The Impact of Sovereign Wealth
Fund Transactions on Firm Value
University of Groningen
MSc International Business and Management
Specialization: International Financial Management
Uppsala Universitet
MSc International Economics and Business
Author: Laurens Marie Student Number: S1610554 Supervisor: Dr. Ing. N. Brunia
The Impact of Sovereign Wealth Fund
Transactions on Firm Value
Abstract
Sovereign Wealth Funds – rapidly growing government-owned investment vehicles – are investing heavily in equity nowadays. In this paper I analyze the impact of Sovereign Wealth Fund transactions on firm values in the context of large shareholder and corporate governance theory. I find that the market reacts significantly positive (negative) on the announcement of an investment (divestment). Additionally, I find that the market reacts stronger to announcements of divestments since 2008. This provides some evidence for the idea that investors look more favorable upon Sovereign Wealth Funds since 2008.
JEL classification: G15, G29, G34, G38
Contents
1. Introduction ... 4
2. Literature Review ... 7
2.1 Background on Sovereign Wealth Funds ... 7
2.2 Sovereign Wealth Funds and Corporate Governance ... 8
2.3 Target Firm Characteristics... 13
2.4 Sovereign Wealth Fund Characteristics... 17
2.5 Long-term performance ... 19
3. Methodology and Data ... 20
3.1 Empirical Analysis ... 20
3.2 Data and Descriptive Statistics ... 27
4. Results and Discussion ... 32
4.1 Announcement Period Returns ... 32
4.2 The announcement returns for subsamples ... 35
4.3 Cross-sectional analysis ... 39
5. Conclusion ... 43
Literature ... 45
Appendix A: Growth of Sovereign Wealth Funds and OECD Countries ... 50
Appendix B: Sovereign Wealth Fund Transparency Index ... 50
Appendix C: Methodology ... 51
Appendix D: Data and Descriptive Statistics ... 54
Appendix E: Results ... 62
1. Introduction
Sovereign Wealth Funds are government-controlled funds that invest and manage their countries’ excess reserves. Even though the relatively unknown Funds have been around for a long time, their number and level of activity have increased dramatically since 2000 and this growth is expected to continue during the next few years (Chhaochharia and Laeven, 2008; Knill et al, 2008). Nowadays, Sovereign Wealth Funds have about $5 trillion of assets under control and this has attracted a lot of attention from politicians and media (TheCityUK, 2012). Especially since the Funds acquired stakes in well-known firms like Volkswagen AG, Porsche, CitiCorp, and Barclays.
Since the Funds usually take large equity stakes, they are comparable to other institutional shareholders – like mutual, pension, and hedge funds – that usually also acquire significant stakes in firms (Dewenter et al., 2010). Because the Sovereign Wealth Funds are controlled and influenced by their governments, the objectives and activities might differ from those of other institutional shareholders. For example, the Funds might pursue strategic or political goals that favor the home country of the Fund. Additionally, many Sovereign Wealth Funds are characterized by a lack of transparency and do not reveal any information on their portfolio, their investments, and the goals of their investments. Because of the governmental influence and the opaqueness, it has been suggested by politicians and media that the Funds might not always pursue goals that are in the benefit of existing shareholders (e.g. Knill et al., 2009). It is therefore not surprising that many countries have been skeptical towards Sovereign Wealth Funds and worry about the investments they undertake (e.g. Drezner, 2008; Dewenter 2010).
However, like other large shareholders, Sovereign Wealth Funds are likely to create value for the existing shareholders because they are able to improve the corporate governance of the firms they invest in (Shleifer and Vishny, 1986). One way of achieving this is by actively monitoring the management of the target firm. Investors could therefore react positively (negatively) on the announcement of an investment (divestment) in a firm by a Sovereign Wealth Fund.
To my knowledge, only six empirical (working) papers have dealt with the impact of transactions by Sovereign Wealth Funds on the firm value. Chhaochharia and Laeven (Working Paper; 2008), Raymond (2008), Bortolotti et al. (Working Paper; 2009), Knill, Lee, and Mauck (Working Paper; 2009), Kotter and Lel (2011) and Dewenter et al. (2010) all found significant positive share price reactions to the announcement of an investment by a Sovereign Wealth Fund, ranging from about half a percent to approximately 2.5 percent during the days surrounding the announcement of the investment. Additionally, Dewenter et al. (2010) investigated the impact of the announcement of Sovereign Wealth Fund divestments and found negative average abnormal returns of -1.4%.
onwards Sovereign Wealth Funds have experienced a significant change in the way they are being perceived (El-Erian 2010). Since that year, host countries look more favorable upon investments by Sovereign Wealth Funds for two reasons. Firstly, in 2008, 24 voluntary principles for Sovereign Wealth Funds were published, in order to increase the understanding of the Funds (International Working Group of Sovereign Wealth Funds, 2008). These principles aim to increase the transparency of the Funds, and they ensure that the Funds comply with the rules and regulations of the host countries. Secondly, liquidity evaporated on the global financial markets during the crisis, since investors were hesitant to buy shares in companies. However, Sovereign Wealth Funds still undertook investments, which were welcomed by politicians, because these investments showed trust in the performance of the targeted firms and the local economy. As a result, politicians and journalists look more favorable upon the Funds since 2008 (Persaud, 2010).
This might be reflected in the market returns on the announcement of a transaction by a Sovereign Wealth Fund. It is therefore of importance to investigate whether the impact of Sovereign Wealth Fund transaction on firm value is different since 2008. I investigate 232 investments and 101 divestments in public companies that were undertaken by Sovereign Wealth Funds in the period January 2004 – July 2011. I find that the share price of a firm in which the Sovereign Wealth Fund announces to invest rises significantly with 1.29% during the [0,+1] window. Similarly, I find that divestments result in abnormal returns of -1.04% during that same event window. Additionally, I find no evidence that indicates that the abnormal returns upon the investments by less transparent Funds result in a more negative market reaction than by transparent Funds. These findings therefore indicate that investors do not worry about Sovereign Wealth Funds. Instead, they see the Funds as financial entities that pursue financial goals and add value to the firm.
Next, I find some empirical evidence that supports the idea that investors look more favorable upon Sovereign Wealth Funds than before 2008. Even though I do not find any statistical differences between the investment returns before and after 2008, I do find that divestments after 2008 result in a more negative reaction of the market than before.
This paper differs from other published articles in several ways. Firstly, my focus is on the period 2004 – 2011, which is different and more recent when compared to the other papers. Because of the developments since 2008, it is of importance to investigate whether the markets react differently to announcements of Sovereign Wealth Fund transactions since 2008. Secondly, to my knowledge no paper has investigated whether the size and the net debt ratio of the target firm are related to the reaction of the market upon the announcement of a Sovereign Wealth Fund transaction.
2. Literature Review
2.1 Background on Sovereign Wealth Funds
The term Sovereign Wealth Fund was first used by Andrew Rozanov in 2005 (Fotak et al., 2008), but Dewenter et al. (2010) trace the Funds back to 1876. In this year, the State of Texas founded the Permanent University Fund, which invested part of its leases on state lands to support the universities in the State. The first Sovereign Wealth Fund that is still active is the Kuwait Investment Authority. This Fund was founded by the Kuwaiti government in 1953 to manage the financial surpluses earned in the oil business. The Funds have thus been around for a long time, but the scope and size of them has changed during the last decennium (Balin, 2008).
Back in 2000, the Funds managed about $1.5 trillion in assets (Balin, 2008). Nowadays, however, they have around $5 trillion in assets under control. This means that the Sovereign Wealth Fund market is already bigger than the hedge fund and private equity market combined, as can be seen in Appendix A.1 (TheCityUK, 2012). Up to 2011, only 23 countries had formed Sovereign Wealth Funds, but more and more countries consider creating Sovereign Wealth Funds. The assets under control by Sovereign Wealth Funds are therefore expected to grow significantly during the following years; back in 2008 Morgan Stanley even projected the Funds to grow to $12 trillion in 2015 (Dewenter et al., 2010). This growth is the direct consequence of the expansion of global foreign reserve assets; especially oil producing countries in the Middle East and emerging markets in Asia experienced cash inflows that exceeded the liquidity requirements of the governments (Weiss, 2008). Consequently, from 2000 onwards countries like China, Qatar, and Saudi Arabia have created Sovereign Wealth Funds to manage part of their reserves.
2.2 Sovereign Wealth Funds and Corporate Governance
Sovereign Wealth Funds often take large stakes in companies and they keep these stakes for longer periods of time. Like other large shareholders (e.g. pension funds, hedge funds, and mutual funds), the Funds can be expected to have an impact on the corporate governance of a target firm (e.g. European Central Bank, 2007; Dewenter et al., 2010; Bortolotti et al., 2008).
Since Sovereign Wealth Funds usually take significant stakes in companies, they will have more incentives to monitor the performance of these firms than small shareholders, who are not big enough to account for the costs that accompany monitoring the managers of the firm. The presence of Sovereign Wealth Funds could lead to more monitoring activity, which increases the value of the target company (Shleifer and Vishny, 1986).
The literature on Sovereign Wealth Funds indeed provides empirical evidence that the presence of the Funds leads to more monitoring activity in target firms. Dewenter et al. (2010), for example, investigated the 5-year post-transaction period of 184 Sovereign Wealth Fund investments and found that 27.2% of these transactions were followed by monitoring events. Examples of such monitoring events are increased senior management turnover, the setting of dividend policy, personnel training, and product pricing (Dewenter et al., 2010). These monitoring activities by Sovereign Wealth Funds and other large shareholders are expected to improve the value of the target company. Therefore, it is likely that the share price will rise upon the announcement of an investment by such a large shareholder.
Chhaochharia and Laeven (2008), Bortolotti et al. (2009), Knill, Lee, and Mauck (2009), Kotter and Lel (2011) and Dewenter et al. (2010) conducted research on the announcement of Sovereign Wealth Fund purchases. All found indeed significant positive share price reactions on the announcements, ranging from about half a percent to approximately 2.5 percent during the days surrounding the announcement of the investment. These results are in line with literature on other large shareholders. Several studies (e.g. Seifert et al., 2004; Mikkelson and Ruback, 1985; Holderness and Sheehan, 1988) found that share prices of target firms went up, when it was announced that a new large shareholder took a stake in them. The gains on the shares range from 0.6% to 7.3% during the days surrounding the announcement of the investment.
The divestments by Sovereign Wealth Funds are likely to have the opposite effects of investments. Since the market expects that the divestments will lead to less monitoring opportunities of the Sovereign Wealth Fund, a divestment might be accompanied by declining firm value. Therefore, I expect that firms dealing with a Sovereign Wealth Fund divestment will experience negative abnormal stock price reactions during the days surrounding the announcement. To my knowledge, Dewenter et al. (2010) published the only paper that did research on the announcement of Sovereign Wealth Fund divestments. They studied 47 divestments by Sovereign Wealth Funds during the period 1996-2008 and found that divestment announcements result in significantly negative average abnormal returns of -1.4% during the [-1,+1] window.
Table 1: Empirical Research on Stock Price Reactions on Sovereign Wealth Fund and Other Large Shareholder Investments
This table presents the results of the impact of investments by both Sovereign Wealth Funds and other large shareholders on the share price of the target firm. The second column presents the period during which the research took place, the third column presents the details on the sample, and the fourth column presents the returns, where AR and respectively stand for Abnormal Returns and Cumulative Abnormal Returns. Information on the methodology that has been used in these papers can be found in Appendix C.1. *** indicates significance at 1%, ** indicates significance at 5% and * indicates significance at 10%.
Authors Period Sample Results
Fotak, Bortolotti, and Megginson (2008)
1989 – 2008 75 events
(62 firms in 23 countries)
AR[0,0]=0.94%***
Raymond (2008) May 2005 – April 2008 50 events
(50 firms in 23 countries)
CAR[-1,+1]=3.85%***
Chhaochharia and Laeven (2008)
1997 – 2007 89 events
(Firms all over the world)
CAR[-10,-5]=1.15%*** CAR[-20,+10]= 1.76%*** CAR[-5,+2]= 0.96%*** Knill, Lee, and Mauck (2009) Inception of Fund
– December 2009
232 events
(Firms all over the world)
CAR[-1,0]=1.37%***
Dewenter, Han, and Malatesta (2010)
Jan.1997- April 2008 202 events
(US and non US firms)
CAR[-1,+1]=1.70%***
Kotter and Lel (2011) 1980 – Feb. 2009 417 events
(326 firms in 45 countries)
CAR[0,+1]=1.32%*** CAR[-1,+1]=2.25%*** CAR[-2,+2]=2.74%*** Mikkelson and Ruback (1985) 1978 – 1980 473 events
(Firms listed on the NYSE/AMEX that received >5%investments) CAR[-1,0]=2.88%*** AR[-1]=2.64%*** AR[0]=0.34%*** Holderness and Sheehan (1988) 1978 – 1982 31 events
(20 listed firms on the NYSE/AMEX)
CAR[-1,0]=7.3%*** CAR[-20,+10] =12.8%*** Barclay and
Holderness (1991)
1978 – 1982 115 events
(Firms listed on the NYSE/AMEX that received >5%investments)
Size Stake and Abnormal Returns: A Nonlinear Relationship
Sovereign Wealth Funds and other large shareholders do not only provide monitoring activities, but they might also seek ways to extract private benefits, thereby expropriating wealth from other (minority) shareholders (Shleifer and Vishny, 1997). Like other large shareholders, Sovereign Wealth Fund could transfer assets and profits from the company to themselves or their home countries by means of self-dealing transactions. Johnson et al. (2000) refer to these activities as tunneling and give examples of how large shareholders can transfer wealth to themselves. A first example deals with transfer pricing: a firm could transfer products or assets to its controlling shareholder at below market prices or it could buy products from the controlling shareholder at above market prices. This would transfer wealth from the target firm to the large shareholder. Moreover, the large shareholder could take out loans and use assets of the target firm as collateral. These tunneling activities will reduce the value of a target firm, whereas the monitoring activities that might be undertaken by Sovereign Wealth Funds – and other large shareholders – increase the value. This implies that the announcement return of larger Sovereign Wealth Fund (or other large shareholder) investments is a nonlinear function of the size of the stake that has been acquired. It is likely that the value of a target firm will increase when the acquired stake stays below some critical level, will reach its maximum when it reaches this critical level, and will decline when the size of the stake exceeds this level (Stulz, 1988; McConnell and Servaes, 1990).
This theory is supported by several studies on both Sovereign Wealth Funds and other large shareholders. Dewenter et al. (2010), who studied Sovereign Wealth Funds, found a nonlinear relation indicating that abnormal returns will increase when a company acquires a stake up to about 45%, but stakes larger than 45% are associated with lower positive abnormal returns.1 These results are in line with other research. For example, McConnel and Servaes (1990) conducted a cross-sectional analysis on 1173 firms in 1976, and on 1093 firms in 1986 and found a nonlinear relationship between the management ownership and the market valuation of the company, as measured by Tobin’s Q. Their results show that Tobin’s Q first increases, until ownership is about 40 to 50%, and then starts to decrease.
I expect the opposite effects for divestments by Sovereign Wealth Funds. The larger the divestment, the more likely it is that the firm targeted by the Sovereign Wealth Fund will experience less tunneling activities, and thus the abnormal returns will become less negative. Dewenter et al. (2010) provide empirical evidence that is in line with this theory. They found that share prices first decrease upon an announcement of a divestment by a Sovereign Wealth Fund, but will become less negative once they start to divest more than 20% of the shares of the target company.
First-time and Subsequent Transactions
It is likely that first-time Sovereign Wealth Fund investments result in higher abnormal returns than when the Fund announces a subsequent investment. When the Fund invests in a firm for the first time, the market might be more surprised and expects the Fund to increase the corporate governance of the targeted company and thus increases the firm value. However, when the Fund already has a stake in the company, the reaction of the market might not be as large, since the Sovereign Wealth Fund already had the opportunity to influence the corporate governance of the target company. Dewenter et al. (2010) investigated 202 Sovereign Wealth Fund investments in listed US and non-US companies and found that initial investments indeed result in larger abnormal stock price returns than follow-up investments.
In case of partial divestments, the Fund can still undertake monitoring activities. However, when it divests its entire stake, it will not have any connections with the target firm anymore and will therefore no longer conduct any monitoring activities. Therefore, I expect that the market reacts stronger in case of entire divestments.
Sovereign Wealth Funds Transactions after 2008
Since early 2008, there have been three developments that have resulted in a more favorable look upon investments by Sovereign Wealth Funds (Persaud, 2010). An explanation might be that the International Working Group of Sovereign Wealth Funds (IWG) published 24 voluntary Fund principles; the so-called Santiago Principles (IWG, 2008). These practices for Sovereign Wealth Funds address the concerns that have been raised by politicians and media and have to increase the global understanding of the Funds (Norwegian Ministry of Finance, 2011). In order to do this, the principles ensure that the Funds are transparent, comply with the rules and regulations of the country they invest in, and make investments based on a risk-reward trade-off (IWG, 2008).
Thirdly, the recent turmoil in the global economy has had an impact on the investment and divestment policies of Sovereign Wealth Funds. According to Bernardo Bortolotti, executive director of FEEM2, Sovereign Wealth Funds approach risk in a different way since the crisis that emerged at the end of 2007. The Funds have become more risk averse and this has resulted in several changes in their transactions patterns (Monitor Group, 2009; 2010a). For example, Sovereign Wealth Funds shifted their attention from the mature OECD markets to the (familiar) emerging markets. In 2007, about 40% of the investments took place in emerging markets, while in 2010 60% took place in these countries (Monitor Group, 2011). Since the Funds have become more risk averse, it is likely they will only make an investment when they are sure there are opportunities to increase firm value by means of monitoring activities. Because investors know the Funds have become more risk averse, they might expect that the Fund has scrutinized a target firm when they buy a stake of it.
It can thus be expected that investors look more favorable upon investments by Sovereign Wealth Funds since 2008. Moreover, because the Funds have become more risk-averse, it is likely that they make more careful investments. As a result, I expect that abnormal returns upon the announcement of an investment by a Sovereign Wealth Fund after 2008 are higher than returns upon announcements before 2008.
For the same reasons, I expect that the market returns on the announcement of a divestment are stronger after 2008. Since that year, investors might look more favorable upon Sovereign Wealth Funds and it is therefore likely that they react stronger on the divestment announcement of a divestment.
2.3 Target Firm Characteristics
The characteristics of the target firm could influence the market reaction on the announcement of an investment by a Fund (Kotter and Lel, 2011). In this section I look at the sector, the free cash flow, the size of the company, and the ownership structure of the target firm and I reason how these characteristics can have an influence on the market reaction.
In case I expect differences between transactions before and after 2008, I will mention this. When I expect that the characteristic has no specific influence in either period, the theory holds for both periods.
Sector
Leaders of countries like the United States, the United Kingdom, France, and Germany, have expressed their fear that the Sovereign Wealth Funds could acquire strategic stakes in companies that are of importance to the governments of the host countries. According to Knill et al. (2008), one major concern of host governments is the national security; a concern that is fueled by the idea that the Funds tend to invest in sectors that are of strategic importance to the country.3 Chhaochharia and Laeven (2008) investigated the investments of eight Funds and indeed found that the Funds favored strategic sectors. For example, the Government of Singapore Investment Corporation had invested 91.1% of its capital in the financial sector at the end of 2007, whereas the Abu Dhabi Investment Authority had invested 60.3% in the oil and gas sector.
Since governments and journalists warn the public about investments in such sectors, it is likely that the market reacts less favorable when a Sovereign Wealth Fund announces an investment in a firm that operates in a strategic sector. However, it can be expected that especially the presence of an opaque Fund in a firm operating in a strategic sector leads to lower abnormal returns than an investment by a transparent Fund in a strategic sector.
However, since 2008 the media and politicians look more favorable upon investments by Sovereign Wealth Funds, as has been mentioned before. It is likely that investors also look more favorable upon investments in strategic sectors after 2008. For example, in 2009, the British ministry of Business Affairs argued that the British economy needed investments to restore trust in companies and the local economy and would therefore welcome investments by Sovereign Wealth Funds in all sectors (Reuters, 2009). Consequently, I expect that investments in strategic sectors by Sovereign Wealth Funds after 2008 result in higher abnormal returns than similar investments before 2008.
In case of divestments, I expect the opposite results. When a Fund announces to leave a strategic sector, the market might react less strongly. However, it is likely that investors look more
3
favorable upon Sovereign Wealth Funds since 2008. Therefore, I expect that sales of stakes in strategic sectors by Sovereign Wealth Funds after 2008 will result in stronger negative abnormal returns than similar divestments before 2008.
Free Cash Flow Hypothesis
When a Sovereign Wealth Fund or another large shareholder invests in a company, it can be expected that the market responds more favorable when a firm with high free cash flows is targeted. According to Jensen (1986), free cash flows represent cash flows that are available after having funded all projects that have positive net present values. When a company has a high free cash flow, corporate management either has to invest the money or has to pay it out to its shareholders. Managers, being the agents of shareholders, are expected to act in the interest of these shareholders. However, paying out the free cash flows to the shareholders might not always be in the interest of the managers, so the free cash flows could lead to conflicts between the parties. Agency theory, which is directed at this relationship between the agent and the shareholder, has analyzed this problem of conflicting interests extensively and argues that it is difficult and expensive for the shareholder to manage the actions of the agent (Jensen, 1986; Eisenhardt, 1989). Reducing the free cash flows by means of paying out the money to the shareholders is one way to prevent possible conflicts (Jensen, 1986). When the management pays out the free cash flows, it cannot fund low-return projects that are not in the interest of the shareholder. Another way to prevent agency conflicts that arise because of high free cash flows, is to issue more debt. Debt reduces the conflicts between the agent and the owner, since the payments of the interest - and eventually the principal – motivate the managers to manage the cash flows effectively. This will prevent them from using the available cash at their own preference.
likely smaller than the abnormal returns of companies with higher cash positions. To my knowledge, the literature on Sovereign Wealth Funds has done no research on the impact of free cash flows on firm value. Kotter and Lel (2011), however, did investigate the impact of cash on firm value. They studied 417 Sovereign Wealth Fund investments in 326 unique firms during the period 1980 – 2009 and found that the market reacts more favorable when a firm has a lower cash to assets ratio.
When a Fund announces a divestment in firm with a high free cash flow, I expect the market to react stronger than when the Fund announces a divestment in a firm with a low free cash flow. Firms with a high free cash flow require monitoring by a firm, as has been mentioned before. When a Fund leaves such a firm, investors realize there will be less monitoring activity that is required to manage the free cash flows effectively. Therefore, the market will react stronger to this announcement, than to the announcement of a Fund to sell a stake in a company with a low free cash flow, since the Fund has fewer monitoring opportunities in this latter firm.
Ownership Structure
It can be expected that other large shareholders already took stakes in the companies the Sovereign Wealth Funds are interested in. If this is the case, it can be assumed these parties already have undertaken actions to improve monitoring activities (Shleifer and Vishny, 1986). When a Sovereign Wealth Fund also makes an investment in a firm in which other large shareholders are present it might have fewer opportunities to increase the monitoring productivity. Consequently, it is likely that Sovereign Wealth Fund investments in companies with fewer shares freely floating result in a less favorable market reaction.
The same applies to divestments by Sovereign Wealth Funds. When a Fund announces it will divest its stake in a firm that is characterized by a high amount of freely floating shares, the market reaction might be more negative than when it announces the sale of a stake in a firm with fewer shares freely floating around. In the latter firms, it might be that there are other large shareholders present that undertake monitoring activities.
Size
additional shares without affecting the stock price of the target company, it might have more incentives to monitor the firm. Gerken (2009) provides empirical evidence that supports this theory. He studied 18,210 large investments in S&P 1500 firms by foreign external shareholders between 1994 and 2005. He found that higher liquidity leads to more corporate governance and monitoring activity.
Since companies with a high market capitalization are more liquid than firms with a relatively low market capitalization, I expect that larger – and thus more liquid – companies offer Sovereign Wealth Funds more incentives to monitor (Bolton and Von Thadden, 1998; Busse and Green, 2001). Eventually, the Funds can profit from their monitoring by acquiring additional shares at a price that does not reflect the improvements that result from the increased monitoring.
2.4 Sovereign Wealth Fund Characteristics
Next to the characteristics of the target firm, the characteristics of the Sovereign Wealth Fund could also influence the reaction of the market upon the announcement of an investment by a Fund. Like institutional investors, Sovereign Wealth Funds are not all the same: each Fund pursues its own goals and plays its own role, especially when it concerns monitoring (Borokhovich et al., 2006; Gillan and Starks, 2003). Because large shareholders all have different reasons to monitor, and the effectiveness of monitoring itself also varies, the impact of Sovereign Wealth Fund investments on the value of the firm might depend on the characteristics of the Fund. In this section I identify two of these characteristics: the country of origin and the transparency of the fund.
In case I expect differences between transactions before and after 2008, I will mention this. When I expect that the characteristic has no specific influence in either period, the theory holds for both periods.
Country of Origin
Sovereign Wealth Funds usually invest a significant portion of their financial assets in the domestic market. An explanation for this domestic focus is the strong connection the Funds have with their own governments. The Funds could use these connections to ensure that the firms they invest in receive a more favorable treatment of the government. For example, the target firm could have better chances of receiving government contracts. The firms that receive an investment from a domestic Sovereign Wealth Fund could thus profit from the connections the Fund has with the government, which would increase the value of the target firm. Several authors (e.g. Facco, 2006; Goldman, Rocholl, and So, 2009) indeed found empirical evidence that shows that companies that have ties with their domestic governments have higher market returns than those companies that have no ties with the government.
However, since 2008 several Sovereign Wealth Funds started to invest in their home markets to support and revive the local economy (e.g. Gilson et al., 2009; Raymond, 2008). For example, the Irish National Pension Reserve Fund acquired 25% stakes in two troubled Irish banks restore trust in the banking sector and to prevent the collapse of the Irish banking system in December 2008 (Belfast Telegraph, 2008). Moreover, the French President Nicolas Sarkozy confirmed that their Sovereign Wealth Fund, the Fonds Stratégique d' Investissement, was founded to develop domestic startup companies and to protect French companies from foreign and speculative investors (The Guardian, 2007). I therefore expect that domestic investments after 2008 will create less value than domestic investments before 2008, since investments
The opposite holds for divestments. Before 2008, investors might have reacted more negatively to divestments by domestic Funds than to foreign divestments, since the Funds could have heard about negative firm information from their governments. After 2008, however, the market might react less strong on domestic divestments by Sovereign Wealth Funds than foreign divestments, since it could be an indication that the domestic Funds no longer see the necessity of having a strategic stake.
Transparency
The objectives of Sovereign Wealth Funds are not always clearly communicated to the outside world, since not all Funds are equally transparent (Truman, 2008). Many Funds are from countries with authoritarian governments that give no insight in the activities and objectives of their Sovereign Wealth Funds. For example, Funds like the SAFE Investment Company (China) and the International Petroleum Investment Company (Abu Dhabi) give no insights in their operations and policies. This opaqueness worries investors, politicians and academics, because it cannot be determined whether the Funds make investments for strategic or commercial reasons.
investments of the Fund are based on financial objectives and lead to higher values of the target firm.
Based on this reasoning, I expect opposite results for divestments. When a transparent Funds divests a share, the market will react less favorable than when an opaque divests a share, since it is likely that the transparent Fund pursues financial goals, while the goals of the opaque Funds are not clear.
2.5 Long-term performance
3. Methodology and Data
This section is split into two parts. The first part introduces the methodology and the second part will describe the data used to conduct the described analyses. The used methodology is partly based on the methodology that has been presented in the Sovereign Wealth Fund literature (Kotter and Lel, 2011; Dewenter et al., 2010), as can be found in Appendix C.1.
3.1 Empirical Analysis
The reactions of share prices to the announcement of a Sovereign Wealth Fund investment are measured by an event study. In this study, the announcement date 0 is the event date and the 15 days prior to the announcement and the 10 days after comprise the event window of [-15,+10]. Furthermore, the 180 days before the event window [-195,-16] are used as the estimation window. According to Kothari and Warner (2005), the estimation period has to contain sufficient observations to estimate the parameters of the model. Recent Sovereign Wealth Fund literature (e.g. Chhaochharia and Laeven, 2008; Knill, Lee, and Mauck, 2009; Kotter and Lel, 2011; Dewenter et al., 2010) uses periods between 140 and 250 days and I decide on an estimation period of 180 days. As suggested by MacKinlay (1997), the following parameters of the market and risk adjusted model are then used to calculate the abnormal returns (AR) for stock j at day t. The parameters are:
̂
[ ]
∑
(1)
̂
[ ]
∑
(2)
̂
∑ t – ̂ mt- ̂m ∑ mt – ̂m
̂
̂
-
̂ ̂
m̂
[ ]∑
– ̂
̂
Where ̂ and ̂ are estimated over the estimation window [-195,-16], where t is the total return on
denominated in US Dollars and are calculated by taking the first difference of the log of the prices of respectively the share and the index, and are adjusted for dividends and stock splits.4
Based on these parameters, the Abnormal Returns (AR) for stock j at day t can be calculated by using the following formula:
– ̂
̂
The abnormal returns have a conditional mean of 0, since it is tested whether the returns differ significantly from the normal expected returns on the announcement date, and have a conditional variance
equal to:
̂
(
̂
̂
)
This variance of the daily abnormal returns has two components. According to MacKinlay (1997), the first component is disturbance variance that is estimated from the residuals of the market model and this disturbance variance can be calculated by using equation (5) and the second component is a sampling error term. Since the estimation period of [-195,-16] is large enough, this sampling error term will converge to 0 (MacKinlay, 1997).
Next, I make an adjustment in measuring the abnormal returns. Following Patell (1976), I standardize the abnormal returns.5 According to MacKinlay (1997), this standardization will result in more powerful tests, since it ensures that an individual company with a high variance will not influence the results of the tests. The Standardized Abnormal Returns (SARs) can then be computed by dividing the daily abnormal returns of each security by the standard deviation of each firm
̂
:
̂
The next step is to test for the significance of the Standardized Abnormal Returns on each individual day. To test this, I conduct a one-sided t-Test that has been suggested by Brown and Warner (1985). This test assumes that the residuals are not correlated and that there is no event-induced variance. Since the abnormal returns are standardized, it prevents a single firm will dominate the results.
Z=
∑√
Where Nt refers to the number of companies at day t.
4
In this study I use the logged changes, since Raymond (2008) argues that it is better to use the logged differences instead of the growth rate, as this will reduce the kurtosis of the series.
5
In order to measure the abnormal returns during several days, I calculate the Standardized Cumulative Abnormal Returns (SCARs) for company j during the event window [t1,t2], as has been suggested by Patell (1976).
SCAR
j,t1-t2=
∑ ( ) √∑ ̂Finally, I test for significance of the Standardized Cumulative Abnormal Returns. First, I sum the standardized cumulative abnormal returns across the period [t1,t2] and I divide the outcome by the squared root of the number of events (N). This one-sided t-Test, which has also been used by Dewenter et al. (2010) and Raymond (2008), is presented in equation (11):
Z=
∑√
However, according to several studies (e.g. Brown and Warner, 1985; Boehmer et al., 1991), the variance of returns might increase during the days surrounding the event. According to Brown and Warner (1985), the underestimation of this event-induced variance might lead to rejecting the null hypothesis too often. Therefore, I conduct a cross-sectional test as introduced by Boehmer et al. (1991) when there is event-induced variance. This test assumes that the residuals are not correlated, but it ensures that event-induced variance does not have an influence on the results in this model. The formula for this test can be found in appendix C.2.
Next to the market and risk adjusted return model, I will conduct tests based on the Constant Mean Return model and the Market Adjusted Return to check whether my results are robust to other methodological choices. Brown and Warner (1985) use the following formula for the Constant Mean Return Model:
Where , representing the average return of stock j, can be calculated with the following formula:
∑
Brown and Warner (1985) use the following formula for the Market Adjusted Return Model:
Non-parametric Tests
Parametric tests should only be conducted when the data are normally distributed. When they are not normally distributed, the distribution of the sample is not representative. Non-parametric tests, however, do not rely on the assumption of normally distributed data and can therefore be used as robustness check. One widely used non-parametric test is Corrado’s rank test (e.g. Raymond, 2008). This rank test is expected to be correctly specified, no matter how skewed the distribution of the data is (Corrado, 1989).
The first required step of the test is the transformation of the abnormal returns of each share into ranks over the both the estimation and event period of [-195,+10]. The lowest value of this period of 206 days will get rank 206, whereas the highest abnormal return of each company will be rank 1. This means that the average rank is 103.5. Based on this information, the test statistic U for each day can be calculated by means of the following formula:
U
∑
√
Where Kjt is the rank of the abnormal return for stock j during the event and estimation window of
[-195,+10].
and where σ K can be calculated by using the following formula:
∑
(
∑
)
However, when testing for cumulative abnormal returns, Serra (2004) suggests using the following formula, that is derived from Corrado’s test:
U
∑ ∑√∑
Testing for differences
I will also test whether differences between subsample means are significant. Since the subsamples are unrelated, they can considered to be independent. However, the population variances and
are unknown and therefore estimating the standard error of the distribution of the sample is necessary (Keller, 2006). The preferred method to calculate this depends on whether the unknown variances of the population are equal or not. In order to test this, I will conduct Levene’s F-Test for Equality of Variances (See Appendix C.3). According to Al Jafari (2011), this test is the most used statistic for testing for differences between means, since it does not require the data to be normally distributed (Levene, 1960). Based on the outcome of this test, I will either conduct an Equal Variances test or an Unequal Variances Test, as has been suggested by Keller (2005).
Equal Variances Test:
̅
̅
√
Where ̅ and ̅ refer to the means of the two subsamples and where and represent the hypothesized means of the subsamples. Moreover, and represent the sizes of the subsamples and is the pooled variance estimator that can be calculated with the following formula:
Unequal Variances Test:
̅ ̅
√
Non-parametric test
Next to these parametric mean tests, I will also conduct the non-parametric Mann-Whitney-U test, which tests for differences between the medians. In order to conduct this test, I first arrange the data of the two groups from low to high in a single ranked set (Keller, 2005). Next, I calculate the score of the Mann-Whitney-U test with the following formula:
∑
Where U is the statistic of the Mann-Whitney-U test, where N1 and N2 represent the number of firms
Multivariate Analysis
The next step will be to investigate what characteristics of the Sovereign Wealth Fund and the target firm are related to the reaction of the market on the announcement of a transaction by a Fund. In order to investigate this, I will conduct a multivariate cross-sectional regression that is based on the variables that have been introduced in chapter 2. The dependent variable will be the cumulative abnormal returns during the [0,+1] window. In line with other research (e.g. Raymond, 2008; Dewenter et al., 2010), I will use the cumulative abnormal returns instead of the standardized cumulative abnormal returns.
This first variable is Stake, which refers to the percentage stake a Sovereign Wealth Fund acquires or divests. The Variable Stake2 measures whether there is a nonlinear relation between the acquired (divested) stake and the dependent variable, as has been suggested by Dewenter et al. (2010).
The variables Size, Net Debt Ratio, Sector, and Floating Shares are related to the characteristics of the target firm. Size refers to the size of the target company and is measured by the log of the total market capitalization in US Dollars, as has been suggested by other research on Sovereign Wealth Funds (e.g. Kotter and Lel, 2011). The variable Net Debt Ratio (defined as debt minus cash, divided by book assets, as suggested by Bates et al. (2009)) will be used to measure the impact of the free cash flow on the dependent variable.6 Sector is a dummy variable that measures the sector the target company is active in. To construct this variable, I assign firms that operate in a strategic sector a 1, whereas companies that do not operate in a strategic sector are assigned a 0.7
Moreover, the variable Floating Shares measures the percentage of shares that is not being held by blockholders.
Next, the variables Home Market and Transparency are related to the Sovereign Wealth Fund characteristics. Home Market is a dummy variable, where 0 and 1 represent respectively a domestic and a foreign Sovereign Wealth Fund making an investment in a company. Transparency is also a dummy variable and is based on the Linaburg-Maduell Transparency Index8, as has been suggested by literature on Sovereign Wealth Funds (e.g. Fotak et al., 2009; and Miceli, 2011)9. This
6 According to Jensen (1986), higher levels of debt reduce the conflicts between agents and shareholders.
These conflicts might arise of high free cash flows. Net debt takes into account the levels of cash and the levels of debt and is therefore a suitable proxy for the free cash flow.
7 In line with Knill et al. (2008), I assume utilities, financials, and telecom firms to be operating in strategic
sectors.
8
Sovereign Wealth Fund Institute (2011), retrieved from < http://www.swfinstitute.org/statistics-research/linaburg-maduell-transparency-index/>
9 Other literature on Sovereign Wealth Funds (e.g. Dewenter at al., 2010; Kotter and Lel, 2011) has used a
Index, which was developed at the Sovereign Wealth Fund Institute by Carl Linaburg and Michael Maduell, is based on ten principles; each principle that holds adds one point to the Transparency Index. The maximum rating a Fund can receive is therefore 10, while the minimum rating is 1. An overview of the principles used by Linaburg and Maduell can be found in Appendix B.1. The creators of this Index recommend that a Fund should have a minimum transparency rating of 8 in order to claim that the Fund is transparent. Therefore, I assign a 0 to the Funds that have a transparency level that is lower than 8, whereas I assign a 1 to Funds that have a transparency level of 8 or higher. This leads to the following multivariate cross-sectional regression:
Where:
CARjt is the cumulative abnormal return during the [0,+1] window for company j
at time t.
ASj is the percentage stake a Fund has acquired or divested in target firm j.
AS2j is the squared stake a Fund has taken or divested in target firm j.
Sjt is the size of company j, measured by the log of the total market
capitalization in US Dollars, on day t.
NDjt is the Net Debt Ratio for company j on day t.
FSjt-1 is the amount of floating shares not being held by large shareholders on the
day before the announcement of the transaction (t-1).
HMj is a dummy variable that takes a value of 1 for a domestic Sovereign Wealth
Funds making a transaction in company j and that takes a value of 0 for a foreign Sovereign Wealth Funds undertaking a transaction in firm j.
Tjt is a dummy variable taking a value of 0 for opaque Funds investing or
divesting in firm j at time t, whereas transparent Funds take a value of 1.
is the random sampling error for firm j at time t.
α,β,γ,δ,ζ,η,θ,ι represent the coefficients of the regression.
Heteroskedasticity
3.2 Data and Descriptive Statistics
In order to test the hypotheses, I create a new dataset of Sovereign Wealth Fund investments and divestments for the period January 2004 – July 2011. I have chosen for this period for two reasons. Firstly, the focus of the currently available papers on the investments of Sovereign Wealth Funds has been on the period up to 2008; the most recent years have not been researched yet, despite the severe changes in the global economy that also impacted the Sovereign Wealth Fund investments (Monitor Group, 2010b). Secondly, according to Kotter and Lel (2011), 75% of the Sovereign Wealth Fund announcements they found dated from 2004 onwards. Consequently, the years up to 2004 provide less data and are therefore left out of the sample to be used in the research.
The first step is to retrieve the list of Sovereign Wealth Funds from the Sovereign Wealth Fund Institute10, thereby focusing on the Funds with a capitalization of over $1 billion, as suggested by Fernandes (2011). The 23 Sovereign Wealth Funds in my sample represent more than 99 percent of the value of the Sovereign Wealth Fund universe (Appendix D.1).11
Next, I identify investments in firms with publicly traded equity by manually searching both LexisNexis and Zephyr, using the name of each Sovereign Wealth Fund and keywords like investment and stake12. For each investment I collect information on the announcement date of the investment,
the rumor date, the amount invested and the acquired share.
This results in 288 events, but several filters are applied to the transactions to get a clean sample (See Appendix D.2). Firstly, investments in initial public offerings are left out of the sample, because the impact of the independent Sovereign Wealth Fund investment cannot be determined when a company issues shares to the market (Kotter and Lel, 2011). Secondly, as suggested by Dewenter et al. (2010), I treat simultaneous Sovereign Wealth Fund investments (e.g. consortium investments) in the same company as one event. For example, the Qatar Investment Authority and Temasek (Singapore) together announced the purchase of a stake in Singapore-based Raffles Medical Group. Of the remaining 276 events, stock price data of 44 targeted companies are not available on either Thomson Reuters DataStream or Google Finance. The clean sample then consists of 232 investments in 190 unique companies, in 40 different countries13.
To identify the Sovereign Wealth Fund divestments in firms with publicly traded equity, I manually searched LexisNexis and Zephyr, using the Sovereign Wealth Fund name and keywords like divestment and sale. This resulted in 116 divestments. However, stock price data for 15 events were
10
Sovereign Wealth Fund Institute, http://www.Sovereign Wealth Fundinstitute.org/
11
Idem
12 Several Sovereign Wealth Funds have wholly owned subsidiaries that undertake transactions on behalf of the
Funds. Therefore I have also included the names of these subsidiaries in my search for transactions.
13
not available on DataStream or Google Finance. The clean sample then consists of 101 events in 88 unique companies, in 18 different countries (See Appendix D.2).
Moreover, I retrieve the specific data on the Sovereign Wealth Funds from the Sovereign Wealth Fund Institute.14 This institute provides information on the backgrounds of the Funds, including the country of origin and the level of transparency. However, data on the transparency is not available for three Sovereign Wealth Funds. Additionally, I retrieve data for the firm characteristics from DataStream. For the investment sample, the net debt ratio is not available for 22 companies, the market capitalization is not available for 15 companies, and the information on the floating shares is not available for 26 firms. Also, information on the acquired stake is missing for 19 target firms. For the divestment sample, the net debt ratio is not available for 17 firms, the market capitalization is not available for 2 companies, while the data on floating shares is not available for 15 firms. Information on the divested stake is missing for 6 companies.
Table 2: Sovereign Wealth Fund Investments and Divestments
This table presents the country or origin and the level of transparency for 26 Sovereign Wealth Funds that made investments or divestments in public companies all over the world during the years 2004 - 2011. Additionally, this table presents information on the number of (domestic) investments and (domestic) divestments that were undertaken by the Funds. These transactions were identified by manually searching both LexisNexis and Zephyr, using the name of each Sovereign Wealth Fund and its well-known wholly owned subsidiaries, and keywords like investment and stake. The information on the Sovereign Wealth Fund was retrieved from the Sovereign Wealth Fund Institute Web Site at http://www.swfinstitute.org.
Sovereign Wealth Fund Country Transparency (Domestic)
Investments
Divestments (Domestic)
Abu Dhabi Investment Authority Abu Dhabi 4 8 2
Australia Future Fund Australia 10 0 1 (1)
China-Africa Development Fund China 4 2 0
China Investment Corporation China 7 16 (3) 3 (1)
Fonds Strategique d'Investissement
France NA 15 (14) 2 (2)
Fundo Soberano do Brasil Brazil NA 1 (1) 0
Government of Singapore Investment Corporation
Singapore 6 31 15
Government Pension Fund Norway 10 25 (21) 22 (16)
International Petroleum Investment Company
Abu Dhabi 4 11 (1) 2
Khazanah Malaysia 5 19 (13) 9 (9)
Korea Investment Corporation South Korea 9 3 (1) 0
Kuwait Investment Authority Kuwait 6 5 (1) 5 (2)
Libyan Investment Authority Libya 2 5 0
Mubadala Abu Dhabi 10 5 (3) 0
Mumtalakat Bahrain 9 1 (1) 1
National Pension Reserve Fund Ireland 10 2 (2) 0
National Social Security Fund China 5 4 (4) 4 (4)
New Zealand Super Fund New Zealand 10 2 (2) 0
Oman Investment Fund Oman NA 3 0
Public Investment Fund Saudi Arabia 4 1 (1) 0
Qatar Investment Authority Qatar 5 17 (3) 2
SAFE Investment Company China 2 5 1 (1)
SAMA Saudi Arabia 4 1 0
Samruk-Kazyna Kazakhstan 6 1 0
State General Reserve Fund Oman 1 1 0
Temasek Holdings Singapore 10 48 (7) 32 (15)
Table 2 provides an overview of the Sovereign Wealth Funds represented in the sample, their country of origin, the number of investments and divestments. The Singaporean Funds Temasek and the Government of Singapore Investment Corporation are heavily represented in my dataset: together they made 79 investments and represent 34% of the total Sovereign Wealth Fund investment sample. Other active Funds are the Norwegian Government Pension Fund (25 investments), the three Funds from Abu Dhabi (together 24 investments), and the Malaysian Fund Khazanah, that made 19 investments. It is also important to note that the five Funds from OECD countries (i.e. from Norway, France, Ireland, South Korea and New Zealand) undertake only 44 investments.
The Singaporean Funds are also heavily represented in my divestment sample: together they account for 47 divestments, representing more than 46% of my sample. Other Funds that divested heavily are the Government Pension Fund (22 divestments) and Khazanah (9 divestments). These four Funds, which are controlled by three countries, made 77% of all the Sovereign Wealth Fund divestments during the period 2004 – 2011.
Appendix D.3 breaks down the investments and divestments by Sovereign Wealth Fund and
target firm country. Almost 50% of the investment and 60% of the divestment transactions take place in Asia and Australia and most of these transactions were undertaken by the Asian Sovereign Wealth Funds. This contradicts suggestions by the media that the Funds invest exclusively in firms from OECD countries. For example, all the 19 investments of the Malaysian Fund Khazanah and 63 of the 79 Singaporean investments were in Asia. Only about 30% of all the transactions take place in Europe; 29 of the investments were undertaken by Sovereign Wealth Funds from the Middle East, while the Asian Funds only invested in 12 European companies. The remaining 42 European investments were undertaken by the Sovereign Wealth Funds from Norway, France, and Ireland. Additionally, the New Zealand Super Fund made both investments in its own country, while the Fund from South Korea invested in its own market, in the United States and in Singapore. This means that the OECD Funds placed all but one investment in OECD markets.
Table 3 presents the data on the characteristics of the Fund itself, and on the target firm characteristics for the investment sample. It becomes clear that the data for variable Size, which measures the market capitalization of the target company, ranges from $5.75 million to $193 billion. In order to reduce this range, I will take the natural logarithm of the total market capitalization, as has been suggested by other research on Sovereign Wealth Funds (e.g. Kotter and Lel, 2011). Next to the descriptive statistics of the variables, I present the descriptive statistics of all single days in the event period for the investments in Appendix D.5.
Table 4 presents the descriptive statistics that are used in the cross-sectional analysis for the divestments. As in the case of investments, the data for the variable Size range from $7.2 million to $230 billion. Therefore, I will also use the natural logarithm of the total market capitalization for the divestment sample in order to reduce the range of this variable. Next to the descriptive statistics for the variables that will be used in the multivariate analysis, I present the descriptive statistics of all single days in the event period for the investments in Appendix D.6.
Moreover, Appendix D.7 shows the distribution of the standardized abnormal returns for both the investment and divestment sample during the estimation period of [-195,-16]. According to Brooks (2008), the data are normally distributed when the Jarque-Bera statistic for normality is not significant. Since the statistics are significant for both the investments and divestments, it can be concluded that the data are not normally distributed. Parametric tests are therefore not appropriate (Brown and Warner, 1980; 1985). However, since both my samples are larger than 100, it can be assumed that they are normally distributed and parametric tests can therefore be used (Newbold, 2003).
Table 3: Descriptive Statististics of the Sovereign Wealth Fund Investments
This table presents the descriptive statistics of the data on the firm characteristics, and of the data on the characteristics of the Fund itself. These transactions were identified by manually searching both LexisNexis and Zephyr, using the name of each Sovereign Wealth Fund and its well-known wholly owned subsidiaries, and keywords like investment and stake. Panel A presents statistics on the characteristics of the firm the Fund invests in. Stake refers to the percentage stake the Fund takes in a company, whereas Stake2 is the squared size of the variable Stake. The Size of the target firm is measured by means of the total market capitalization, which is measured in billion US Dollars.
The variable Net Debt Ratio refers to the amount of debt as percentage of total assets. Sector is a dummy variable that refers to the sector the target firm operates in. In this variable, companies that were assigned a 0 are firms that are not operating in a strategic industry, whereas I assign a 1 to companies that do operate in a strategic sector. Panel B shows the variables that are related to the characteristics of the Sovereign Wealth Fund. Home Market is a dummy variable, where 0 and 1 represent respectively a domestic and a foreign Sovereign Wealth Fund making an investment in an company. Transparency, which is based on the Linaburg-Maduell Transparency Index, is also a dummy variable, where 0 represents the Funds that have a transparency level that is lower than 8, and 1 represents Funds that have a transparency level of 8 or higher.
Panel A Observations Mean Median Maximum Minimum
Stake 213 0.0937 0.05 1 0.0001
Stake2 213 0.0292 0.0025 1 0.0000
Size (in mln $) 217 11,040 2,229 193,035 5.7500
Net Debt Ratio 210 0.1403 0.1684 0.9446 -0.8804
Strategic Sector 232 0.4784 0 1 0
Floating Shares 206 0.7147 0.7750 1 0.1000
Panel B
Home Market 232 0.3362 0 1 0
Correlation and Multicollinearity
Moreover, I test for correlations between the independent variables that will be used in the multivariate regression. Appendix D.8 and Appendix D.9 present the results of the correlation tests for respectively the investment and divestment sample. For both samples, I find very high correlations between the variables Stake and Stake2, which is not surprising, since the latter variable is derived from the variable Stake. Moreover, in case of the divestment sample I find that the variables Sector and Size are the only ones that have a relatively high correlation of 0.53, which is significant at the 1% level.
Since a few variables show a high correlation, I also test for multicollinearity and the results for these tests can be found in Appendix D.10 and D.11 for respectively the investment and divestment sample. According to Belsley et al. (1980), a general rule of thumb is that there is no multicollinearity when the Variation Inflation Factors does not exceed 10. Since this is the case for all the variables, I do not have to drop any variables.
Table 4: Descriptive Statististics of the Sovereign Wealth Fund Divestments
This table presents the descriptive statistics of the data on the firm characteristics, and of the data on the characteristics of the Fund itself. The divestments were identified by manually searching both LexisNexis and Zephyr, using the name of each Sovereign Wealth Fund and its well-known wholly owned subsidiaries, and keywords like
investment and stake. Panel A presents statistics on the characteristics of the firm the Fund invests in. Stake refers to
the percentage stake the Fund takes in a company, whereas Stake2 is the squared size of the variable Stake. The Size
of the target firm is measured by means of the total market capitalization, which is measured in billion US Dollars. The variable Net Debt Ratio refers to the amount of debt as percentage of total assets. Sector is a dummy variable that refers to the sector the target firm operates in. In this variable, companies that were assigned a 0 are firms that are not operating in a strategic industry, whereas I assign a 1 to companies that do operate in a strategic sector. Panel B shows the variables that are related to the characteristics of the Sovereign Wealth Fund. Home Market is a dummy variable, where 0 and 1 represent respectively a domestic and a foreign Sovereign Wealth Fund making an investment in an company. Transparency, which is based on the Linaburg-Maduell Transparency Index, is also a dummy variable, where 0 represents the Funds that have a transparency level that is lower than 8, and 1 represents Funds that have a transparency level of 8 or higher.
Panel A Observations Mean Median Maximum Minimum Stake 95 0.0656 0.0230 0.6968 0.0002
Stake2 95 0.0172 0.0005 0.4855 0.0000
Size (in mln $) 99 14,908 3,196 229,981 7.220
Net Debt Ratio 84 0.1349 0.1706 0.4986 -0.6112
Strategic Sector 101 0.4752 0 1 0
Floating Shares 86 0.6485 0.7150 1 0.0400
Panel B
Home Market 101 0.5050 0 1 0
4. Results and Discussion
This chapter discusses the outcomes of the tests that are conducted to test the hypotheses that were introduced in chapter 2. Firstly, I show the results of tests on the total investment and divestment samples and will present robustness checks on these tests. The second part deals with testing subsamples of the investment and divestment samples. The third section discusses the cross-sectional regression analyses that further examine the hypotheses that were introduced in chapter 2.
4.1 Announcement Period Returns
Figures 1 and 2 give an overview of the impact of respectively Sovereign Wealth Fund investments and divestment announcements on the share price of a target firm. Both figures show that on average the market reacts positively on the day a Sovereign Wealth Fund announces an acquisition, while the price of the share falls on the announcement date of a divestment. An overview of the daily average abnormal returns for both samples can be found in Appendix E.1.
Figure 1. This graph plots the average abnormal returns of a firm during the days surrounding the announcement of a
Sovereign Wealth Fund investment at event date 0. The full sample consists of 232 observations during the period January 2004 - July 2011. Daily abnormal returns are used in the Market and Risk Adjusted Returns model in which the MSCI World Index is the market proxy.
Figure 2. This graph plots the average abnormal returns of a firm during the days surrounding the announcement of a
Sovereign Wealth Fund divestment at event date 0. The full sample consists of 101 observations during the period January
-0,80% -0,60% -0,40% -0,20% 0,00% 0,20% 0,40% 0,60% 0,80% 1,00% -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 A b n o rm al R e tu rn s
Days Event Period
-0,80% -0,60% -0,40% -0,20% 0,00% 0,20% 0,40% 0,60% -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 A b n o rm al R e tu rn s
