The global flow of funds in 2003 was pretty easy to figure out once the BIS annual report come out. The US ran a current account deficit of $530 billion. Most other countries - though not Australia, the UK and a set of Eastern European economies -- ran current account surpluses. Most of the US current account deficit was financed by $442 billion in dollar reserve accumulation by the world’s central banks ($487 billion if you include the $45 billion transferred to a couple of Chinese state banks).
Overall reserves increased by a bit more than $442 billion in 2003. The world’s central banks bought around $60 billion of euros and yen, and the value of the central banks’ pre-existing holdings of euros and yen increased by around $100 billion because of moves in the dollar/ euro and dollar/yen. So total reserves increased by $615 billion. The world’s central banks ended 2003 with about $3 trillion in reserves, $2.1 trillion in dollars, and $0.9 trillion in euros, yen and everything else.
But the basic flow of funds was simple. The central banks of the world captured most of the world’s current account surplus and invested most of that surplus in dollars, financing the US in the process. Since Asian reserves went up by more ($475 billion) than Asia’s current account surplus ($310 billion), Asian central banks actually did more than their share.
This year is a bit more of a conundrum. The US current account deficit has widened significantly, to at least $650 billion. If you lump together Europe’s deficit countries (UK, Eastern Europe) with Europe’s surplus countries (everyone else -- Euroland, Switzerland, the Nordics), Europe as a whole ran a surplus. Australia and New Zealand have deficits, but they are too small to put a huge dent in the global flows. More or less all of every major region’s current account surplus ultimately has to find its way back to the US. The US likely accounted for something like $650-660 billion of the $690 billion total current account deficit run up by deficit countries.
Global reserves continued to increase. Asia’s reserves are up by at least $535 billion in 2004. Global reserves probably increased by $700 billion, to 3.8 trillion (JP Morgan’s estimate, reported here). Only $70-80 billion of that likely comes from valuation gains. Central banks probably bought around $620 billion in new reserves.
But did it go into dollars? That would be by far the easiest way to fund the US current account deficit.
Alas, we know that some central banks were not buying many dollars. Thailand cut the share of dollars in its reserves from 80% to 50%, according to the FT, which implies that at the margin, it was buying euros. And the Thais claim others -- dollar peggers like Malaysia and Hong Kong as well as India and Singapore -- tried to follow suit. The $2 billion in valuation gains Malaysia reported in 2004 on its reserves (which rose $22 billion in 2004 to $67 billion) certainly suggest that Malaysia had a substantial share of its reserves in euros (and yen) by the middle of this year. Interesting.
But someone still has to lend to the US, not to Europe.
ABN Amro’s Norfield estimates dollar reserves only increased by $300 billion in 2004. That is certainly too low. Recorded inflows to the US from central banks in the first three quarters were $263 billion, and given the scale of global reserve accumulation in q4, it would be a surprise if the US did not get at least another $60 billion in reserve financing in q4. That puts the US at $320 billion -- and we know that in the past, BIS data on dollar reserve accumulation has exceeded recorded inflows into the US by a substantial margin. I would be surprised by a number smaller than $400 billion.
But if the world’s central banks bought $400 billion in dollars and $200 billion in euros (and pound and Swiss franc) in 2004, that still implies a $200 billion inflow into "Europe" writ large. That $200 billion inflow did not go to fund a broad European current account deficit. According the IMF’s data, "Europe" writ large (including the UK and the deficit countries of eastern Europe) had a current account surplus (the fall WEO estimated a surplus of $85 billion). Combine the postulated central bank inflow into "Europe" with "Europe’s" current account surplus, and Europe has also almost $300 billion to lend to the rest of the world.
In some way or another, that had to find its way back to the US. The US need for financing is giant sucking sound absorbing all the world’s spare savings.
How could that have happened?
Private European investors could have been buying dollar assets, and taking the currency risk. They may have concluded the euro had already risen as much as it was going to, and its next move would be down (so far, they have been right in 2005 -- even if they were wrong for 2004).
Private European investors could have been buying dollar assets, and buying insurance against moves in the euro/dollar from someone else. The risk the dollar might fall v. the euro would be hedged -- but that implies someone is supplying the hedge, and is taking the dollar risk?
US firms might have sold euro denominated debt, in effect providing the hedge European investors wanted?
European private investors could have used the inflow of Euro reserves to finance the accumulation of private assets elsewhere in the world, say Asia. And then private investors in Asia used the inflow from Europe to buy US assets, and thus finance the US current account deficit.
At the end of the day, the sheer size of the US current account deficit implies that almost all of the rest of the world’s current account surplus has to be lent back to the US. So if the world’s central banks are providing less of the financing the US needs, and supplying financing to Europe that Europe does not need, then some private actors somewhere in the world have to be taking the risks associated with lending to the US in dollars.
And they have to be lending to the US in a reasonably big way. With $400 billin in financing from the world’s central banks, the US still need $250 billion in financing from private investors to run a $650 billion plus current account deficit. Actually, the US needs $400 billion, $250 billion for the current account another $150 billion or so to finance net equity outflows (both FDI and portfolio equity) ... I do wonder which private investors have been willing to make such a large bet on the dollar at current interest rates.
One last point. If Norfield is right and dollar reserve accumulation is only $130 billion, in 2005, there is no way the US will be able to fund a current account deficit of $800 billion -- yet barring some big improvement in the monthly trade balance, that’s what we are looking at -- and it could be worse. $130 billion in financing from the world’s central banks would be a HUGE change. I have no idea how Norfield came up with that estimate: it may reflect the world’s central banks wishes, but it seems inconsistent with the United States’ need for cheap funding.