Dwindling foreign demand for Treasuries …

Floyd Norris of the New York Times highlights something that was also on my radar screen:  a sharp fall off in foreign demand for US treasury bonds. For the first time in years, the US budget deficit is being financed by domestic investors.

Foreign inflows to the US haven’t fallen off, to be sure.   But the composition of these inflows has changed.  Foreigners are buying more agencies and US corporate debt and fewer Treasuries.  First quarter data is here; the TIC data from April and May do not indicate that the story has changed. 

Why?

Well, one theory is that foreign central banks have stopped financing the US.   I don’t believe that.  I do believe that the set of central banks now adding to their reserves is far different than the set of central banks adding to their reserves in 2003 and 2004.  Oil exporters have replaced Japan and (to a lesser extent) many Asian countries other than China.    And as I almost constantly note, the oil exporters generally don’t show up in the US capital inflows data directly.   Their presence is felt, but not directly observed.

In 2005, there was a sharp fall in (recorded) central bank demand for Treasuries, a fall off the coincided with the end of Japanese central bank purchases in the fall of 2004 (it took Japan a while to place the huge sums of dollars bought in late 03/ early 04 in the bond market).   However, there was no fall-off in overall foreign demand for Treasuries.   Some of that demand came from the Caribbean – think hedge funds.  But lots came from London – think oil state central banks and investment funds. 

In 2006, both central banks and private (or not so private) investors abroad stopped buying Treasures.   What happened?     Here is my guess:

• Caribbean holder of Treasuries sold in the first quarter.  Net sales were around $8.5b.  See the Treasury bulletin.  Hedge funds unwinding curve flattening trades?   
• Some oil exporters reduced their Treasury purchases.   Russia isn’t the culprit: it never bought treasuries, only agencies for some reason – but it has reduced its agency purchases this year.     But the others must have shifted their portfolios around a bit – either buying more agencies and corporate debt, or just parking their cash in bank deposits.  Don't look in the offiical data to confirm this -- OPEC holdings of Treasuries are up.  But the total observed flows here are so small relative to the oil surplus that it is likely that most of these flows show up in the data from London.  Looking only at what shows up in the US data as "OPEC" holdings can be misleading.
• Some central banks may have opted to hold short-term t-bills rather than longer-term bonds.   Though the TIC data doesn’t support this story in the aggregate – there hasn’t been a surge in overall central bank holdings of Treasury bills.
• Norway sold.
• Japan seems to be shifting its reserve holdings slowly out of the Treasury market and into the Agency market.  At least that is my interpretation of the data from the New York Fed’s custodial accounts.  Look at the 52 week changes in Agency and Treasury holdings.  This story is supported by the q1 data in the Treasury bulletin: Japanese investors sold $19b of Treasuries and bought $11b of agencies.  That data aggregates the activity of the central bank and private investors, but it is suggestive. 

Sum it all up, and all these changes seem to have had an impact – net flows into the Treasury market from abroad this year have been negligible.    If you take out the Caribbean, some of this effect goes away.   China clearly has been a net buyer ($15b in q1).    But there still have been something of a shift.

Read Norris.   And look at his graphs in the Saturday New York Times if you have a chance.  I was planning to do something similar, but he beat me to the punch in a big way.  Though there are a couple of fun twists that could be added to the Norris graphs.