Economics of Sovereign Wealth Funds

Chapter 10 Sovereign Wealth Fund Investment Strategies: Complementing Central Bank Investment Strategies

Udaibir Das, Adnan Mazarei, and Han Hoorn
Published Date:
December 2010
  • ShareShare
Show Summary Details

The global financial crisis has two relatively obvious implications for sovereign wealth funds (SWFs). Not only will the compression in global imbalances depress the future growth trajectories of assets under management by SWFs, the sharp correction in global equities in 2008 may have undermined the appetite and political tolerance for risk-taking by some SWFs, at least in the short run. However, notwithstanding these shocks in the short run, SWFs’ general proclivity and capacity for risk-taking—in the long run—should not be significantly altered because the motivations for generating investment returns on sovereign wealth should still prove to be compelling, especially after the sharp recovery in equity prices in 2009 and in early 2010. Also, official reserves have continued to build in the world, and petrodollars have burgeoned again, providing a powerful boost to SWFs’ asset bases. This chapter highlights some of the traits of SWFs’ long-term investment strategies that should remain relatively unaltered.

The financial crisis has also reminded central banks of the importance of having adequate foreign reserves in liquid form, held in the intervention currency. Given that the U.S. dollar is the dominant intervention currency in the world, the decade-long trend of reserve diversification away from U.S. dollars by central banks may not necessarily continue. Indeed, the diversification out of U.S. dollars seems not to have taken place as dramatically as some may think. The lower share of U.S. dollars is partly a consequence of the dollar’s depreciation for much of the period since 2002, rather than of active dollar selling. At the same time, while the global financial crisis has few direct implications for the desirable currency composition of SWFs’ portfolios, indirectly, to the extent that central banks stop diversifying away from the intervention currencies, SWFs may need to compensate by diversifying away from the U.S. dollar and the euro more rapidly than otherwise. Thus, governance of some SWFs and their investment strategies may not be as independent as they once were, and greater coordination with corresponding central banks may be the trend going forward. In other words, greater specialization by central banks and SWFs is likely, with less competition between them as they pursue distinct investment strategies, with central banks focusing on safety and liquidity while SWFs focus on investment returns.

This chapter has six sections. First, an updated forecast of the likely growth path of the collective assets under management (AUM) by SWFs is presented. Second, the case is made that SWFs will likely continue to pursue higher risk–higher return strategies, despite the global financial crisis. Third, a distinction is made between SWFs’ different risk profiles based on the sources of their funds. Fourth, a collective model portfolio is suggested for SWFs. Fifth, the “sibling rivalry” between some central banks and SWFs that had been brewing before the global financial crisis is discussed. It is argued that the rivalry will likely abate as central banks and SWFs, coerced by the crisis, become more complementary as they specialize in different tasks. Sixth, the chapter draws attention to a closely related category of funds—sovereign pension funds (SPFs)—which in the aggregate have even larger AUM than do SWFs. SPFs should increasingly exhibit the same traits highlighted for SWFs.


The volatility in the financial markets in 2008 and 2009 led to meaningful swings in the AUM of SWFs. Because of marked difference in the ways in which equities, bonds, and alternative investments have performed since 2008, different investment portfolios should have led to big differences in performance between SWFs. This author has long argued and assumed that, in the aggregate, SWFs are likely to have a portfolio with a 45:25:30 split between equities, bonds, and alternative investments (Jen, 2007). For the SWF community as a whole, therefore, the guesstimate is that the total loss since end-2007 may have been about 11.1 percent. While there should have been a wide range of performances, depending on actual portfolio weightings, most SWFs are likely to have suffered modest net investment losses during this period. SWFs with more aggressive profiles may have suffered double-digit investment losses (15 percent or more), while those with conservative profiles may have fared much better (5–6 percent losses or less).

How will the total AUM of SWFs evolve over time? This exercise first computes the likely trajectories of the current account balances of the world’s largest holders of foreign reserves.1 Assuming that the net financial flows are nil, it then projects the likely trajectories of the official reserves and the SWFs’ share of those reserves. A further assumption is made that central banks with large foreign reserves will increasingly allocate excess reserves to their respective SWFs or invest a bigger portion of their reserves in higher risk—higher return markets. The forecasts are in Figure 10.1. By 2015, SWFs are likely to have about US$11 trillion in AUM, and the “crossover” point where the AUM of SWFs exceeds that of central banks’ reserves occurs in 2016.2

Figure 10.1AUM of the World’s SWFs Could Reach US$22 Trillion by 2020

Source: Bluegold Capital LLP Research.


Though the growth profile of the AUM of SWFs may be modestly tempered, SWFs will most likely remain an increasingly important group of investors—a major source of support for risky assets—over the coming years. Despite the recent balance of payments shock, many emerging-market economies still have sizeable excess reserves positions. The three-month import-coverage rule and the Greenspan-Guidotti Rule, combined, suggest emerging markets may have roughly US$2.0 trillion in excess reserves (about US$1.5 trillion in China).

SWFs’ preferences for equities over bonds reflects their essential makeup, seeing that they were created to move out on the risk-return curve. This part of the thesis will not change. Furthermore, SWFs’ greater capacity to express long-term views—resulting from their superior liquidity and lack of leverage—should continue to permit them to capitalize on opportunities that private institutional funds may find too illiquid. Moreover, the structural conversion of wealth by oil-based SWFs from underground real wealth (oil) to above-ground financial wealth will certainly continue.

Figure 10.2 shows the cumulative nominal investment returns on various assets. As can be seen, the performances of these asset classes had been extraordinarily diverse until the onset of the global financial crisis. While past performance does not guarantee future performance, this is still a useful figure. For most of the period since 1987, energy and emerging-market equities have been outperformers, while government bonds have been laggards, though over the entire period, the passive return on bonds has matched that on global equities. Even with the violent sell-off in global equities and real estate in recent quarters, the returns in these asset classes are only modestly lower than the return on global bonds, which have enjoyed an arguably unsustainable safe-haven bid. To be precise, US$100 invested in 1987 would have grown to, in March 2010, US$914 in emerging-market equities, US$725 in energy, US$495 in real estate, US$465 in bonds, and US$476 in global equities. Furthermore, with rising concerns about market risks of sovereign bonds, investing in equities and commodities remains a compelling proposition for SWFs.

Figure 10.2Relative Buy-and-Hold Returns of Various Asset Classes

Source: Bluegold Capital LLP Research.

That the performances of various asset classes can be so diverse should not be news. Investors are concerned not just with investment return, but also with the associated risk of various investments. Figure 10.3 shows a scatter chart of the risk-return profiles of various asset classes (covering the period 1987 to 2009). These points indicate the risk-return position if 100 percent of the portfolio was invested in one asset. In addition, the figure displays the efficient frontier of a long-only portfolio, given that most SWFs are constrained to long-only investments. Emerging-market equities and oil mark one end of the efficient frontier, that is, high return and high volatility, while sovereign bonds mark the other end. While the global financial crisis has forced a comprehensive re-think of how historical returns and volatility should be interpreted, and has taught us that a better risk concept is needed, the point of Figure 10.3 is how fundamentally different government bonds are when compared with other assets.

Figure 10.3Risk-Return Profiles of Different Asset Classes

Source: Bluegold Capital LLP Research.

Note: Data compiled from December 1987 to December 2009. GSCI is a benchmark for investment in commodity markets.

Thus, notwithstanding the global financial crisis, the motivations for diversifying national wealth from foreign sovereign bonds still appear compelling.


Whether SWFs raise the world’s risk-return tolerance depends on the sources of the SWF’s assets. Oil-based SWFs should have the biggest impact on the world’s risky assets, followed by SWFs funded from surpluses derived from trade in goods. SWFs based on capital inflows should have minimal—and possibly even negative—effects on the world’s aggregate risk-taking preference. The net overall effect on the world’s risk-return profile depends on the relative size of these three types of SWFs.

Based on the sources of the accumulation of their assets, SWFs could be divided into those funded by oil and other commodities (Type 1), goods-trade surpluses (Type 2), and net capital inflows (Type 3). A similar classification can be made for countries with large reserve accumulations that do not (yet) have an SWF. As measured by AUM, some 46 percent of the largest reserve holders and SWFs (with total assets of US$6.0 trillion) are Type 1, 36 percent are Type 2, and 18 percent are Type 3. Type 1 SWFs should have the highest risk-return profile, because they essentially have no distinct liabilities. Type 2 SWFs should have a marginally lower appetite for risk than Type 1 SWFs because of their domestic bond liabilities. But the underlying savings-investment surpluses should still imply future consumption and the entitlement of the surplus countries to accumulate claims on foreign assets, including equities. Being net creditors, the SWF and reserve managers should not have tight liquidity constraints. This idea—that reserve holdings should be invested in foreign (higher-yielding) equities—applies particularly well to a country such as Japan.3

Type 3 SWFs should not be aggressive risk-takers, given the relatively fickle nature of capital inflows. Because Type 1 and Type 2 have much larger AUM than Type 3 SWFs, in the aggregate, SWFs should in theory raise the world’s risk-return profile.4 India is a good example of a Type 3 country. Its official reserves reached US$276 billion at end-June 2010, having risen sharply from US$137 billion at end-2005. However, more than all of the reserve increases have come from capital flows, because India has run persistent trade deficits. Brazil is another example of a Type 3 country. Thus, if India and Brazil establish their own SWFs, it is likely they will not have particularly high risk tolerances relative to the SWFs from the Middle East, for example. Of course, the overall risk tolerance of an SWF is not solely determined by the source of the assets; other considerations are also important for the portfolio construction, such as the desire to diversify risk (e.g., an oil-based SWF investing in risk assets with negative correlations to oil).

Figure 10.4 shows the efficient frontier and the single-asset returns over the period 1987 to 2007. Type 1 SWFs, as shown in this figure, should move from the 100 percent–oil point to somewhere along the market line (the line tangent to the efficient frontier and intersecting with the risk-free interest rate). Type 2 and Type 3 SWFs, however, should move to a point on the market line, but away from the 100 percent–government bond point. The three types of SWFs will end up at different points along the market line. Specifically, they should be in the order indicated in Figure 10.4.

Figure 10.4How SWFs Affect Financial Risk-Return Balance

Source: Bluegold Capital LLP Research.

Note: Data compiled from December 1987 to December 2009. GSCI = a benchmark for investment in commodity markets. G10 cash rate represented in this figure by US$3 million interbank rate.

Oil-based SWFs, therefore, should invest at a higher risk-return balance than the rest of the world, because, based on Figure 10.2, crude oil is a low-return, high-volatility investment. It makes sense, in theory, for oil exporters to convert their wealth from underground real assets (oil) to above-ground financial assets. The recent decline in oil prices may constrain the ability of the Gulf Cooperation Council SWFs to invest as aggressively as they could before the global financial crisis, but the motivation for them to take risk should be unaltered, as argued above. Whether Type 2 SWFs will take more or less risk as a result of the crisis is difficult to tell at this point. Type 3 SWFs, however, are likely to further reduce their appetites for risk, because the sudden halt of private capital inflows during this crisis has underscored the fickle nature of these flows.


What could the model portfolio look like for SWFs in the aggregate? Although different SWFs have different preferences and mandates, it is useful to consider the portfolios of established pension fund SWFs to help envisage the likely evolution of SWFs’ portfolios in the coming years. Given current allocations and taking into account that SWFs are usually from countries that are already rich in reserves and have large exposures to sovereign bonds, it seems plausible that, on average, large SWFs may eventually have portfolios with about 25 percent of total assets in bonds, 45 percent in equities, and 30 percent in alternative investments.

The investment portfolios of SWFs evolve over time. Norway’s Government Pension Fund–Global (GPF) is a good example. As it has matured, it has raised its exposure to equities and alternative investments. The growing size and the long-term nature of the GPF’s investments permitted Norges Bank Investment Management to try to capitalize on the liquidity premium by entering into less liquid but higher-return investments. Other SWFs will follow the same path, in the view of this chapter’s author.

But established large pension funds are further along in this maturation process. Established large pension funds such as Caisse de dépôt et placement du Québec (CDQ) of Canada (with US$191 billion in AUM) and ABP of the Netherlands (with US$325 billion in AUM) already have portfolio structures with significant and balanced exposures to bonds, equities, and alternative investments—which include not only real estate, infrastructure, and private equity, but also commodities and hedge funds. The portfolio structure of SWFs in general will probably evolve toward that of large pension funds and beyond, in terms of risk-taking. Many of these large pension funds have roughly 35 percent in bonds and 35 percent in equities. Because countries that have SWFs are usually rich in official reserves, almost all of which are held in sovereign bonds, it is likely that SWFs have lower preferred exposure to bonds, compared with central bank reserves. Indeed, the Government of Singapore Investment Corporation announced on September 23, 2008, that its portfolio allocation consisted of 26 percent bonds, 44 percent equities, and 25 percent alternative investments—not too different from the 25:45:30 expectation.


Signs of more competition between reserve managers at central banks and SWFs within the same countries became increasingly apparent during much of 2007; the two types of entities had increasingly similar investment strategies and may have been competing over the management of foreign assets. Specifically, more and more central banks may have begun to invest in foreign equities and corporate bonds to enhance their own investment returns, to prevent capital from being transferred to the SWFs.5 With central banks not wanting to part with their reserve holdings, large reserve holders in Asia were on their way to adopting portfolios similar to that of the Swiss National Bank, with exposures to equities and corporate bonds.6 As young SWFs try to establish themselves, there is a period during which central banks are in a more advantageous position (more skilled personnel and longer experience). Indeed, some central banks tried to exploit their advantages and invested more aggressively in risk assets to raise the hurdle that SWFs needed to overcome to justify further fund transfers from the central banks. Up to the onset of the financial crisis, Japan, China, and the Republic of Korea might have already been managing their reserves more aggressively than they had in the past.

The crisis, however, may fundamentally alter the investment strategy of central banks and SWFs. Instead of competing against one another—that is, having increasingly similar investment strategies—central banks and SWFs could have more complementary investment strategies going forward, with each specializing in different tasks: central banks concentrating more on preserving liquidity and security of their assets, and SWFs focusing more on enhancing investment returns. Four fundamental arguments support this notion.

First, when some emerging-market economies came under balance of payments pressures (from a sudden stop in private capital inflows, a sharp decline in exports, and a fall in investment earnings from overseas assets), large-scale interventions were necessary. (The IMF has had to deploy unprecedented financial resources to stabilize parts of emerging-market economies.) Reserve hemorrhaging, in practice, occurs primarily through a particular intervention currency, which in most cases is the U.S. dollar. The crisis highlighted the importance of central banks having adequate foreign reserves, not just in the intervention currency but also in the most liquid form of asset in that currency. U.S. agency bonds and mortgage-backed securities, for example, were temporarily illiquid, and may have been a complication for some central banks that needed U.S. dollar cash. Another reserves-management problem exposed by the financial crisis is that many emerging-market central banks do not mark their assets to market. This created the situation whereby central banks avoided selling assets that had suffered market losses, thus avoiding realizing the losses that were never reported. The U.S. agency bonds and mortgage-backed securities were, again, a good example—central banks short of U.S. dollar cash resisted selling their non-U.S. treasury asset holdings, not because they were not liquid, but because of the valuation problem. In short, the crisis raises a serious question about the relative weights central banks should place on liquidity, security, and return in relationship to their objectives in reserves management.

Second, SWFs, in contrast to central banks, should have very distinct mandates and not be constrained by the need to have adequate resources in the intervention currency. SWFs’ portfolio management decisions should attach different relative weights to liquidity, security, and returns, compared with those of central banks. To the extent that central banks need to stop diversifying out of U.S. dollars, the corresponding SWFs could diversify more aggressively out of dollar assets to achieve the desired overall currency exposure from the nation’s public-sector perspective. This observation indicates that central banks and SWFs should not have a competitive relationship. Figure 10.5 suggests that, although reserve managers may be encouraged to retreat into primarily sovereign bonds as a result of the crisis, SWFs should continue to diversify their portfolios. To the extent that central banks halt, if not reverse, the process of U.S. dollar diversification, SWFs may follow Path B in Figure 10.5 more than Path A in their diversification processes.

Figure 10.5Central Banks to Invest in Equities and Corporate Bonds

Source: Bluegold Capital LLP Research.

Note: G5-G10 = Australia, Canada, Switzerland, New Zealand, Norway, Sweden.

Third, while central banks should have independence in devising monetary policy, the case for their having both independence and total freedom in managing foreign reserves is weak. The years ahead will likely witness a clearer delineation of responsibilities and objectives between central banks and SWFs, and rivalries between these two types of institutions, and thus overlapping investments, will likely abate.

Fourth, the analysis for this chapter shows that, as SWFs gradually rebalance toward the model portfolio suggested above, there will likely be large sales of both U.S. dollar and euro assets, and purchases of yen and emerging-market assets. Based on the market capitalizations of various asset classes in different currencies, a benchmark currency basket was calculated that is consistent with the 25:45:30 model portfolio. This benchmark’s currency composition is 43 percent in U.S. dollars, 18 percent in euro, 13 percent in yen, 9 percent in pound sterling, and 17 percent in other currencies. For comparison, the currency exposures of developing countries’ official allocated reserves is 58 percent in U.S. dollars, 31 percent in euro, 2 percent in yen, 6 percent in pound sterling, and 3 percent in other currencies.7


SWFs are important, but SPFs are at least as relevant. The AUM of SPFs is approaching US$5.3 trillion—about a third larger than that of SWFs. Also, starting from a position of a relatively low-risk portfolio and a high “home bias,” though there will be exceptions, SPFs as a group are likely going to raise their risk profiles (i.e., hold more equities and fewer bonds) and increase their foreign content (i.e., hold more foreign assets). The same demographic trends (declining fertility and mortality rates) that have fueled SWFs will also lead to an evolution in SPFs. These prospective trends will likely have meaningful implications for financial markets. The Republic of Korea’s National Pension Service has just acquired a share in London’s Gatwick Airport and real estate in Canary Wharf. China’s National Social Security Fund is expected to grow to US$300 billion in five years’ time. Australia’s Future Fund is doing well and will also grow as a result of the expected fiscal surpluses Australia will run in the coming years. Japan’s US$1.35 trillion Government Pension Investment Fund is dormant now, but could be revived if politics allow. Table 10.1 shows the top 16 SPFs in the world.

TABLE 10.1Sovereign Pension Funds
Domestic-foreign allocationAsset allocation
CountryFund nameAssets (US$ million)Domestic





United StatesSocial Security Trust Funds2,419,000100010000
JapanPension Bureau (GPIF)1,345,687802078212
Korea, Rep. ofNational Pension Service240,000901077203
AustraliaAge Pension237,0001684n.a.n.a.n.a.
ChinaHK Monetary Authority139,700n.a.n.a.n.a.n.a.n.a.
CanadaCanadian Pension Plan127,60026374314326
ChinaNational Social Security Fund114,000937n.a.n.a.n.a.
SpainFondo de Reserva de la Seg. Soc.84,00077239640
AustraliaFuture Fund (2006)49,3007228254332
FranceFonds de reserve pour les retraites43,0006733504010
CanadaRegie des Rentes du Quebec32,000821882180
IrelandNational Pension Reserve Fund30,6000100226315
SwedenPPM (Premium Pensions System)25,700396238566
New ZealandNew Zealand Superannuation15,600485248520
Average329, 012604058358
Sources: Various national pension Web sites; International Financial Services, London.Note: n.a. = not available.

Equities include private equity in some countries, when transparency of portfolio constituents are high.

Others include real estate, hedge funds, cash, inflation-linked assets, and alternative assets.

Canadian Pension Plan (CPP) has 26% Canadian equities and the rest foreign (Financial Highlights on CPP Web site). We assumed the same weights for fixed income.

Sources: Various national pension Web sites; International Financial Services, London.Note: n.a. = not available.

Equities include private equity in some countries, when transparency of portfolio constituents are high.

Others include real estate, hedge funds, cash, inflation-linked assets, and alternative assets.

Canadian Pension Plan (CPP) has 26% Canadian equities and the rest foreign (Financial Highlights on CPP Web site). We assumed the same weights for fixed income.


Although the global financial crisis may have somewhat moderated the pace of SWFs’ asset accumulation, total AUM of SWFs are still likely to approach US$11 trillion by 2015. Total AUM of SWFs is currently a little more than US$3.6 trillion; emerging economies have, in addition, US$2.0 trillion or so of excess reserves, part of which could potentially be transferred to SWFs. In the long run, SWFs’ investment strategies should not be significantly altered by the crisis—the arguments for seeking higher returns by accepting a bit more risk are still compelling. Although the source of capital to a large extent dictates SWFs’ risk tolerance, in the aggregate, SWFs could move toward a model portfolio with 25 percent bonds, 45 percent equities, and 30 percent alternative investments. Furthermore, the crisis exposed shortcomings in the currency-diversification strategies of some central banks. A halt in U.S. dollar diversification by central banks is now a distinct possibility, with SWFs, in contrast, accelerating their U.S. dollar diversification. In general, central banks and SWFs will likely adopt more complementary investment strategies in the years ahead. Finally, a closely related category of funds—SPFs—is worth monitoring. Not only are their AUM larger than those of SWFs, they are likely to exhibit similar investment styles as SWFs.


The IMF’s World Economic Outlook forecast series for the period 2010–14 is extrapolated beyond 2015.

Total official reserves are about US$9.3 trillion, and will likely exceed the US$10 trillion mark by the end of 2010. This analysis assumes that the total returns on official reserves will be about 3 percent a year, and those for SWFs will be 7 percent a year.

Japan’s investment earnings on foreign asset holdings are now larger than its trade surplus: Japan’s official reserves easily generate US$30 billion to US$50 billion a year in investment returns. This feature of Japan’s external account makes it as much a Type 1 economy (with fiscal reserves from investment earnings) as it is a Type 2 economy (one that will see its trade surplus recover with the global economy).

To properly answer the question of how global financial assets may be affected by SWFs, one needs to ask the counterfactual question of not only where the ultimate portfolios will end up, but also how the assets in question would have been deployed had they not flowed into the SWFs.

Although this might have seemed unusual (most central banks only invest in sovereign bonds), it is not unorthodox. The IMF’s Balance of Payments Manual explicitly acknowledges foreign equities as legitimate reserve assets. Furthermore, the Swiss National Bank has been investing in equities and corporate bonds since 2004.

At end-2008, the Swiss National Bank had around 146 billion Swiss francs (Sw F) worth of foreign reserves, Sw F 47 billion of which was held in foreign exchange reserves, Sw F 31 billion in gold, and the rest in claims from Sw F transactions and other assets. While the bulk (64 percent) of the bank’s foreign exchange reserves were still held in government bonds at end-2008, its corporate bond and equity holdings were substantial—22 percent and 11 percent of the total, respectively.

COFER database, composition as of end-2009.

    Other Resources Citing This Publication