The recent custodial data from the Fed leaves no doubt that foreign central banks have dramatically reduced their Treasury holdings in August.    Thanks to Russ Winter, I realized that the New York Fed reports two numbers for foreign custodial holdings – the average holdings over the course of the week, and the number on the end of the (reporting) week.   Using the end of week data, the Fed’s custodial holdings of Treasuries fell by $44.2b from August 1 to August 23.   That’s big.    Most of the fall came in the past two weeks. 

Custodial holdings of Agencies rose by $12.6b – offsetting some of the fall in Treasury holdings.   But overall central bank custodial holdings still fell significantly – by close to $30b.   That hasn’t happened for a while.

And, obviously, central banks reduced Treasury holdings didn’t exactly imply reduced demand for Treasuries.   Treasury yields fell (and prices rose).  The ten year yield fell from 4.8% or so to 4.6%; T-bill yields went from around 5% to around 4% with a little detour down to 3% on Monday.

That, on the surface, seems like a refutation of the argument that central bank demand has played a key role in keeping Treasury yields down over the past few years. 

So what is happening?

Well, there obviously has been a bit of a liquidity crisis, as investors lost confidence in a lot of CDOs — and financial firms that borrowed in the money market to purchase CDOs.   The total stock of outstanding commercial paper fell by about $200b over the past two weeks – and a lot of money that wasn’t reinvested as commercial paper matured seems to have flowed into the Treasury market. 

Foreign central banks, judging from the Fed’s data, helped meet that demand.  

Some may have just recognized a good trading opportunity – and sold their Treasuries at profit.   Some may shifted into cash.  Why roll over short-term t-bills at 3% when too-big-to-fail banks are offering more?   And some may even have shifted into risky assets – though my guess is that not many were quite that bold. 

But something else was going on as well.   The same flight from risk observed in the ABCP market took place globally – money that previously had been bet on emerging markets was taken off the table.   And a few funds probably needed to cash in their paper profits on their emerging market investments to cover losses elsewhere.  

The net result:  a big change in the global flow of funds.   A lot of (private) money had been flowing out of the US, Europe and Japan into emerging markets.   Gross (private) capital inflows to emerging markets set a record in 2006 – and almost certainly were above that pace in the first half of 2007 (see the IIF’s data). 

But as most readers of this blog know (I hope), emerging economies collectively run a large current account surplus.  While some countries in Eastern Europe run external deficits and need ongoing inflows, total (private) capital inflows to the emerging world far exceeded Eastern Europe’s need for financing.  Most emerging economies used the inflow to build up their reserves – not to finance current account deficits.

The results: private capital flows into the emerging world were matched, almost dollar for dollar, by official capital flows out of the emerging world.  A private investor looking to buy Russian equities needed to trade dollars (or euros or yen) for rubles, and Russia’s central bank bought dollars and issued rubles to meet that demand. 

Russia’s central bank then bought Agencies and Treasuries and the like.   So did many other central banks. 

And in fairness to Macro man (who is convinced that central banks are the reason for euro and pound strength) I should note that Russia holds a relatively high share of euros and pounds in its reserves.  To the extent that those betting on Russia borrowed dollars to buy ruble denominated assets, they were indirectly supporting the euro.   Something over 50 cents of every dollar going into Russia was invested in euros and pounds.    

Of course, something under 50 cents of every euro flowing into Russia was invested back in dollars.  The overall impact of inflows into Russia (and the resulting central bank outflows) on the euro/ dollar depends heavily on how that trade was “funded”, to use the market argot.

Over the past couple of weeks this entire process went into reverse.  

Private money moved out of emerging economies.  Emerging economy central banks sold some of the treasuries (and bunds) that they had bought with these inflows. 

Look at Russia.   Central bank data from the second full week of August shows a $5b fall in Russia's reserves.  The euro’s 1.4% fall against the dollar that week — and the resulting fall in the dollar value of Russia's euros — explains about 1/2 the fall.  The rest reflects capital outflows from Russia that were financed by the sale of Russian central bank assets.  Earlier this week, Russia sold another $3b to stabilize the ruble – perhaps because a few Russian consumer lenders seem to have a large fx mismatch on their balance sheet.

Brazil hasn’t sold reserves – but the pace of its reserve accumulation has slowed noticeably.   That is true across a range of emerging economies.  

Under these conditions, I wouldn’t expect a fall in central bank purchases to push up Treasury yields.   Treasuries are rising precisely because private investors who previously shunned (low-yielding) Treasuries in favor of CDOS (and ABCP that financed CDOs) and emerging economies reversed course.  

Or, put differently, if Treasuries weren’t rising in value (i.e. their yield wasn’t falling), central banks wouldn’t be selling. 

I suspect that emerging economies as a whole are still adding to their official assets – though obviously not their custodial account at the New York fed.    Custodial holdings are an imperfect proxy for reserve growth at best.

However, the growth in official assets is now coming exclusively from those emerging economies with large trade and current account surpluses, not from those economies that previously had been building up reserves to offset private inflows. 

Who then?

China.  It still has a large current account surplus. 

And the Gulf.  Oil prices are still rather high.   But the Gulf tends to invest through oil investment funds that don’t hold Treasuries at the Fed.

Russia obviously benefits from high oil prices too.   It though has attracted a lot more foreign portfolio investment in the first half of 07 than the Gulf.  And for now, portfolio outflows exceed its current account surplus.