Economics of Sovereign Wealth Funds
Chapter

Chapter 13 Sovereign Wealth Funds and Financial Stability: An Event-Study Analysis

Author(s):
Udaibir Das, Adnan Mazarei, and Han Hoorn
Published Date:
December 2010
Share
  • ShareShare
Show Summary Details

Since the beginning of the global financial crisis in the summer of 2007, financial stability has been at the forefront of policy discussions. Sovereign wealth funds (SWFs) have become dominant players in global financial markets over the same period because they have injected significant capital into major financial institutions. In some countries, SWFs were instructed by their governments to invest in domestic financial institutions and the stock markets to support battered stock prices. Research on the financial-stability implications of these funds has been slowly emerging, hampered by lack of data on SWFs’ asset allocations.

Many arguments have been put forth about the potential positive and negative effects of SWFs on global financial markets. Some argue that SWFs can play a stabilizing role in global financial markets. First, many commentators point out that as long-term investors with no imminent call on their assets, and with mainly unleveraged positions, SWFs are able to sit out longer during market downturns or even go against market trends. In addition, SWFs in some countries, particularly in the Middle East, have recently supported domestic equity markets and financial institutions. Second, large SWFs may have an interest in pursuing portfolio reallocations gradually, to limit any adverse price effects of their transactions. Third, SWFs could, as long-term investors and by adding diversity to the global investor base, contribute to greater market efficiency and lower market volatility. Fourth, SWFs’ investments may enhance the depth and breadth of markets.

Although SWFs appear to have been a stabilizing force thus far, as a result of their size, there are circumstances in which they could cause volatility in markets. Having large and often unclear positions in financial markets, SWFs—like other large institutional investors—have the potential to cause market disturbances. For instance, actual or rumored transactions may affect relative valuations in particular sectors and result in herd behavior, adding to volatility. Deeper markets, such as currency markets, can also be affected, at least temporarily, by rumors or announcements about changes in currency allocations by central banks or SWFs. To the extent that SWFs invest through hedge funds that rely on leverage or are subject to margin requirements, such investments may inadvertently magnify market changes. For markets to absorb flows from any major investor class without large price fluctuations, they need to be able to anticipate the broad allocation and risk-preference trends of such investor classes. Opacity about such trends can lead to inaccurate pricing and volatility. Both theoretical and empirical research are underway with regard to the financial-stability implications of SWFs.

Recent capital injections by SWFs have further intensified the debate about the impact of SWFs on financial stability. SWFs from East Asia and the Middle East were frequently in the news, as major mature-market financial institutions required additional capital. In total, SWFs have contributed more than US$50 billion of such capital since November 2007. The capital injections by SWFs have augmented the recipient financial institutions’ capital buffers and have been helpful in reducing various firm-specific risk premiums, at least in the short term, because the injections curtailed the need to reduce bank assets to preserve capital. The announcements of capital injections from SWFs have assisted in stabilizing share prices and the elevated credit default swaps spreads, at least over the short run (IMF, 2008). In most cases, after the announcement of new SWF capital injections, the initial share price reaction was positive, with announcements of asset write-downs offset by concurrent capital injections from investor groups in which the SWFs had significant roles. Although other factors are not taken into account, this initial evidence supports the view that SWFs could have a volatility-reducing impact on markets.1

This chapter, using an event-study approach based on a hand-collected database, endeavors to deepen the analysis of SWFs’ impact on financial stability by differentiating between various scenarios comprising investments and divestments in advanced and emerging economies, financial and nonfinancial sectors, and SWFs with higher and lower levels of corporate governance. The overall findings suggest SWFs have no significant destabilizing effect on equity markets. This empirical study contributes to the emerging academic literature that seeks to analyze the behavior of SWFs in financial markets.

The chapter proceeds as follows: The first section briefly reviews the literature and some conceptual issues. The second section outlines the event-study approach and describes the data. The third section presents empirical results, and is followed by a concluding section.

LITERATURE REVIEW

SWFs are defined as special-purpose investment funds or arrangements owned by the government. They are often established out of balance of payments surpluses, official foreign currency operations, proceeds of privatizations, fiscal surpluses, or receipts from commodity exports. Their total size has been estimated at US$2 trillion to US$4 trillion,2 but many of them have probably seen large unrealized losses from the crisis combined with a sharp reduction in oil prices. These unrealized losses have been higher for SWFs with higher shares of equities in their investment portfolios or large, illiquid positions in private-equity or hedge funds. Given that SWFs typically have fairly long investment horizons, they are likely to sit out these unrealized losses.

The lack of publicly available data on SWF asset allocations has led to a strand of IMF research on the theory side. Lam and Rossi (2010) develop a theoretical model that aims to examine the impact of SWFs on global financial stability during periods of stress. Their findings indicate that SWFs have a risk-sharing role in financial markets. As part of the IMF-coordinated process of the Santiago Principles, which provide generally accepted principles and practices for SWFs, Hammer, Kunzel, and Petrova (2008) examine the asset-allocation and risk-management frameworks of SWFs based on a detailed survey. The results show that SWFs have specific investment objectives in place, adopt asset approaches (mean-variance style) in determining their asset allocation strategies, use common risk measures (e.g., credit ratings, value-at-risk models, tracking errors, duration, and currency weights) for risk management, and have explicit limits in their investment classes and instruments.

Simulations of SWFs’ asset allocations have been undertaken by Kozack, Laxton, and Srinivasan (see Chapter 14 of this book). Specifically, they create two stylized diversified portfolios, one mimicking Norway’s SWF (the Government Pension Fund—Global) and the other representing some well-established SWFs. They then conduct a scenario analysis of the effect of diversification of sovereign assets. Although the calibrations are highly sensitive to the underlying model assumptions, the findings indicate that advanced economies will see lower capital inflows, while emerging-market countries will be the primary beneficiaries. The quantitative results of this work are consistent with the back-of-the-envelope calculations of Beck and Fidora (2008), which imply a net capital outflow from the United States and the euro area and net inflows to emerging-market countries. In the same vein, Miles and Jen (2007) and Hoguet (2008) point out that there is scope for the global equity risk premium to fall and for real bond yields to rise if SWFs allocate their assets to equities. In addition, as SWFs increasingly diversify into global portfolios, their activities may place some pressure on the dollar (see Chapter 15 of this book).

Some empirical research using equity-market indicators and an event-study approach has examined the role of SWFs as major institutional investors. For instance, in an event study, Chhaochharia and Laeven (2008) find that the announcement effect of SWF investments is positive. They report that share prices of firms respond favorably when SWFs announce investments, partly because these investments often occur when firms are in financial distress, but the long-run performance of equity investments by SWFs tends to be poor (see Fotak, Bortolotti, and Megginson, 2008, for similar results). Kotter and Lel (2008) show that the cumulative abnormal return of SWF investments has an announcement effect similar to that of investments by hedge funds and institutional investors such as the California Public Employees’ Retirement System on stock returns. In addition, investments by more transparent SWFs have a 3.5 percent larger cumulative abnormal return, suggesting that voluntary SWF disclosure might serve as a signal device to investors. In addition, Kotter and Lel (2008) also obtain a significant negative, but small, announcement impact from SWFs’ divestitures. Beck and Fidora (2008) conducted a country case study of Norway’s SWF and asked whether its exclusion of companies that violate the ethical guidelines of the Ministry of Finance exerts price pressures on those companies. Their findings suggest no significant negative abnormal returns following the divestiture by the SWF of these companies.

To summarize, existing research on SWFs suggests that they can be a stabilizing force in global financial markets. Event studies do not find a destabilizing impact from SWF investments and divestments in equity markets, while simulations of SWF asset allocations imply only a gradual shift with modest economic effects. With SWFs improving their transparency and disclosure procedures over time, the availability of historical data on SWF transactions would provide researchers with the necessary information to examine further their implications for financial stability.

DATA AND METHODOLOGY

The empirical research conducted for this chapter uses an event-study approach to assess whether stock markets react to SWFs’ announcements of investments in and divestments of firms. The objective is to investigate the information content of these announcements. Based on 166 publicly traceable, hand-collected events of investment and divestment (both announcement and action) by major SWFs during 1990 through 2009, this section evaluates the short-term financial impact of SWFs on selected public equity markets in which they invest and divest. The impact is further analyzed by different sectors (financial and nonfinancial), actions (buy and sell), market types (developed and emerging), and level of corporate governance of the SWF (higher and lower scores). The results are expected to suggest how stock markets react to capital investments and divestments by SWFs and present implications of SWFs’ actions for stability in global financial markets. The investigation of divestments is of particular interest because abnormally large stock price reactions (relative to the market) may be destabilizing to the degree that others mimic or act on SWFs’ divestment behavior. Such actions might be particularly troublesome if prices slip below other investors’ predefined target levels, thus prompting forced sales of their positions.

Data

Several SWFs that have bought or sold shares of firms in the advanced and emerging stock markets are included in the study. Among them are the Abu Dhabi Investment Authority, the China Investment Corporation, the Government of Singapore Investment Corporation (GIC), the Kuwait Investment Authority, the Korea Investment Corporation, the Libyan Investment Authority, Mubadala (in Abu Dhabi), the Qatar Investment Authority, and Temasek (in Singapore). Information on the events was gleaned from the SWFs’ Web sites and various financial news reports such as Factiva, a division of Dow Jones & Company. Target firm actual total returns (and price indices) and country stock market returns (and price indices) were obtained from the Datastream International database.3 This search resulted in a total of 166 investment and divestment events in 115 unique firms, with some firms receiving multiple SWF investments between 1990 and 2009.4 This information was then combined with firm-level and country-level data collected from Bloomberg, and SWF-specific data from various sources, including Truman (2008).5

Table 13.1 shows the number of SWF investments and divestments by country of target firm. Table 13.2 displays the distribution of the sample by the identity of the acquiring SWF, showing that Singapore’s two SWFs (GIC and Temasek) dominate the sample. Figure 13.1 shows the ratios of SWFs’ investments (buy) and divestments (sell) to the full sample as well as similar ratios for the subsamples—financial and nonfinancial sectors, developed and emerging markets, and SWFs with high and low levels of corporate governance.

TABLE 13.1Country of Target Firms
CountryNumber of events
Australia6
Austria1
China17
Egypt2
France8
Germany7
Iceland1
India13
Indonesia5
Italy6
Japan2
Korea, Rep. of3
Malaysia7
Pakistan4
Philippines1
Portugal2
Singapore22
Spain3
Sweden2
Switzerland2
Taiwan Province of China1
Thailand2
United Kingdom31
United States17
Vietnam1
Total166
Source: Authors’ compilations.
Source: Authors’ compilations.
TABLE 13.2Events by Acquiring SWF
SWFNumber of EventsCountry
Abu Dhabi Investment Authority26United Arab Emirates
China Investment Corporation11China
Government of Singapore Investment Corporation38Singapore
Kuwait Investment Authority14Kuwait
Korea Investment Corporation1Korea, Rep. of
Libyan Investment Authority2Libya
Mubadala2United Arab Emirates
Qatar Investment Authority23Qatar
Temasek49Singapore
Total166
Source: Authors’ compilations.
Source: Authors’ compilations.

Figure 13.1Ratios of SWF Investments and Divestments

Sources: SWF Web sites; Factiva; authors’ calculations.

Note: The SWFs with high-level corporate governance refer to those whose total score is higher than 40; low level refers to a total score equal to or lower than 40 (Truman, 2008).

Methodology

If markets are rational, the effects of an event should be reflected immediately in stock returns and prices. Thus, a measure of an event’s impact can be constructed using stock prices and returns observed over a relatively short period. To benchmark the returns of the stock relative to the event, the overall stock market returns, in percentage changes, for the country corresponding to the target firm are used. Specifically, the following steps were taken to implement the event study:

  • Determination of the selection criteria for the inclusion of SWFs. SWFs were chosen if data were available on their actions, and on stock prices and total returns. The sample contains several SWFs that have bought or sold stakes in financial firms and nonfinancial firms.

  • Collection of a number of investment and divestment events and compilation of a list of firms and event dates by searching publicly available databases to find news announcements of SWFs’ actions.

  • Identification of the event window for SWFs’ investments and divestments. Because the event date can be determined with precision, the short-term analysis employed a five-day (seven-day) event window, comprising two (three) preevent days, the event day, and two (three) postevent days.6 Inclusion of the preevent days allowed rumors that preceded the formal announcement to enter the assessment. Postevent days were also included because prices in illiquid markets may take a couple of days to adjust to new information. To test for robustness, the event window was expanded to nine days: four preevent days, the event day, and four postevent days.

  • Definition of the estimation period. Following Peterson’s framework (1989), the market model was estimated based on the 200 trading days ending 30 days before the announcement of each investment or divestment. Ending the sample before the event ensures that the “normal” behavior of returns is not contaminated by the event itself. To test for robustness, the estimation periods were varied (separate estimations at 100 days and 300 days), and price indices were used instead of total returns to each firm and the economy.

  • Estimation of a normal return during the event window in the absence of the event, using a one-factor ordinary least squares regression equation:7

  • where Rit is the return of stock i in period t, Rmt is the overall stock market returns in period t, αiand βi are regression coefficients, and eit is an error term, i and t are individual stocks and time, respectively.

  • Calculation of the abnormal return within the event window. Having calculated estimates of αi and βi with the data from the estimation period, abnormal returns (ARit) were calculated by differencing the actual and estimated returns,

where Rit* is the estimated return.

The abnormal return observations for each day in the event window must be aggregated to draw overall inferences for the event of interest. The aggregation can be along two dimensions—through time and across securities—simultaneously.

The individual securities’ abnormal returns, in the 5-day case, can be aggregated for each event day, t = –2, –1, 0, +1, +2, during the event window. Given N events (a total of 166 in the entire sample), the sample aggregated average abnormal returns (AAR) for period t is

  • The aggregated average abnormal returns can then be aggregated over the event window to calculate the cumulative average abnormal return (CAAR).

  • Testing whether the abnormal return is statistically different from zero. Because the number of observations in the event window (5 or 7 days) is limited, t-tests were used rather than the Z score, because Z scores usually require at least 50 observations to obtain statistically robust results.8

EMPIRICAL RESULTS

Table 13.3 presents the AAR and CAAR for the (—2, +2), and (—3, +3) windows. In general, the AAR is positively associated with SWFs’ buy actions and not significantly negatively associated with SWFs’ sell actions in the full sample. The results also suggest that the share-price responses to SWFs’ investments in developed economies are statistically significant at the 5 percent confidence level, while those in emerging economies are not. In addition, SWFs’ investments in the financial sector have a larger impact on share prices than do their investments in the nonfinancial sector. These differences in responses may be due to the relatively more liquid equity markets in developed economies and in the financial sector.

TABLE 13.3Stock Market Reactions to Announcements of SWF Investments and Divestments(Total Returns in Percent)
Event windowt-statistic of AARMean of AARt-statistic of CAARCAAR
Panel A: Buy only, 134 events from 101 firms
(–2, +2)4.31**0.273.33**0.77
(–3, +3)3.75**0.224.45***0.96
Panel B: Sell only, 32 events from 23 firms
(–2, +2)0.000.00–0.31–0.07
(–3, +3)–0.08–0.02–1.21–0.19
Panel C: Buy and sell in developed economies only, 87 events from 55 firms
(–2, +2)4.29**0.214.95**0.72
(–3, +3)2.88**0.186.17***0.94
Panel D: Buy and sell in emerging economies only, 79 events from 60 firms
(–2, +2)1.200.171.670.34
(–3, +3)0.980.111.140.20
Panel E: Buy in developed economies only, 72 events from 51 firms
(–2, +2)5.47**0.304.44**0.91
(–3, +3)3.13**0.245.82***1.21
Panel F: Sell in developed economies only, 15 events from 9 firms
(–2, +2)–0.49–0.19–1.24–0.44
(–3, +3)–0.56–0.16–2.99**–0.77
Panel G: Buy in emerging economies only, 62 events from 50 firms
(–2, +2)2.070.242.31*0.60
(–3, +3)2.37*0.202.94**0.69
Panel H: Sell in emerging economies only, 17 events from 14 firms
(–2, +2)0.440.160.990.23
(–3, +3)0.370.091.620.29
Panel I: Buy in financial sector only, 41 events from 24 firms
(–2, +2)2.72*0.705.13**2.53
(–3, +3)3.09**0.666.38***3.40
Panel J: Sell in financial sector only, 5 events from 3 firms
(–2, +2)
(–3, +3)
Panel K: Buy in nonfinancial sector only, 93 events from 77 firms
(–2, +2)0.310.04–1.78–0.18
(–3, +3)–0.15–0.01–4.06**–0.35
Panel L: Sell in nonfinancial sector only, 27 events from 20 firms
(–2, +2)0.000.00–0.31–0.07
(–3, +3)–0.08–0.02–1.21–0.19
Panel M: Buy by high-level governance SWF only, 76 events from 59 firms
(–2, +2)–0.11–0.02–0.20–0.02
(–3, +3)0.180.020.840.07
Panel N: Sell by high-level governance SWF only, 26 events from 19 firms
(–2, +2)0.360.120.850.17
(–3, +3)0.270.061.040.15
Panel O: Buy by low-level governance SWF only, 58 events from 45 firms
(–2, +2)2.68*0.683.03**1.89
(–3, +3)2.21*0.504.05**2.22
Panel P: Sell by low-level governance SWF only, 6 events from 4 firms
(–2, +2)–1.23–0.73–1.96–1.61
(–3, +3)–1.15–0.53–3.67**–2.39
Source: Authors’ estimates.Note: — in panel J means there were no qualified observations before or after the corresponding event dates. *** significant at 1 percent level; ** significant at 5 percent level; * significant at 10 percent level.
Source: Authors’ estimates.Note: — in panel J means there were no qualified observations before or after the corresponding event dates. *** significant at 1 percent level; ** significant at 5 percent level; * significant at 10 percent level.

Different scenarios were tested using these events. Panel A of Table 13.3 reports the AAR and CAAR during the event windows of SWF investments for the entire sample of 134 investment observations during the period 1990 to 2009. The mean AAR is 0.27 percent and 0.22 percent for windows of (–2, +2), and (–3, +3), respectively, around the announcement date, and the CAAR is 0.77 percent and 0.96 percent, respectively. The sign-test statistics for the AAR are also highly significant for the two windows. Panel B reports the AAR and CAAR during the event windows of the announcements of SWF divestments for the entire sample of 32 observations during the period between 1990 and 2009. The mean AAR is 0 percent and –0.02 percent for the windows (–2, +2), and (–3, +3), respectively, and the CAAR is –0.07 percent and –0.19 percent, respectively. The sign-test statistics for the AAR and the CAAR are insignificant for the two windows.

Panel C reports the AAR and CAAR during the event windows of SWF investments and divestments for the developed-economy sample of 87 observations during the period between 1990 and 2009. The mean AAR is 0.21 percent and 0.18 percent for the windows (–2, +2), and (–3, +3), respectively, and the CAAR is 0.72 percent and 0.94 percent, respectively. The sign-test statistics for the AAR and the CAAR are highly significant for the two windows. Panel D of Table 13.3 reports the AAR and CAAR during the event windows of SWF investments and divestments for the emerging-economy sample of 79 observations during the period between 1990 and 2009. The mean AAR is 0.17 percent and 0.11 percent for the windows (–2, +2), and (–3, +3), respectively, and the CAAR is 0.34 percent and 0.20 percent, respectively. The sign-test statistics for the AAR and CAAR are insignificant for the two windows.

The impact is further analyzed on the investments and divestments in different market types (developed and emerging), different sectors (financial and nonfinancial), and level of corporate governance of the SWF (high and low). In general, according to the AAR, investments in developed economies (Panel E) and in the financial sector (Panel I) are statistically significant. In addition, the positive (negative) impact of the CAAR for the investments (divestments) by low-level governance SWFs are significantly larger than those by high-level governance SWFs. This could indicate that the improvement of corporate governance in SWFs would be helpful in reducing the impact on market volatility.9

The event window of (–4, +4) was used as a robustness check to test the impact of SWFs’ actions. In addition, the estimation periods were varied to 100 and 300 days. Finally, price indices for each firm and the economy were used instead of total returns. The results were robust to different event windows, different estimation periods, and the use of price indices. (See Table 13.4 for the results using price indices.)

TABLE 13.4Stock Market Reactions to Announcements of SWF Investments and Divestments(Using Price Indices, Returns in Percent)
Event windowt-statistic of AARMean of AARt-statistic of CAARCAAR
Panel A: Buy only, 134 events from 101 firms
(–2, +2)4.09**0.263.46**0.75
(–3, +3)3.84**0.224.71***0.98
Panel B: Sell only, 32 events from 23 firms
(–2, +2)0.240.071.320.26
(–3, +3)0.160.031.670.24
Panel C: Buy and sell in developed economies only, 87 events from 55 firms
(–2, +2)5.05**0.255.45**0.87
(–3, +3)2.83**0.206.37***1.10
Panel D: Buy and sell in emerging economies only, 79 events from 60 firms
(–2, +2)0.980.141.580.28
(–3, +3)0.940.101.330.20
Panel E: Buy in developed economies only, 72 events from 51 firms
(–2, +2)5.5**0.314.44**0.95
(–3, +3)3.22**0.255.66***1.23
Panel F: Sell in developed economies only, 15 events from 9 firms
(–2, +2)–0.12–0.041.020.28
(–3, +3)–0.20–0.050.670.14
Panel G: Buy in emerging economies only, 62 events from 50 firms
(–2, +2)1.940.212.37*0.53
(–3, +3)2.58**0.193.41**0.69
Panel H: Sell in emerging economies only, 17 events from 14 firms
(–2, +2)0.470.171.050.24
(–3, +3)0.400.101.800.32
Panel I: Buy in financial sector only, 41 events from 24 firms
(–2, +2)2.46*0.665.45**2.45
(–3, +3)2.91**0.656.82***3.42
Panel J: Sell in financial sector only, 5 events from 3 firms
(–2, +2)
(–3, +3)
Panel K: Buy in nonfinancial sector only, 93 events from 77 firms
(–2, +2)0.350.04–1.56–0.17
(–3, +3)–0.10–0.01–3.83**–0.35
Panel L: Sell in nonfinancial sector only, 27 events from 20 firms
(–2, +2)0.240.071.320.26
(–3, +3)0.160.031.670.24
Panel M: Buy by high-level governance SWF only, 76 events from 59 firms
(–2, +2)–0.220.04–0.50–0.07
(–3, +3)0.110.020.680.06
Panel N: Sell by high-level governance SWF only, 26 events from 19 firms
(–2, +2)0.380.120.920.18
(–3, +3)0.270.060.910.14
Panel O: Buy by low-level governance SWF only, 58 events from 45 firms
(–2, +2)2.72*0.693.04**1.91
(–3, +3)2.26*0.514.07**2.26
Panel P: Sell by low-level governance SWF only, 6 events from 4 firms
(–2, +2)–0.40–0.231.210.75
(–3, +3)–0.32–0.131.640.88
Source: Authors’ estimates.Note: — in panel J means there were no qualified observations before or after the corresponding event dates. *** significant at 1 percent level; ** significant at 5 percent level; * significant at 10 percent level.
Source: Authors’ estimates.Note: — in panel J means there were no qualified observations before or after the corresponding event dates. *** significant at 1 percent level; ** significant at 5 percent level; * significant at 10 percent level.

CONCLUSION

This chapter uses an event-study approach to assess whether and how stock markets react to SWFs’ announcements and actions of investments and divestments in firms. Based on 166 publicly traceable events collected on investments and divestments by major SWFs over the period 1990–2009, the analysis evaluated the short-term financial impact of SWFs on selected public equity markets in which they invested. The impact was further analyzed for different sectors (financial and nonfinancial), actions (buy and sell), market types (developed and emerging), and level of corporate governance of the SWF (high and low). Overall, this event study did not find any significant destabilizing effect of SWFs on equity markets, which is consistent with the emerging academic literature that uses the event-study methodology.

However, any assessment of the longer-term impact and the potentially stabilizing role of SWFs as major institutional investors will require a broader set of data and a more rigorous empirical assessment. In addition, the long-run impact of SWFs’ investments could be influenced by macroeconomic and financial conditions. SWFs’ recent investments in some U.S. and European financial institutions could not buffer those institutions from further large losses. Therefore, it will be hard to draw conclusions for overall global and regional financial stability, or stability in markets other than equity markets, from these event studies. Other methods to examine the empirical impact of SWFs would require more detailed knowledge of SWFs’ investments and their timing and amount—data that is presently not available. Hypothetical market responses to SWFs’ investments require a thorough understanding of the way in which asset allocations are constructed and the size, depth, and breadth of the corresponding markets.

REFERENCES

    ArmitageSeth1995“Event Study Methods and Evidence on Their Performance,”Journal of Economic SurveysVol. 9No. 1 pp. 2552.

    BeckRoland and MichaelFidora2008“The Impact of Sovereign Wealth Funds on Global Financial Markets” ECB Occasional Paper 91 (Frankfurt:European Central Bank).

    ChhaochhariaVidhi and LucLaeven2008“Sovereign Wealth Funds: Their Investment Strategies and Performance” CEPR Discussion Paper No. 6959 (London: Center for Economic Policy Research).

    FotakVeljkoBernardoBortolotti and WilliamMegginson2008“The Financial Impact of Sovereign Wealth Fund Investments in Listed Companies” (unpublished; Norman, Oklahoma: University of Oklahoma).

    HammerCorneliaPeterKunzel and IvaPetrova2008“Sovereign Wealth Funds: Current Institutional and Operational Practices,”IMF Working Paper WP/08/254 (Washington: International Monetary Fund).

    HoguetGeorge R.2008“The Potential Impact of Sovereign Wealth Funds on Global Asset Prices,”VisionVol. 3No. 2 pp. 2330.

    International Financial Services London2010“Sovereign Wealth Funds 2010.”IFSL Research.Available via the Internet:http://www.ifsl.org.uk/output/ReportItem.aspx?NewsID=20.

    International Monetary Fund2008“Box 1.2. Do Sovereign Wealth Funds Have a Volatility-Absorbing Market Impact?” in Global Financial Stability Report (WashingtonApril).

    KotterJason and UgurLel2008“Friends or Foes? The Stock Price Impact of Sovereign Wealth Fund Investments and the Price of Keeping Secrets,”International Finance Discussion Papers No. 940 (Washington: Board of Governors of the Federal Reserve System).

    KozackJulieDougLaxton and KrishnaSrinivasan2010“The Macroeconomic Impact of Sovereign Wealth Funds”Chapter 14 in this volume.

    LamRaphael W. and MarcoRossi2010“Sovereign Wealth Funds—Investment Strategies and Financial Stress,”Journal of Derivatives and Hedge FundsVol. 15 pp. 30422.

    MilesDavid and StephenJen2007“Sovereign Wealth Funds and Bond and Equity Prices,”Morgan Stanley ResearchJune1.

    PetersonPamela Drake1989“Event Studies: A Review of Issues and Methodology,”Quarterly Journal of Business and EconomicsVol. 28 (Summer) pp. 3666.

    TrumanEdwin M.2008“A Blueprint for Sovereign Wealth Fund Best Practices” Peterson Institute Policy Brief Number PB08-3 (Washington: Peterson Institute).

With the continuing increase in banks’ losses and write-downs during the subprime crisis, the rescue of Bear Stearns, the collapse of Lehman Brothers, and U.S. government intervention into major financial institutions, the longer-term share price development of banks that obtained initial capital injections from various SWFs has been obviously negative. But the short-term reaction to SWFs’ financial support has been perceived as positive by the financial market in most cases.

For instance, a report by International Financial Services, London (2010) revealed that SWFs’ total assets stood at US$3.8 trillion in 2009 and are likely to reach US$5.5 trillion by the end of 2012, a substantially lower estimate than had been made previously.

Datastream is the only data vendor that provides total return stock market indices for all the relevant countries, correcting index returns for dividend payments, stock splits, and other such changes.

Some events were discarded for invested companies that are not publicly listed, and therefore no data for stock price and total return were available.

The corporate governance score of each SWF is from the “total” score in Truman (2008). Those scores higher than 40 were designated “high” in this chapter’s econometric analysis, while those equal to or lower than 40 were designated “low.”

Several event windows (specified in parentheses) were chosen to test robustness.

Because the market model is most commonly used in research to generate expected returns and no better alternative has yet been found, despite the weak relationship between beta and actual returns (Armitage, 1995), the market model was used in the analysis to predict “normal” return. To test for robustness, a three-factor model can also be employed.

The t-test is of interest to examine whether there are differences of the abnormal returns over time and especially across types of markets. The event-study approach shows the explicit impact of SWF actions because the methodology is based on individual purchases and sales of publicly available equities.

This result is in line with the positive market responses to the investments in the entire sample—SWFs with low-level corporate governance accounted for the majority of the sample of SWF investments.

    Other Resources Citing This Publication