- Blog Post
- Blog posts represent the views of CFR fellows and staff and not those of CFR, which takes no institutional positions.
Please allow me a bit of latitude. China won’t add $500b to its reserves in 2007 for the simple reason that the central bank will sell some of its reserves to the state investment company. Moreover, China might opt to ship some of its spare dollars over to the state banks. But rather than spell those qualifications out, I am going to talk of “reserves.”
There is just no way to get around the fact that China bought a ton of foreign exchange in the first quarter. Its reserves went up $135b in the quarter – topping its previous record quarter ($85 or so in q4 04). Only about $5b of the increase came from valuation gains v. $20b or so in late 2004, so China actually bought a lot more fx in the market now than then -- $130b v $75b (by my estimates). $130b in a quarter works out to $520b for a year.
To paraphrase Macro Man – who is almost as reserve-obsessed as I am – China accrues foreign exchange reserve faster than Britney Spears accrues tabloid headlines.
Jon Anderson of UBS is more measured, but he still wrote:
.. "We're shocked. From trend reserve growth of around $20b a month in 2006, January, February and March rolled in at nearly US$50b a month, or $140b for the first quarter as a whole. This is more than just a jump. The magnitude and abruptness of the acceleration are simply stunning."
Roughly $100b of China’s $130b in reserve growth probably flowed into dollar assets. That roughly what it would take to keep the dollar share of China’s reserves constant. China might have even bought a few more dollars. Macro Man reports that he didn’t see 20b euros worth of dollar sales from the PboC in first quarter (actually he did the calculation for the PboC and a couple of its partners in crime … but the basic point holds).
$100b in a quarter is $400b annualized. To state the obvious, that is a lot. Especially for a single institution in a single country. Talk about a concentrated position. Talk about power and interdependence.
$400b is enough to have almost fully financed the United States 2000 current account deficit. That was back during the .com boom when private investors were dumping money at the US.
It is enough to finance roughly ½ the United States much bigger 2007 current account deficit.
It is two times as much financing as the Japanese provided the US in 2003 or 2004 – back when the Japanese were intervening on a scale that was considered unprecedented.
It is roughly two times the IMF’s effective lending capacity.
It works out to $30-35b a month. The IMF’s biggest bailout loan – to Brazil – was roughly comparable in size. And that was disbursed over two years. So China’s central bank provides as much to the US in a typical month as the IMF provided to one of the world’s biggest emerging economies over two years.
I could go on. The big surge in reserves in the first quarter either reflects a pick up in “hot money” flows – perhaps more accurately described as non-FDI capital flows -- or an unwinding to past swap contracts. If the surge comes from a pick up in non-FDI inflows, it requires a very large swing indeed: there was an outflow of $20b in q4 (a rather suspicious outflow if you ask me) and an inflow of around $50b in q1 (by my calculations -- I use an estimate of the current account rather than the goods trade surplus in my calculations, which produces a lower number than is often thrown about). $50b is about a ¼ of what the US needs to cover its current account deficit – that is real money on a global scale.
The other possible explanation -- as both Richard McGregor of the Financial Time and Andrew Batson of the Wall Street Journal report -- is that the banks unwound a lot of their existing swaps with the PBoC. Jon Anderson of UBS notes that The Bank of China reported $40b in outstanding swaps in its end-2006 report and China Construction $30b in the middle of 2006. The unwinding of some of these swap lines may help to explain the surge in reserve growth in q1. Basically, the theory is that the PBoC kept its reported reserves down in 2006 by shifting some of its reserves over to the banks -- and then unwound than shift in q1, leading reported reserve growth to soar.
A swap can only be unwound once, so that source of reserve growth shouldn't be projected forward.
But you don’t actually need to expect that q1 pace of non-FDI capital inflows will continue for the entire year to think that China may add close to $500b to its reserves this year. Chinese export growth picked up at the end of last year and the first part of this year. If Chinese export growth doesn’t slow over the course of the year and Chinese import growth doesn't pick up, I am looking at a $320b trade surplus/ a $400b current account surplus. China tends to have a larger current account surplus in the second half of the year for seasonal reasons, and the surplus is also trending up over time. Throw in $50b in net FDI inflows and the q1 inflows -- whether hot money flows/ or the unwinding of past swap contracts -- and you are at $500b.
Some – including some senior US officials – like to argue that China cannot afford to sell its dollar reserves, since it can only sell at a loss. But I am not sure China can afford to continue to add to its reserves at this pace.
China almost certainly has a bit more than $1.2 trillion in reserves and reserve-like assets. $60b – if not more – is over at Central Huijin to recapitalize three state banks. Stephen Jen has estimated that Huijin may hold as much as $100b. I had thought that an additional $50b likely has been shifted to the balance sheets of the state banks through swaps, but some of that may have been unwound. So let's say China now holds $1.3 trillion in reserves and reserve-like foreign assets.
By the end of the year, on current trends, the government’s holdings of foreign assets likely will reach $1.7 trillion. Unless something changes, China is set to reach $2 trillion in the middle of 2008.
To top it off, China isn’t the only country adding to its reserves. Far from it. Christian Menegatti and I believe the non-Chinese reserves grew about as fast as Chinese reserve in the first quarter. Think $250-60b in global reserve growth/ $200b in dollar reserve growth.
Annualized, that works out to a $1 trillion increase in global reserves and a $800b increase in dollar reserves. Throw in $50b in dollar flows from the world’s oil investment funds, and you can make a plausible case that the world’s central banks and oil investment funds – the official sector – will finance the ENTIRE US current account deficit. I am a bit more pessimistic than most, but even I don’t expect the current account deficit will be much bigger than $900b in 2007.
Not all those official flows will show up in the US data. A lot of central banks are adding to their offshore dollar deposits. The world’s banks then lend those deposits out to various private actors who want to buy US assets. Some reserves are given over to private managers. And some central banks make use of London custodians. The survey data suggests that the TIC data (which is used in the BoP data) under-counted official purchases by about $100b back when the world was adding about $600b to its reserves annually. Scale that up to $150b for a world with $1 trillion in annual reserve growth.
The survey data also suggests that total private purchases of US debt are somewhat overstated. Either that or a lot of debt is just disappearing into a black hole. The “flow data” showed about $825b in total purchases of US long-term debt from mid June 05 to June 06 ($215b official, $610b private). The stock data suggests total purchases of more like $610b ($345b official, $270b private).
The flow data has its problems. But the data from q3 and q4 still suggests that private purchases of US bonds have fallen significantly while official purchases have picked up significantly relative to the previous four quarters. That seems unlikely to have changed in the first quarter. The US has grown more – not less – dependent on central bank financing as the US economy slowed.
I never expected China would be on track to add $500b to its reserves in a year, or the world a trillion. Never. Back in 2004, I though the world’s central banks wouldn’t sustain a $600-700b pace of increase. The numbers are so large that in some sense they are hard to believe. But I have double and tripled checked the numbers. They are in the data. The COFER data (especially if you add in SAMA’s non-reserve foreign assets). The IMF WEO data. The national data when the big central banks are added up. This is real.
This weekend in Washington, when the world’s finance ministers take a break from discussing Paul Wolfowitz in the corridors and deliver their prepared remarks, I suspect that many will talk about how we live in a world where private markets dwarf the funds available to the official sector, and in a world marked by floating exchange rates. That is safe rhetoric – the kind of rhetoric that echoes the conventional wisdom and never gets any staff person in trouble. Treat it with a grain of salt. It ain’t the entire truth.
Net official flows out of the emerging world absolutely dwarf the net private flows into the emerging world. Think a trillion of official flows going out (subtracting Japan from the global total but adding in the gulf’s investment funds) and maybe $300b in private flows going in. The world’s biggest current account deficit has been – in my judgment – financed almost exclusively by the official sector over the last two quarters.
And it is a bit hard to square $1 trillion in global reserve growth (the quarterly growth in both q4 06 and q1 07 really does look to be at that pace) with the argument that most central banks let their exchange rate float.
UPDATE: Logan, in the comments, notes that the state banks (or perhaps central Huijin) had the option of converting the "dollar" equity injections from 2003 and 2005 into yuan at the beginning of this year, and hypothesizes that this might explain the "surge" in q1 reserve growth. That is the best explanation I have heard. There are two sources for the fx swaps on the state banks balance sheet. As part of the recapitalization process, the PBoC promised to protect the yuan value of the dollar equity injections done in 2003 and 2005. Then in late 2005, the PBoC started doing swaps with the banks to withdraw yuan from circulation. These swaps likely continued in 2006 (and perhaps even in q1 2007). The unwinding of the swaps linked to the equity injection wouldn't have any impact on money growth and other monetary variables. The unwinding of the fx swaps that were part of the sterilization effort in late 2005 and 2006, by contrast, would be a de facto monetary expansion.
Consequently, I suspect Logan has found the source of the surprisingly strong q1 reserve growth: the conversion of the dollar equity injection into yuan fits the available facts better than any other explanation I have seen. The complete conversion of the equity injections would bring "underlying' reserve growth down to around $70 -- which is broadly consistent with the q4 data.
Update 2: jkh (in another comment) correctly points out that the unwinding of swaps linked to the recapitalization would still generate a monetary expansion unless offsetting actions were taken. Of course, China could have rolled the unwinding of the swap and the sterilization of the resulting increase in the money supply up in a single transaction, which would explain why Jon Anderson couldn't find enough "visible" sterilization to offset a $130b plus reserve increase.