from Follow the Money

Managing too much: China’s reserves were up $29b in November, $28b in December, $78b in q4 and $247b for the year

January 15, 2007

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

More on:


China’s reserves reached $1,038.7 billion ($1.0387 trillion) at the end of November – an increase of $29 billion over October.     China then added another $27.6b in December, bringing its reserves up to $1,066.3 billion ($1.0663 trillion) at the end of the year.

That is a “headline” increase of $247.3b for the year.   Adjust that total for the rising dollar value of China’s estimated holdings of euros and pounds, and the valuation adjusted increase is about $222b.   That would be a bit under China’s estimated $240b 2006 current account surplus. 

In the fourth quarter, China’s reserves increased by about $78b.  That sounds impressive.  It certainly seems to have generate some concerns inside China.  Jiao Jinpu of the PBoC warned that:

The problem of excess liquidity in the financial system cannot be solved in the short-term and will continue this year as the country's trade surplus keeps expanding and US dollar depreciation causes more capital inflows into China. He noted that the central bank will face increasing difficulty in setting monetary policy, as a narrowing China-US interest rate gap limits room for the PBoC to adjust interest rates.

But to my mind, the PBoC actually got off easy in q4.  A $78b increase in  Chinese reserves in the fourth quarter is actually on the wimpy side.    A 20% euro/5% pound/ 5% yen portfolio would have increased in value by about $9b – so the valuation adjusted increase is around $69b.

China’s trade surplus was $68b.   The current account surplus tends to be about $15b larger (on a quarterly basis) than the (custom’s) trade surplus.    China tends to attract about $5b in (net) FDI inflows a month, or $15b a quarter.    Sum that up, and China’s foreign assets should have increased about $100b in the fourth quarter. 

Roughly $70b of that appeared on the PBoC’s balance sheet.  But $30b didn’t.

The distribution of the increase in China’s reserves between November and December also surprised me a bit.  Japan – which has, I would guess, a lower fraction of its reserves in euros, pounds, and the like than China – saw its reserves go up by $15b in November solely on the back of valuation gains and interest payments.   Its reserves then fell by $1b in December, both because the dollar rose a bit against the euro and because US rates rose, reducing the market value of Japan's long-term bonds. 

I consequently expected a much bigger increase in China’s reserves in November, and a smaller increase in December.

China’s November trade surplus was $22.9b (call it $23b) .    Throw in another $5b a month to get the current account surplus – and another $5b a month in (net) FDI inflows.  That suggests – absent any hot money flows, any acquisition of foreign asset by state companies and state banks or any valuation gains – a monthly reserve increase of at least $30b. 

And we know China had some valuation gains in November.  The euro appreciated by around 3.8% against the dollar in November, rising from 1.2773 to 1.3261.    If one assumes (to make life simple) that reserve currencies other than the yen moved in tandem with the euro and rounds the euro/dollar move to 4%, valuation changes should have pushed Chinese reserves up by $16b if China had 40% of its reserves in euros and similar currencies currencies, $12b if China had 30% of its reserves in euros and similar currencies and $8b if China had 20% of its reserves in euros and similar currencies …

It has gotten a bit harder to infer portfolio composition from (estimated) valuation gains recently, since Chinese reserve growth no longer is predictably a bit above the trade balance.     Recently reserve growth has roughly tracked the trade balance – something that implies ongoing “private” capital outflows (or the acquisition of foreign assets by state companies).    Fluctuations in those flows can generate unpredictable monthly variation in reserve growth.  

But the absence of a bigger headline increase in November still calls into question some estimates of China’s reserve portfolio.  Both Deutsche bank and Morgan Stanley have estimated that China may only hold 60% of its reserves in dollars.   That seems unlikely to me – if China really did have than many euros (or things that moved like euros) in its portfolio, its reserves should have increased by more in November.   $16 (valuation) +$23b (the trade surplus) = $39b.  That is a bit more than $29b.

For that matter, my estimate that 30% of China’s reserves are in currencies other than dollars (and 25% are in euros and currencies that move with the euro) is a bit hard to square with the absence of a bigger increase in China’s reserves in November.   I could hide behind the “unpredictable monthly variation in reserve growth” argument, but to be honest, right now I would probably say that 70% is on the lower bound of China’s likely dollar share.      

Conversely, the $28.6b December increase is only a bit less than one would expect given China’s $21b December trade surplus, its estimated $26b December current account surplus and an estimated $34b overall increase in foreign asset accumulation given strong FDI inflows (FDI inflows seem to have topped $8b in December).   Valuation wasn’t a big factor in December.  

And since China’s reserve growth in q3 and q4 has been running at about the pace of its trade surplus, the December increase seems a bit on the high side, relative to recent trends.   Like I said, fluctuations in month to month reserve growth now make it a bit harder than it used to be to try to infer portfolio composition from implied valuation gains. 

Here is some rough math for the year.

China’s reserves increased by $247.3b this year.   That is a bigger headline increase than in 2005 – but in 2005 valuation changes held Chinese reserve growth down while in 2006 valuation changes pushed China’s reserve growth up.    A 20% euro/ 5% pound/ 5% yen portfolio would have increased in value by $25b over the course of the year.   That implies valuation-adjusted reserve growth of $222b.    

Keeping a constant 70% dollar portfolio share implies dollar reserve growth of around $173b.    Keeping a constant 20% euro/ 5% pound/ 5% yen portfolio would imply euro purchases of $30b, pound purchases of $6-7b, and yen purchases of $13b.   The yen purchases are larger – relatively speaking – because the value of Japan’s existing yen didn’t rise over the course of the year.   

In 2005, China’s valuation adjusted reserve growth was about $238b (my estimate) – and China also (per the IMF) shifted about $29b to the banks: $15b for the recapitalization of ICBC and $14b in various swaps.   That works out to total foreign asset accumulation of $267b.

China’s 2005 current account surplus was $60b more than its (customs) trade surplus.   A $178b (customs) trade surplus in 2006 consequently translates into a $240b current account surplus – well above the 2005 current account surplus ($161b). 

Throw in $70b in FDI inflows to China in 2006.   One would expect China to have added about $310b to its total foreign assets.  If Chinese FDI abroad reached $20b, that would imply a $290b increase in China's holding of foreign debt.

$290b is higher than $222b.  There is a gap of $70b.   And I am being conservative in calculating the $290b expected increase in Chinese foreign assets – I am, for example, leaving out the portfolio equity inflows associated with the IPOs of ICBC and the Bank of China, inflows that we know have been held offshore. 

So what happened?   Who  -- other than the PBoC – has been building up its holdings of dollar and euros?

That is the question.  There was a $45b increase in “private” holdings of foreign debt in the balance of payments data for the first half of 2006.  A similar increase in the second half of the year would make the overall balance of payments add up.    

It is possible this reflects a surge in truly private Chinese demand for foreign assets as capital controls have been lifted.  

But I sort of doubt it.  M2 growth has been fairly strong – suggesting that lots of Chinese money is still flowing into RMB bank accounts.   RMB deposit growth is also faster than overall deposit growth, suggesting that domestic fx deposits are continuing to fall as a share of total deposits.  That hardly suggests huge demand for dollars.  Chinese money that didn’t go into the banking system looks to have gone into the stock market – not into foreign markets.    The acceleration of the pace of RMB appreciation in the second half of 2006 reduced the appeal of foreign bank accounts.   There aren’t anecdotes of huge sums of Chinese money moving offshore.   

And we are conservatively looking about $70b – a big sum.

My guess is that state banks, state insurance companies, state pension funds and state companies account for most of the increase …. But that is a guess. 

Still, the fact that Chinese reserve growth didn’t increase along with China’s current account surplus certainly made the PBoC’s job easier.  Look at the following chart – which projects the $45b increase “private” holdings of foreign debt continued in the second half, generating a $90b outflow for the year (the light blue bar).  I also included FDI outflows (the tan bar) in the chart.


Let’s see if “private” (probably state) institutions continue to add to their dollar balances in 2007 even as a faster pace of RMB appreciation generates bigger valuation losses on those dollars.   And let’s see if Chinese firms can step up their investment abroad.

I am sure the PBoC is watching carefully.  They are looking at the same data that I am.  They know that current rates of export and import growth imply a $300b, if not bigger, Chinese current account surplus in 2007.    And they know net FDI inflows are likely to continue.   And they consequently know that absent dollar purchases by state banks, state pension funds and state firms, the increase in the PBoC’s reserves needed to sustain China’s current exchange rate regime could approach $400b.

(p.s. I should also recommend Stephen Green's article looking at the PBoC's profits from its reserves in 2006, he calculates that the PBoC got far more interest on its dollars and euros than it paid out in RMB, and that the valuation gains on the PBoC's euros and pounds offset the valuation losses on its dollars, so the PBoC made money in 2006.   That's true.   But, I would note, over time, the RMB needs to appreciate against both the euro and dollar -- the PBoC shouldn't bank of valuation gains on its euros consistently offsetting the losses on its dollars.  At some point, it will take losses on both)


More on: