- Blog Post
- Blog posts represent the views of CFR fellows and staff and not those of CFR, which takes no institutional positions.
Central banks are not selling dollars. Some analysts tend to put a bit too much emphasis on changes in the dollar’s share of global reserves – and a bit too little emphasis on the increase in total dollar holdings.
Parmy Olson of Forbes, for example, focuses on the small fall the dollar’s share of global reserves in q4 of 2006 rather than the change in the stock, arguing that a change in the dollar's share of global reserves is evidence that central banks are selling dollars.
“The International Monetary Fund recently showed that the world's currency reserve holders have indeed been selling dollar assets. In the fourth quarter of last year, the dollar's share of the $3.5 trillion in global reserves fell to 64.2%, from 64.6%, as central banks opted for more euros and British pounds.”
Alas, the IMF’s data actually shows the opposite: central banks added to their dollar assets in q4. The total stock of reserves among those countries that report detailed data to the IMF increased by around $180b, to $3307b. The dollar's share slid a bit, but total dollar holdings among those central banks that report data to the IMF increased by $85b in q4. The IMF's data shows holdings of $2076b in q3 and around $2162.5 in q4.
And that total doesn’t include any increase in China's dollar holdings. China is among the countries that do not report data to the IMF (see the charts I prepared for the Peterson Institute’s China conference). If the likely increase in its dollar holdings is added in, central banks added far more than $85b to their dollar holdings in q4.
I am nit-picking because this, in my view, is an important issue.
The big story over the past year in my view has been the rise in the pace of reserve accumulation and the increase in the share of the US deficits financed by central banks, not small changes in the dollar's share of total reserves.
Net central bank purchases of dollars almost certainly reached a record high The IMF’s data is pretty clear on this point.
Between q2 2006 and q2 2007, those central banks that report the currency composition of their reserves to the IMF added $367.6b to their dollar reserves. Countries that do not report detailed data to the IMF added another $496b to their reserves, and it is a safe bet that a lot of that increase was used to buy dollar-denominated assets. That $496b increase in reserves is also a bit of understatement as well – as it doesn’t include a roughly $50b increase in the non-reserve foreign assets of the Saudi Monetary Agency.
Suffice to say that central banks have been big net buyers of dollars.
Analysis of China’s portfolio sometimes suffers from a similar problem.
Too much attention is paid to the data the US Treasury reports on China’s holdings of Treasuries, even though we know that China has shifted its purchases toward Agencies – and have good reason to think that some of the Treasuries that the US data suggests have been sold to buyers in the UK actually have been sold to the PboC and other central banks.
But I won’t wallow around in the data mud for any longer.
Setting the point about central bank “dollar sales” at a time of record "dollar purchases" aside, the Olson article – which draws heavily on Stephen Jen’s recent work on Sovereign Wealth Funds – is quite interesting.
Jen believes that the shift toward Sovereign Wealth Funds implies a shift in official flows toward Asia and emerging economies – and a shift away from both the US and Europe. I suspect Jen is right. Sovereign wealth funds aren’t restricted to classic reserve assets and Asian equities are a lot more attractive than Asian bonds. Most sovereign wealth funds have signaled that they are looking to increase their exposure to emerging Asia.
That though begs another question. While sovereign wealth funds want to invest in Asia and emerging markets (and particularly in Asian emerging markets), both Asia and the world’s emerging markets run current account surpluses. They don’t need the money. The US by contrast does. If the shift from reserves to sovereign wealth funds means a shift from financing the US to financing emerging Asia, both the US and a host of Asian economies need to adjust. Otherwise, one country’s sovereign wealth fund will just end up adding to another country’s reserves.
And – as Andrew Osborn and Joanna Slater of the Wall Street Journal highlighted in an excellent frontpage article on Wednesday -- most emerging economies already are attracting more funds than they feel like they can manage. Neither India or Thailand exactly wants more inflows -- or even more reserves.
Not now. Not when they already have more than enough trouble managing all the funds coming into their economies. One solution would be more currency appreciation. But Osborn and Slater note that countries are afraid to appreciate faster than China.
[according to Raghuram Rajan, a finance professor at the University of Chicago and a former IMF official.] "In general people are worried about losing their competitive edge...especially if demand from the U.S. slows down and you're fighting one another for the crumbs”
The situation is complicated by China keeping a tight rein on its yuan, through restrictions on capital flows and massive government buying of dollars. Countries that compete with China for exports are reluctant to let their currencies strengthen rapidly.
As a result, a host of countries are looking for ways to reduce the pace of their reserves growth – whether by liberalizing controls on outflows, by tightening controls on inflows, or by setting up sovereign wealth funds to take some assets off the central banks books.
All this helps, but it doesn’t really solve the core issue. So long as the emerging world resists the kind of currency appreciation that would eliminate its current account surplus, someone in the emerging world has to finance the US and Europe.
Not every emerging economy can set up a sovereign wealth fund that just invests in other emerging economies. Not if Bretton Woods 2 is going to last. More on this later.