The dollar and the world’s central banks (once again)

Macro Man is back from vacation, and I have little to add to his explanation for the dollar’s rebound.

The gap between the United States economic performance and the rest of the world’s economic performance looks set to narrow -- which makes the dollar a bit less unattractive and other countries’ assets a bit less desirable. The gap between US rates and other large countries interest rates may fall, as other G-10 central banks start to ease. Finally, I have long been convinced -- in part because of a good Goldman Sachs paper on the topic -- that high oil prices are bad for the dollar. A weak dollar may also be good for oil, though I am less convinced on that point. I would put it differently: the Fed’s easing was bad for the dollar and it was good for oil, in part because a host of rapidly growing countries with subsidized oil prices followed the Fed and adopted extremely easy monetary policies that helped (for a while) spur oil demand.

UPDATE: Things look a bit different on Thursday than on Wednesday, with oil back up ..

That said, the US external deficit remains far larger than the deficit private investors abroad want to finance at current US interest rates. The June TIC data (released last Friday) was rather weak. Indeed, right now the US is having trouble consistently producing assets the rest of the world wants to buy. But CDOs composed of tranches of mortgage backed securities based on subprime loans practically have to be given away. The stock of US banks and broker dealers hasn’t seemed like such a good deal recently. Foreign demand for US stocks has dipped recently.

Of course, the US doesn’t rely exclusively (or even heavily) on private demand for its debt and equity for financing. In some sense, it cannot. Not when a set of surplus countries’ still have undervalued currencies. The FT leader notes: "global imbalances between the US and currencies pegged to the dollar are not yet fully resolved. A faster appreciation of Asian currencies still makes a lot of sense, as does a re-peg of oil exporters’ exchange rates to a basket of currencies." The big surplus countries right now -- China and the oil exporters -- channel almost all of their surplus into their central banks and sovereign funds. The growth in China’s government assets exceeds its (still large) current account surplus. The same was true of Russia in the second quarter (the third quarter may be different). And almost all of the Gulf’s oil surplus is channeled through SAMA and the Gulf’s three big sovereign funds (ADIA, KIA and QIA). The concentration of Chinese and Gulf foreign assets in state hands implies that the buildup of official claims from the surplus countries will play a large role financing the deficits in the deficit countries.

Central banks aren’t quite as keen on Agencies as they used to be. But central banks still are buying a lot of Treasuries.

Macro Man’s notes one additional reason for the dollar’s rally: central banks are not selling dollars for euros quite as rapidly as they once did. He calls this "addition by subtraction."

What might explain this?

One answer is obvious: many central banks are no longer buying dollars in the market, and thus have no need to sell dollars for euros to meet their portfolio targets. Many Asian central banks -- with the obvious exception of China -- have been selling dollars recently, as they fight pressure for their currencies to depreciate.

I would bet that the latest data on Russia’s reserves (out tomorrow) will show that Russia too has been selling dollars. And given that Russia keeps less than 50% of its reserves in dollars, Russia generally needs to sell dollars for euros whenever its reserves are rising.

But two big players likely are still adding to their reserves and sovereign funds: China and the countries of the Gulf. So why might they have stopped selling dollars for euros?

Well, it seems -- as Macro Man notes (in the comments) that speculative pressure on the Gulf currencies has disappeared recently. The Gulf releases data with a huge lag, but my guess is that the unwinding of speculative bets on the GCC currencies has led money to move out of the region, and these outflows have offset ongoing inflows from still high oil prices. So for the time being, the Gulf might not be adding to its foreign assets quite as fast as you might think.

And then there is China. China’s current account surplus and ongoing FDI inflows imply something like $40 billion of Chinese reserve growth a month, barring large "hot money" outflows. That incidentally is more than the Gulf with oil at $125, let alone with oil at $110. So why hasn’t China been more active in the euro-dollar market?

Two ideas come to mind:

One is that China concluded that selling dollars for euros -- whether to meet an existing portfolio target in the face of rapid reserve growth or in an effort to reduce the dollar’s share of its portfolio -- was a mug’s game. The more China sold, the more it pushed the dollar down -- and the faster money poured into China in the expectation that the RMB wouldn’t follow the dollar down. Supporting the dollar by ending dollar sales might help change the dynamics that had pushed Chinese reserve growth up to such high levels ...

The other is that China’s decision to halt RMB appreciation against the dollar (the RMB has actually depreciated recently against the dollar, though not enough to keep it from appreciating significantly against the euro) combined with tighter controls on capital inflows have reduced hot money inflows into China, and thus slowed the exceptionally rapid growth in China’s foreign assets.

Take your bet. China hasn’t released data for July, so there is no real way of knowing.

If the funds moving into China to bet on the RMB had previously been held as dollars -- and if China was selling some of the dollars coming in for euros to meet its portfolio target -- the net flow from hot money inflows into China and associated reserve outflows might have been euro positive. And a fall in those flows conversely might be euro negative.

I am not sure on this; it is just a guess.

One final point: the IMF’s COFER data suggests that those emerging markets that report data on the currency composition of their reserves have held the dollar share of their reserves roughly constant (at around 60%) since the end of 2003, despite the dollar’s slide. That implies that they have bought proportionally more dollars when the dollar is falling than when the dollar is rising, and conversely, bought more proportionally more euros when the euro is falling than when it was rising.

If that pattern holds, central banks should increase the share of their new reserves doing into euros to keep the dollar’s share of their portfolio from rising.

Then there is a second dynamic: the IMF’s data suggests much stronger overall growth in the emerging world’s reserves when the dollar is under pressure than when the dollar is rising (graphs and data here, and here). There is a bit of noise -- the price of oil matters, and oil rose in 2005 along with the dollar pushing up overall reserve growth -- but it is a fairly strong correlation. It makes sense too. The same forces that pushed the dollar down v the euro were present against other currencies, but in Asia and elsewhere central banks stepped in to keep their currencies from rising and the dollar from falling.

On one hand reserve growth likely has slowed -- which means that central banks will be selling fewer dollars for euros on an ongoing basis to meet their portfolio targets. On the other hand, past central banks that rigorously target a constant dollar share in their portfolios will need to put a larger share of their reserve growth into euros now than the euro is falling than when the euro was rising.

That gets us back to the debate over whether central banks were -- back when the dollar was falling -- supporting the dollar against the euro (by adding proportionally more dollars to their rapidly growing reserves, and thus supplying more financing to US than ever before) or hurting the dollar against the euro (by selling more dollars for euros than ever before to meet their portfolio targets).

Or to be concrete, suppose China added $300 billion to its reserves in 2005, and put 40% into euros and 60% into dollars. And suppose it intervenes in the first instance exclusively by buying dollars. It would have needed to sell $120 billion dollars for euros. Now fast forward to 2007. Now assume that China added $800 billion to its foreign assets over the last four quarters, and put 25% into euros, implying dollar sales of $200 billion. What mattered more, the larger absolute sale of dollars or the rise in China’s dollar reserve growth from (in a two currency world) from $180 billion to $600 billion?

These numbers are made up. I obviously don’t know.

In any case, it will be interested to see what the IMF data shows for the third quarter of 2008 when it finally comes out in December.

And even then it won’t really provide the answers, as the countries that are adding to their reserves now are also among the countries that do not seem to report any data on the currency composition of their reserves to the IMF.