from Follow the Money

Trillionaires R Us

October 26, 2006

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

More on:

Monetary Policy


The rewards for guestimating the currency composition of China’s reserves pale relative to the rewards for lending your credibility to China’s effort to clean up and recapitalize its banking system.    

But my estimates of the currency composition of China’s reserves have been widely cited.  So I feel rather exposed to the risk that they may prove to be way off.  Kind of like Goldman should ICBC go belly up when China’s current boom ends …

So here is the due diligence behind my estimate.

It is primarily based on data that is now a bit dated – the Treasury’s survey of portfolio investment in the US at the end of June 2005.   That survey revealed far higher Chinese holdings of US debt than could be inferred from the monthly flow data.  Total Chinese holdings of US debt were $524b at the end of June 2005, and China added $184b to its portfolio between the end of June 2004 and the end of June 2005. (See table 5)     

The $184b increase made sense to me, given China’s reserve growth– Chinese inflows in the monthly data had long seemed too small relative to China’s reserve growth.   The flow data is now picking up a higher fraction of Chinese reserve growth, for reasons that I don’t fully understand.   But until the next survey comes out, we won’t know quite how well the TIC data from July 2005 on has tracked Chinese purchases.    My guess is that the monthly flow data still somewhat understates total Chinese purchases. 

Indeed, there is no particularly strong reason why one would expect the TIC flow data to track Chinese purchases perfectly.  If China buys a Treasury bond from a European pension fund, that transaction would never register in the TIC data.   And if a custodian bought on behalf of SAFE, it wouldn’t show up as a Chinese purchase either. 

China also has dollar deposits in the international banking system – though as the BIS notes, SAFE isn’t considered a central bank so SAFE’s deposits are not considered official – as in central bank – deposits.    

Based on those two data sources, I would be surprised if substantially less than 70% of China’s total dollar holdings were held in dollars.   I would not be surprised if China’s dollar holdings were a bit higher than 70% of its portfolio --  though it is quite possible that some of China’s dollar reserves are invested in dollar-denominated debt from emerging economies.    

Conversely, I think the most likely explanation for China’s enormous $100b reserve increase in q4 of 2004 is that China does hold some of its reserve portfolio in euros and pounds.  Sure, hot money inflows also probably picked up in q4 2004, but at least some of the gain presumably came from currency moves.  So I would be surprised if China holds as few euros and pounds as the Japanese.

Prasad and Wei demonstrated, in my view, that capital gains (and losses) from currency moves explain at least some of the errors and omissions in China’s balance of payments data for 2002, 2003 and 2004 (See Table 9 of their paper).   They haven’t extended their analysis through 2005, but their argument still holds.   The dollar’s rally lowered China’s reported reserve growth – and reduced the error term.

Two additional points: 

First, the Economist notes that most of China’s reserves are in Treasuries.   That is true.  And it also note, correctly, that China seems to be shifting into higher-yielding dollar debt.  China’s portfolio of Agencies and mortgage backed securities has been growing faster than its Treasury portfolio.   Given the size of China’s overall reserve portfolio, $339 in Treasuries is if anything a little on the small side.   If my calculations are right, China’s total holdings of dollar are close to $700b.  And its holdings of agencies and mortgage-backed securities – when all the data from the next survey comes in and the totals are summed up – may match its Treasury holdings.   They weren’t that far behind back in mid-2005.  See Table 4 of the Survey.

Second, the Economist notes that the real issue isn’t what to do with China’s existing trillion.  It is how to avoid accumulating the next trillion.

The only real solution to the poor return on China's reserves is to stop accumulating them. That requires policy reforms to reduce China's massive saving, which drives its current-account surplus, and a more flexible exchange-rate system. But before that happens, China's reserves could well hit $2 trillion.

Surprise, surprise: We at RGE agree!

More on:

Monetary Policy