It is not too hard to figure out what Japan is doing with its reserves.
From the beginning of 2003 to September 2004, Japan’s official reserves increased by $360 billion. Over that time frame, Japan’s holdings of treasuries increased by $342 billion. That number also includes private Japanese purchases of treasuries, but it still tracks the reserve data pretty closely. There is a bit of a lag between the purchase of dollar reserves and the purchase of treasury bills with those dollars, but the two go up together. There is a reason why treasury dealers say Japanese intervention leads to higher demand for short-term treasury bills.
China poses a bit more of a mystery. From beginning of 2003 til September 2004, the People’s Bank of China’s (PBoC) reserves increased by about $268 billion (including the roughly $45 billion used to recapitalize state banks). China’s recorded holdings of treasuries increased by $56 billion over this time, so only 20% of so of China’s increased reserves are showing up in higher recorded Chinese holdings of treasuries. This year reserves are up $106 billion while holdings of treasuries are up only $16.5 billion, so the fraction of Chinese reserves going into Treasuries seems to be declining over time.
So what is China doing with the remaining $210 billion of reserves assets that it has added over the past two years? Some of it may be going into "disguised" buying of Treasuries - through various asset management firms or foreign broker dealers. Holdings of treasuries by Caribbean banking centers are up are up $50 billion this year; UK holdings are up by a comparable amount. Some of the demand from the Caribbean represents demand from US hedge funds based there. But some may be coming from central banks working through foreign broker dealers or even investing in hedge funds (remember, the Bank of Italy invested in LTCM).
Some of China’s growing reserves are no doubt going into other US assets - agencies, asset backed securities, even corporate bonds -- to try to get higher yield. Reported official buying is of these assets is small, but central bank may be buying through intermediaries.
Is much going into Euro denominated assets? To date, probably not - people in the market would know if the PBoC was moving into Euro. Moreover, China is large enough that its numbers impact on the global statistics: in 2003, the world added roughly $60 billion in euro and yen reserves, and $440 billion in dollar reserves. Since China accounted for about a quarter of the overall increase in the world’s reserve, it seems likely that its reserve acquisition did not deviate too much from the global breakdown of roughly 90% dollars/ 10% other currencies.
Let’s look at the data in slightly different way, by looking at China’s total stock of reserves rather than what China is doing with the reserves it has bought over the past two years. The Economist this week estimated that 80% of China’s $515 billion in reserves are in dollars, which works out to $412 billion. China’s recorded holdings of Treasuries are around $175 billion. That leaves around $240 billion in other dollar assets - not chump change.
China’s reserve management is fundamentally constrained by its peg to the dollar. Suppose China started moving its reserves from dollar debt to euro debt - that would have the same impact as if China took some of the dollars now coming into the PBOC (after all, reserves going up by over $10 billion a month) and used them to buy euros. It would put pressure on the euro/dollar. If China kept the yuan pegged to the dollar, then yuan would depreciate along with the dollar and the PBOC would need to start to buy more dollars to support the peg ... China would end up needing to continue to buy large amounts of dollars to support the peg.
All this is fancy way of saying that if China were to start pegging to the euro, it could buy euro assets to keep the yuan/ euro rate fixed. But so long as it is pegged to the dollar and there is pressure to appreciate v. the dollar, supporting the peg requires China to buy lots of dollar assets.
There is a scenario where China might be able to hold fewer dollar assets and still keep the peg to the dollar... namely, if the European Central Bank (ECB) bought China’s dollar reserves. Suppose the PBoC starting selling dollars/ buying euros, but the ECB then stepped in, buying dollars/selling euros to avoid euro appreciation. This would let China buy dollars to support its peg and then sell its dollars to the ECB for euros -- avoiding any pressure on the euro/dollar. But it is not likely that the ECB is any more interested in the PBoC in buying $150 billion plus of dollar debt a year.
China indeed faces a real dilemma. Not only does it have a large stock of dollar reserves, but that stock is growing. Yet the value of that stock is also likely to fall in the future. China’s over $500 billion in reserves are currently equal to about 1/3 of China’s GDP, so a 33% real appreciation of the yuan would generate capital losses equal to 10% of China’s GDP. That is a big loss, by any measure. Defending the peg right now requires reserve accumulation of $150 billion a year (in the third quarter, China’s reserves increased by almost $45 billion). If that continues, in four years, China’s reserves would easily exceed $1.1 trillion. China’s GDP is rising too - if the exchange rate stays constant and the IMF’s growth path is right, GDP will be @ $2.35 trillion in 2008. Reserves would then be equal to about 47% of China’s GDP.
A 33% real appreciation and reserves to GDP of 45% produces a capital loss of 15% of GDP. But that probably underestimates China’s future losses, since over time, the scale of the real appreciation China needs also is rising. China’s economy is becoming more productive very quickly -- and remember that Japan’s economic miracle in the 50s, 60s and 70s was accompanied by substantial real appreciation in yen. A 50% real appreciation produces a capital loss of above 20% of GDP. That is real money, even for fast growing China.
This is why what Larry Summers has called the balance of financial terror is fundamentally unstable. The "terror" aspect comes from the fact that is China were to start to sell (or just stop buying), the value of its existing holdings would fall rapidly and US interest rates would go up sharply. The balance refers to the fact that to date, China has opted to hold its Treasuries and other dollar assets, not to sell. But for the balance to stay stable, China literally needs to double its bet - and double the central bank’s expected loss -- over the next four years.