 922 billion US dollars, and the mean hovers around 269 billion US dollars. Cultural distance varies between 0.0000, which would be domestic investments, and 5.2248, with a mean of approximately 1.357. Lastly, the smallest firm included in the dataset in terms of