from Follow the Money

Just how much money do sovereign wealth funds have?

April 6, 2008

Blog Post
Blog posts represent the views of CFR fellows and staff and not those of CFR, which takes no institutional positions.

More on:

Financial Markets

Capital Flows

A ton, according to Morgan Stanley’s Stephen Jen and a host of Wall Street investment banks

Not all that much, according to Milken Institute’s Christopher Balding (as reported by Bob Davis in the Wall Street Journal’s economics blog). The Gulf – which is home to the big sovereign funds has only about $300b in reported US assets. Its reported reserves aren’t that big. And its supposedly big sovereign funds haven’t stepped up to inject more capital into the US financial system since January.

The truth, I suspect, lies in the middle. Some Wall Street numbers sometimes are a bit too high for my taste. But Felix is also right: Christopher Balding’s estimate seems significantly too low.

The most likely reason why sovereign funds haven’t made more investments in US financial firms is that they lost a ton of money on their investments, not a lack of cash. Their losses (unrealized) haven’t escaped notice in their home countries. They may feel like they were, to paraphrase Landon Thomas’ reporting in the New York Times “made fools of.” It probably didn’t escape sovereign funds notice that a lot of banks paid out record bonuses right after securing big commitments from sovereign funds – and that some of the same firms looking for capital last fall are still short of capital now ….

Thomas also notes that the Gulf investors aren’t happy with the criticism that their investment has received in the US. They believe the US government should be thanking them, not asking them to more transparent. Thomas’ reporting appeared in the Deal Journal "Leveraged Planet" special from last week -- it is well worth reading.

Why am I confident that Balding’s estimate is too low? Simple: the US data understates the financial holdings of the Gulf. There is a relatively well-known reason for this too – as Rachel Ziemba explains – the big Gulf funds and the Saudi Monetary Agency have effectively outsourced the management of much of their wealth. They put their money in the hands of a Swiss bank or a London asset manager. And the Swiss bank or London asset manager invests on their behalf. The US data only sees the investment from Switzerland or London, not the source of the money. And frankly the UK facilitates all of this by refusing to publish any detailed data on capital flows in and out of the UK. I titled one of my early papers on petrodollars “disappearing into London”

There is another reason why I am confident that the Gulf has large assets that don’t show up in the US data: the IMF’s balance of payments data.

If you compare the increase in Gulf holdings of US assets over the past five years ($193 billion, once $32b in estimated valuation gains from rising equity markets are stripped out, per the US Treasury survey) to the IMF’s data showing “official” outflows from the Gulf over the last five years ($731b; the WEO statistical appendix table A13) the flows to the US are way too low to be credible.* Only 27% of the Gulf’s official outflows show up in the US data -- a very low total for a region whose currency is linked to the dollar. Put a bit differently, if the US data is right there is no risk that any actions by the US will drive Gulf money out of the US – as the US is already massively under-represented in the Gulf’s portfolio.

* I compared the increase in the US assets of the "Asian" oil exporters from June 2002 to June 2007 to cumulative official outflows -- including reserves - in the IMF balance of payments data for the Middle East from 2003 through 2007. I could have adjusted the US data to turn the June survey data into a December estimate based on the monthly flows, but the reward to effort ratio was too low. I also assumed -- probably incorrectly -- that all of the increase in the Gulf’s holdings came from official investors.

Moreover, it is reasonable -- based on the balance of payments data -- to think that the Gulf now needs about $50 a barrel oil to pay for its imports. if oil stays around $100 and oil production in the GCC around 15 mbd, simple math suggests that the GCC counties will have roughly $275b to invest abroad in 2008. A similar calculation suggest that the Gulf had around $70b to invest in q1. Once the IMF releases a bit more 2007 data next week I’ll be able to make a more precise estimate.

On the back of these numbers, it seems pretty likely the US data significantly understates the Gulf’s holdings of US assets -and that the US data correspondingly provides a poor quite to the size of many Gulf funds. On the other hand, many of the estimates for the current size of sovereign wealth funds are in my judgment a bit too high.

Setting questions about the dividing line between a central bank and a sovereign fund aside (Should Russia’s stabilization fund, the investment portfolio of Hong Kong’s monetary authority and the non-reserve assets of the Saudi Monetary Authority be considered sovereign funds?), the debate over the size of sovereign funds largely boils down to a debate over the size of the Gulf’s funds. Singapore’s GIC is the only other real source of uncertainty.

And even in the Gulf, some things are pretty well known.

We roughly know how much Kuwait has. It reports its data annually – and we should get a new data point for the end of q1 2008 fairly soon. It has $215b in its sovereign fund in q1 2007. That is higher now – probably more like $275-300b. It also has $20b in reserves. We also roughly know how much the Saudi Arabian Monetary Agency manages. SAMA’s monthly report shows that it has about $300b in non-reserve foreign assets plus $60b from Saudi pension funds. The data is reported in Saudi riyal, but the conversion isn’t hard.

There are some thorny questions raised by the “private” assets of various wealthy Saudis, including some wealthy Saudis who might be considered sovereign. These assets are likely to be substantial, though there is next to no data that allows a reasonable estimate of their size.

But the biggest source of uncertainty stems though different estimates of the size of the Abu Dhabi investment authority (ADIA). It is wildly considered to be the world’s largest largest sovereign fund. It is also among the most secretitve. So far ADIA has refused to disclose its size.

Indeed, as the IMF notes in its paper on sovereign wealth funds (see paragraph 32, p 17), a number of countries fail to report even basic balance of payments data in order to disguise the size of their sovereign assets.

The Gulf’s refusal to meet basic international norms for balance of payments data disclosure is one major reason why I am not terribly sympathetic to the Gulf’s sense that it isn’t as welcome as it should be as an investor.

The Gulf funds have insisted on a level of secrecy that has precluded building the kind of public record that would prove to outside observers that they invest apolitically. Given strong suspicions that the Gulf’s currency policy is driven as much by politics as economics, such not totally unfounded. Investing in the US when it doesn’t make economic sense could be considered a politically driven investment decision.

The Gulf’s desire for secrecy also has made it hard to narrow down ADIA’s potential size. I of course don’t know how much ADIA has. But I had to bet, I would bet it has somewhere between $600b and $700b. That is a huge sum – but it is still well below the $900 to $1000b estimates that often appear in the press. It also a lot bigger than the $250b number sometimes cited.

There is a fair amount of informal guidance that suggests some of the high end estimates are too high. Landon Thomas – who seems to have developed a good set of sources on these questions – has reported that some high end estimates are a bit high ("bankers, former employees and analysts familiar with the fund peg it at $650 billion to $700 billion — an amount that is still over 15 times the size of the Fidelity Magellan Fund."). The IMF’s Mohsin Khan has also indicated that ADIA doesn’t have a trillion dollars, or even $900b.

Rachel Ziemba and I have been trying to build a model for some of the Gulf funds. Our (still preliminary) work is based on an estimate of Abu Dhabi’s oil revenues, an estimate of the broad contours of its portfolio and publicly available data on how various indexes that correspond with its portfolio have done.

Our methodology creates a range of sources of error. We could have underestimated the amount of money that ADIA is getting out of Abu Dhabi’s oil. ADIA obviously could have outperformed many indexes. We haven’t, for example, included a separate estimate of the performance of ADIA’s PE investments -- and those investments likely outperformed a simple index from 2003 to 2007. Leverage works in a rising market. But the biggest source of error is that we don’t really know how much money ADIA started the decade with.

The Wall Street Journal reported ADIA had around $150b in 2000. If that is right, it is hard to see how ADIA now has anything close to $700b (absent truly superb investment performance) right now. The numbers work better if ADIA had more like $250b in late 2000.

Our model has a lot of limitations, but it has one virtue: it lets us make an informed guess about how ADIA has evolved over time. Call it marking ADIA to a (bad) model.


The rapid growth of ADIA’s portfolio from 2003 on reflects both the recovery of global equity markets and a surge in oil revenues. We estimate that ADIA has received between $30 and $40b in new oil money to invest from 2005 on, and in 2005 and 2006 ADIA’s assets increased by around $100b a year once its capital gains are added in. They probably increased by more -- as we haven’t allowed for some potential sources of excess returns. That is why I personally would bet ADIA has a bit more than the chart suggests.

Bottom line: The Gulf’s oil exports are real. The external surpluses that have resulted from the recent surge in oil prices are also real. The foreign assets that have been bought by that surplus are real. The Gulf countries penchant for secrecy and aversion to transparency is also real. That makes determining the precise size of the Gulf’s assets hard – but there is pretty good evidence that the big Gulf governments combined had well over a trillion in external assets at the end of 2007 – and probably more like $1.5 trillion.

As importantly, that sum should rise by something like $300b in 2008. The Gulf is now attracting substantial capital inflows in addition to running a large external surplus.

More on:

Financial Markets

Capital Flows