Why not run a bigger trade deficit? (The August TIC data)

After all, the market seems willing to finance one!

That is the message of the TIC data.   Monthly inflows of $90 billion or so are large enough to sustain a current account deficit of $1080 billion, or a $820 billion deficit and lots of US FDI abroad.  Delphi can expand its operations in China ... 

Part of the reason for the strong headline TIC number is that US residents sold $17 billion of "foreign bonds," and brought those funds home.  But net foreign purchases of US debt were also strong - about $88 billion. 

We know that global reserve accumulation has been running at about $50 billion a month, or around $600 billion a year (number is adjusted for valuation gains).  China alone added around $20 billion a month to its reserves in the third quarter.  And last I checked, the big oil exporters were running big current account surpluses and adding to their reserves (and the state run oil investment funds) in a big way.  

China added $20 billion to its reserves in August, so it alone could have supplied the $4.4 billion in net inflows from central banks if it just invested 25% of its reserve increase in the US. 

Indeed, it probably accounted for the lion's share of that inflow.  According to the US data, Chinese investors, both public and private, bought a bit over $5 billion purchases in long-term treasuries, agencies and corporate bonds.  Demand for corporate bonds was particularly high.  China's central bank also increased its short-term Treasury holdings by around $4.3 billion (that would not show up in the headline TIC number, which only shows foreign purchases of long-term debt) and bought around $0.85 billion in foreign bonds from US investors - think dollar denominated bonds issued by other governments and the World bank.  

A 2/3 dollar, 1/3 non-dollar spit (see Morgan Stanley; Joachim Fels seems to think most central bank reserves remain in dollars) and $50 billion or so in global reserve accumulation implies roughly $33 billion a month in dollar reserve growth.  Some of that will show up in the banking data: both the onshore and offshore dollar deposits of central banks are growing.  But it also seems to be quite unlikely that central banks (and government run oil investment funds) only bought $4.4 billion of long-term US debt.

I'll give a concrete example, working not off the August data but rather off the data for the first half of the year that the Treasury has published.   Want to guess the total net purchases of long-term US debt securities in the first half of the year by Russia and the OPEC countries?  

Try around $5 billion.

Remember, those countries probably run a current account surplus of $200 billion in the first half of the year, maybe a bit less, and the global current account only balances if 75% of the entire world's current account surpluses makes its way back to the US to finance the US deficit.  That - ballpark - suggests something more like $150 billion in petrodollar financing for the US  (For more on petrodollar recycling, see this NY Times article).  Lots more than $5 billion. 

So what is going on?  Well, we don't really know.

The oil exporters may be stuffing the petrodollars in offshore dollar bank accounts, helping to finance private inflows. 

They may be buying US debt through intermediaries, so what registers as a Treasury bond bought by a private investor in the UK actually is owned by an investor in the Gulf.

Or they may be buying lots of euro debt, driving euro yields down, widening the differential between US and European long-term rates and thus inducing private European investors to shift into dollars. 

Probably a bit of all three.

But it would be a mistake, I think, to conclude from the TIC data that the US is no longer dependent on foreign central bank support.  The gap between the known increase in global reserves and the recorded central bank inflows to the US is simply too big.