- Blog Post
- Blog posts represent the views of CFR fellows and staff and not those of CFR, which takes no institutional positions.
Premise. China moved. It has no impact on the bond market a week later (Ten year yield = 4.2 one week after the move v. 4.15% before). Therefore China has no impact on the bond market. Note - yields are a bit higher today, ISM and all.
Reality. China did not really move. 2.1% is peanuts. A small move - this is one point where I agree with Professor McKinnon - only whets speculators' appetite. China's reserve accumulation, and its direct impact on the bond market, will if anything increase.
Survived SARs correctly highlights all the difficulties managing an exchange rate regime that sure looks like the result of a messy internal political compromise inside China.
China is still spending a ton of money to keep its currency from appreciating, adding to its reserves and therefore still has a ton of money to invest in various markets around the world. Remember, China runs a bigger current account surplus in the second half of the year than in the first, for seasonal reasons. China's 2005 current account surplus -- close to $150 b according to Goldman -- may be close to its 2004 reserve accumulation. Without a fall off in hot money flows, China's reserve accumulation will pick up in the second half of the year. Adjusted for valuation, China's 2005 reserve accumulation is likely to be around $295 billion v. $195 b or so in 2004. The "raw" increse in China's reserves is likely to be at least $275 billion, even with valuation losses on China's euro and yen reserves.
Whatever impact China is having on the bond market, it is still having it. To quote John Berry:
If the dollar cost of Chinese exports doesn't go up, then there is no immediate impact on trade flows or the amount of dollars being accumulated by the Chinese government. And that suggests that another reason for the excitement that followed the revaluation announcement -- the possible impact on the purchase of U.S. Treasury securities and therefore on U.S. interest rates -- also was misguided.
No doubt Asian central banks are not the only factor influencing Treasury rates. They are one of several factors. I have an immense amount of respect for the work the Goldman team has done on global adjustment, but I am not convinced by Bill Dudley's argument (in the WSJ article) that central banks have not been all that important an influence on the US market.
Though to be clear, central banks are less important this year than in 2003 and 2004. US interest rates are now higher than interest rates on the bonds of most other potential reserve currencies. Less important though is different from not important.
And just to job everyone's memory, remember that China's average annual reserve increase between 98 and 2000 was under $10 billion a year. $200 b plus increases in Chinese reserves are something that is kind of new.
Or look at it in another way: in 2004, the global increase in reserves - almost all reserves to be invested in the US and Europe - was about $625 billion, adjusted for valuation effects. The global current account surplus - if you net Spain's deficit against Germany's surplus and look at the eurozone as a whole - was about $800 billion. In other words, most of the world's "surplus" savings (savings that does not finance domestic investment) shows up in the coffers of the world's central banks. Yes, that means netting out offsetting private flows, but I think that kind of netting makes sense ...
The global reserve increase this year will be at least $500 b - mostly in China and the world's oil exporters. The global current account surplus will be $1000b or so. So reserves do matter less. But remember, much of the "surplus" that is not showing up in reserves is in Japan and in places like Saudi Arabia. Japan's past intervention has shaped the market. And the line between the ruling family and the state is rather blurry in Saudi Arabia. I am not sure I would consider increases in the kingdom's private assets totally "private" flows.
Let's review the bidding - and look at various ways central banks can influence yields in the US treasury market without directly buying US assets.
1) By spending 15% of its GDP to keep its exchange rate from appreciating, China is holding down the global price of consumer goods (and soon more than consumer goods - think MRIs; check this Inc. story via the Big Picture) and keeping inflation low. That helps keep bond yields low.
2) Competition from China - in a world where the central bank is spending 15% of its GDP to keep Chinese wages, in dollar and euro terms, below what they otherwise would be -- certainly contributes (one can debate by how much) to slow real wage growth in the manufacturing sectors in the US and Europe. Richard Freeman argues that the "doubling' or more of the global labor supply from the integration of China, India and the former Soviet Union is one reason why corporate profits (and right now corporate savings) are currently quite high as a share of national income.
China's impact on the world economy can best be understood as what economists call a "positive supply-side shock". ... with twice as many workers and little change in the size of the global capital stock, the ratio of global capital to labour has fallen by almost half in a matter of years: probably the biggest such shift in history. And, since this ratio determines the relative returns to labour and capital, it goes a long way to explain recent trends in wages and profits.
If the corporate savings glut is keeping bond yields down, as JP Morgan argues, that too has something to do with China.
3) China's demand for euro reserve assets keeps euro yields down, and thus drives up the differential between US and European rates - making US assets more attractive.
Let's consider the last channel a bit more.
(warning -- lenthy data analysis follows)
China - counting the reserves transferred to the state banks -- had $448 billion in reserves at the end of 2003. It now has around $771 billion in reserves, and I suspect it will have close to $930 billion in reserves at the end of the year. That is an increase of a bit less than $500 b over two years.
Suppose China started with a 65/30/5 dollar/euro/yen portfolio at the end of 2003 (given its accumulated stock of dollars, it would be hard for China to have had fewer than 65% dollars at the end of 2003. And further suppose it will end 2005 with a 60/35/5 dollar/euro/yen portfolio - i.e. a tiny bit of diversification.
China would have added, by my calculations, China would add $265-270 billion to its dollar reserves (so far, only $109 billion has showed up in the US data ... ) and $190 billion (roughly 160 billion euro) to its euro reserves over two years.
Euro is shorthand for a broader set of European currencies. But 160 billion euro - 80 billion euro a year -- would be over 1/3 of total foreign demand for euro debt over this period. In 2004, inflows into long-term euro denominated debt were around 210 billion, and inflows into long and short-term debt combined were about euro 230 billion.
China reserve accumulation - as the Economist argues - is big enough to shape a range of markets.
I intend to put my various Chinese reserve accumulation calculations up this week, with nice charts and graphs. But in the interim, a few more points on my "where are China's reserves" obsession:
Bank of America (in the Wall Street Journal) estimates that China holds about $452 billion of US securities - or about 60% of its total reserves. That is consistent with the numbers that emerge if you sum up Chinese purchases over the past few years. Look at the numbers in Table 10 of Wei and Prasad. Summing their numbers over the past 14 years along with $40 b in recorded flows this year (through May) produces a total of around $380 billion - or about 50% of China's reserves. To get $450, Bank of America just uses the reported Chinese holdings of Treasuries, which are about $60 b higher than the sum of the flows. This type of analysis provides the basis for the argument that China already has diversified its reserves.
I continue to wonder if this methodology works - namely because the small flows into the US reported in the US data, if accurate, imply very large flows into other markets, and I don't see that in the data.
Two concrete examples.
The fraction of China reserves that showed up as Chinese purchases of US assets - recorded purchases - fell way off in 2003. Reserves went up by $162 billion (including the dollars transferred to two state banks). Recorded flows from China - private and official flows combined -- into the US were $68-69 billion. Wei and Prasad estimate valuation gains added $20 billion or so to China's reserves in 2003, based on a 70/20/10 dollar/ euro/ yen reserve split. So China's had about $140 billion in actual new capacity to buy reserve assets in 2003. If China only bought $68-69 billion of US assets, it necessarily bought $70 billion of non-dollar reserve assets.
That though is inconsistent with the BIS data for 2003. The BIS calculated that all the central banks of the entire world only bought $61.5 billion of non-dollar reserve assets in 2003. One of the two numbers HAS to be off.
Or consider q4 of 2004, when China added $91 billion to its reserves. Say $11 billion of that came from valuation gains. Recorded inflows into the US were around $24 billion ($23.8 b by my math). That leaves $56 billion to invest in other markets, or to build up China's various bank accounts.
In q4 total foreign (all foreign) purchases of euro debt - short and long-term - totaled euro 26.3 billion, or around $34 billion (at an average exchange at the time of purchase of 1.3 billion) - far short of what would be implied to absorb all of China's "non-dollar" reserve increase. That total includes foreign sales of short-term debt, but the flows into long-term debt also fell well short of what would be needed to absorb China's reserves.
Of course, I am ignoring the possibility that China built up its euro and other bank accounts, but my point is clear. If China's reserves are not going into dollars, they have to be going somewhere.
I certainly don't know the composition of China's reserves.
But until we can do a better job accounting for where China is putting its money, if not in dollars, I would be cautious. The Bank of America number is a minimum. Lots of analysis is working off the assumption that almost all of China's reserve increase that does not show up in the US data is being invested in other currencies rather than invested in dollar assets in ways that don't show up in the US data. That is an assumption. Right now, it is one that I at least find hard to square with other known data points.
But if true, the implied non-dollar flows are enormous, particularly in relation to overall capital inflows to other markets. Big enough to drive down yields in other markets to the point where dollar yields become more attractive to private investors.