What is a sovereign wealth fund?

A sovereign wealth fund (SWF) is an organisation that manages financial and other assets on behalf of a government or nation. But what about the holdings of foreign exchange reserves by central banks? These assets, often much larger in size, are not regarded as SWFs. So what’s the difference? This question arose for me last week from meetings in Thailand and China, each of which has a central bank but only China has a SWF – in fact it has three.

The difference is in what the assets are held for. A SWF typically has a long term investment horizon, possibly generations in length, and aims to maximise some risk-adjusted return, accepting some risk and illiquidity along the way. A central bank that manages foreign exchange reserves usually puts the emphasis on safety and liquidity, since these reserves are held against the risk of a sudden loss of foreign investor confidence of an abrupt shortage of foreign exchange liquidity. Returns are secondary or irrelevant.

Foreign exchange reserves

Most central banks hold some assets in the form of foreign reserves, meaning financial assets denominated in foreign currency, plus gold. These assets are held as precautionary savings, to manage the risk of a shock or disruption to the foreign exchange market. If for some reason a country suffers a sudden shortage of foreign currency, it can use its reserves to carry on buying essential imports and to stabilise the exchange rate, while buying time to deal with the underlying cause of the disruption (such as a loss of foreign confidence arising from an overheating economy).

Countries’ demand for foreign reserves dramatically increased after the Asian financial crisis of 1997. Several countries, starting with Thailand, suffered an abrupt loss of confidence by foreign lenders and investments, who withdrew funding. Many Asian fast growing economies depended on foreign funding for their banks and had borrowed heavily in foreign currency, so the region saw a disastrous combination of bankrupt banks and a huge increase in the cost of servicing the debt as the domestic currency plummeted. Indonesia saw a fall in GDP of about 20% on top of a huge increase in the cost of foreign debt.

Determined not to let this happen again, emerging economies around the world started building foreign exchange reserves over the next decade. The only way to do this is to run a persistent balance of payments surplus, because this means more foreign exchange flowing in than flowing out, with the surplus going into reserves. And since a surplus must imply somebody else is running a deficit, this increase in reserves had as its counterpart the large balance of payments deficits run by the US in the decade leading up to financial crisis of 2007-09.

Since these reserves need to be available at short notice to manage any disruption in the foreign exchange market, whether caused domestically or by global crises, they are held in safe, liquid assets such as the bonds of the rich economies, chiefly the US, some Eurozone countries, the UK and Switzerland.

Those countries like the US and UK which can borrow in their own currency need not hold foreign reserves and typically don’t intervene in the exchange markets any more, though Japan has been known to do so. The US holds very little in foreign exchange reserves but it does hold a lot of gold, which could be converted into foreign exchange though it is clearly not as liquid as government bonds. (Whether gold is as safe as bonds is a controversial issue.)

The largest foreign exchange reserves in the world are those of China, currently in the region of $3.3 trillion. China was only slighly affected by the Asian financial crisis because it controlled foreign financial flows (and still controls does even now), though it did suffer the bankruptcy of the provincial government-owned Guangdong International Trade and Investment Corporation (GITIC) in 1999. These reserves represent the accumulation of years of balance of payments surpluses, and the deliberate holding down of the RMB currency value until recently, to keep China’s exports competitive. The reserves are held by China’s central bank, the People’s Bank of China, and managed by its subsidiary the State Administration of Foreign Exchange, nicely abbreviated to SAFE.


SWFs are mostly set up by countries with natural resources. The aim is to save some of the income from the taxes levied on the oil, copper, iron ore or whatever to shift consumption into the future. This is done for efficiency and equity reasons. The efficiency reason is that the economy may not be able to safely handle all of the income at present and there is a risk of a “Dutch disease” crowding out of other parts of the economy. The equity argument is that it is not fair for the current generation to enjoy all the benefits of the natural resources. By investing some of the proceeds for the future, later generations may also benefit.

A few countries without natural resources also have SWFs, chiefly Singapore and China. A persistent balance of payments surplus means these countries have large amounts of foreign exchange which could either be used to buy more imports now (a consumption and investment boom) or invested for the future. Arguably in both cases the surpluses reflect an underlying undervalued exchange rate. The rise in the RMB and move to a more neutral balance of payments means that China’s reserves are no longer rising and in the longer term may fall if the government allows more offshore investment by Chinese residents (which would also threaten the property market, but that’s a different story).

Accurate data on SWFs is not always available. The consensus is that the largest fund is probably the Government Pension Fund Global of Norway, at about $700bn, though Abu Dhabi Investment Authority (ADIA) is of broadly similar size but doesn’t disclose its asset total. Norway is uniquely transparent in the way its SWF operates and provides long term projections of the fund, based on investment performance and on the amount of new oil and gas revenue that will go into it. On its high scenario, the fund will have $3.3 trillion in 2030, about $650,000 per person. Even adjusting for inflation, this would provide a reasonable modest income for the entire population without anyone having to do any actual work. So my pension advice to anyone is marry a Norwegian. There is an interesting analysis of the Norway model, contrasted with the Yale investment model, by my colleagues David Chambers and Elroy Dimson here. (Some key differences are that Norway avoids all “alternative” investments, minimises transactions costs and sticks to rigid rebalancing of assets. It seems to work.)

China has three SWFs. The “official” one is China Investment Corporation (CIC) which has around $500bn invested. The National Council for Social Security fund has around $150bn, which is to provide future pension returns. But most of this money is invested domestically. Lastly there is the SWF part of the State Administration of Foreign Exchange (see above). This is where the definitions get blurred. SAFE manages in total about $3.3 trillion. A part of this, which is estimated by various people to be in the range of $500bn, is separately invested in a SWF.

But even the “reserves” part of SAFE is now apparently looking at assets other than the traditional safe government bonds that are normally held in foreign exchange reserves. The government is shifting the priority from safety and liquidity to seeking a higher return on its assets. This implies a shift from government bonds to equities, corporate bonds, lower quality government bonds and real estate. The actual asset allocation is not disclosed but there are apparently well sourced press articles that document SAFE investments in UK real estate.

If the Chinese government had decided that only a part of its $3.3 trillion need be managed as traditional foreign reserves then SAFE could mutate into by far the world’s largest SWF. China’s foreign reserve management needs are much less than $3.3 trillion. China’s imports are about $150bn a month (recent months have been distorted by the lunar new year). So even if China stopped exporting and had no net foreign investment, it could keep paying for imports for nearly two years from its reserves. Just six months of imports would be a very safe level of reserves, so three quarters of the total reserves could be invested in SWF assets without any risk to China’s foreign exchange management needs. If that were to happen, then SAFE would become a SWF with assets in the region of $2,500bn, or more than three times the size of the Norwegian SWF. If you add China’s other SWF assets it’s clear that China in total has by far the largest amount of assets managed by or on behalf of a state. (And this excludes the large and growing funds being lent by other state owned and directed financial institutions such as China Development Bank, subject of a forthcoming blog).

The significance of this shift would be that even if the foreign exchange mix of assets didn’t change (i.e. the proportion held in dollars was maintained) the asset mix would. Instead of investing around $1,000 billion in US government bonds, China might invest a similar amount in US equities, corporate bonds and real estate. That shift would potentially be good for the US, though it might lead to some interesting price movements.

Thailand doesn’t currently have a SWF but apparently the government thinks it would be a good idea, though perhaps for the wrong reasons. The central bank, the Bank of Thailand, is resisting the idea of diverting some of its $178bn into a separate SWF. It may have reason to be concerned. In China, where the central bank is admittedly far less independent of the government than it is in Thailand, CIC was created in 2007 by diverting a portion of PBOC’s reserves. CIC, though part of the government sector, reports into the Ministry of Finance, not to PBOC. So, while there are perfectly good reasons for arguing about the correct asset allocation for a government, there are also institutional rivalries around the formation of new centres of financial power.

7 Responses

  1. Samir Prakash

    Simon – thank you for another excellent post. Your pieces are little gems of clear, concise writing.

    A question, well, two: You note that Norway’s SWF is very transparent. Roughly, what is the break-up of its investments, by asset class and by continent/country? Second: does being more open about its investments expose it to any new risks, or increase the magnitude of any risks that would exist in any case?

    • Simon Taylor

      Thanks for your nice comments. All the information you could want is included on the Norwegian SWF’s website. As most of their investments are in public equities and bonds there is no reason why the transparency should harm them.

  2. Ilyas

    Simon, a very topical article. I was reading something about the decision making that might go into a Russian bail out of Cyprus (unlikely) and the way in which strategic objectives can be pursued by govts through their SWF arm.

    However my question is not on Russia. Do you think that the Middle Eastern SWF structures are really discreet from FOREX and Central Bank management ? Qatar, Saudi Arabia and the UAE have simply vast investment arms that manage and invest money on behalf of entities that are wholly govt owned.

    • Simon Taylor

      If you take the UAE, Abu Dhabi Investment Authority states that it is quite independent of the government of Abu Dhabi and even more so from the Central Bank of the UAE. ADIA doesn’t disclose its assets under management in its annual review but various sources place the amount in the region of $650bn. The Central Bank’s foreign assets at the end of 2011 (the latest data I could find) were about $45bn. So the SWF for Abu Dhabi alone manages far more money than the central bank. But the central bank needs its reserves to manage the exchange rate of the Arab Emirates Dirham (AED) which is pegged against the dollar. So the two institutions have distinct roles and funds, at least in theory.

      Saudi Arabia is different. The Saudi Arabian Monetary Authority combines both functions. As of 2012 Q1 it had total net foreign assets of about $570 billion, divided between gold, foreign deposits and foreign securities. The CIA Factbook estimates these rose to $623bn by the end of 2012. Strictly speaking SAMA is not a SWF but it appears to invest in equities and real estate, along with the more liquid assets usually held as reserves. As Saudi foreign assets amount to about four years worth of imports, a very large share of the assets could safely be invested in less liquid assets without prejudice to the exchange rate management and reserve role. Treating all of the foreign assets as SWF assets, as some rankings of SWFs do, seems to me to overstate somewhat the scale of Saudi SWF assets, but the line is even more blurred than in China. SAMA is not, of course, independent of the government.

  3. Wing Hung

    I read somewhere however that the Thai govt doesn’t think SWF is a good idea.

    Need to ask some of our Thai classmates.

  4. Richard Carney

    Simon, Great piece — clear and concise. Thank you! I’m quite late in picking up on this thread, but I hope you would be open to responding to another question: do you think sovereign wealth fund equity investments would be managed differently when most of their assets come from excess foreign exchange accumulated from national exports compared to oil-export reserves? For example, would you expect the former to engage in more passive investments (smaller more diversified equity stakes) due to their need for greater liquidity in case of the need for foreign exchange?

    • Simon Taylor

      The goal of a SWF is independent of the source of funds. More important than whether the funds come from a general trade surplus (China) or from natural resource export earnings is how the SWF is set up and governed. In some countries, a dip in foreign earnings has led to the SWF being used as a short term source of funds, just like a reserve fund. If that’s likely to happen then it makes sense to hold liquid assets, and in effect the SWF is partly a form of reserve fund. In some countries (Norway is the best example) this sort of treating the SWF as an emergency piggy bank is ruled out by clear rules and effective institutions, allowing the SWF to invest in long term assets (which may include less liquid ones that would not be suitable as reserves).

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.