Do not forget about Norway - but also don’t put too much weight on the details of the TIC data.

The May surprise in the TIC data:  foreign central banks withdrew financing from the US.  The sold more Treasuries (14.3b) than they bought Agencies and corporate debt.  Net central bank flows into long-term debt were a negative $1.4b. The reason: Norway.   Norway – presumably its oil fund – sold $11.77b in Treasuries in May.

The other surprises: relatively strong outflows from the US continued in May, though US investors bought foreign bonds ($14.3b) rather than equities ($4.9b, down from an average of around $10b for the past few months); and private investors abroad suddenly rediscovered the joy of holding Treasuries.  After several months of muted purchases, they bought $27.5b in May.

On the other hand, I am increasingly frustrated with the TIC data, since it rather clearly isn’t picking up much of what is going on in the world

The big problem: the folks who have money and who must be doing the buying just aren’t showing up in the data.

That’s true on several levels.  A set of central banks I track added $90b to their reserves in May.    China and Russia added around $50b between them.  A bit of that was valuation, but not most of it.   Yet even after netting out Norway, official inflows only totaled a positive 10.4b.   That’s only 20% of China and Russian headline reserve growth, and an even smaller share of my estimate of global reserve growth.  And the difference isn’t showing up in the short-term data.   Chinese short-term claims are up $0.7b but Russian short-term claims are down $1.6b.    

The Bank of Japan seems to be moving out of Treasuries into Agencies.  But recorded inflows from Japan consistently have seemed small relative to Japan’s current account surplus.    This month isn’t a particularly good example: Japan bought $1.7b of Treasuries, $8.7b of Agencies and $0.3b of US corporate debt, for total purchases of around $10.7b.  Q1 is a better case.  In the first quarter, Japan sold $19.1b of Treasuries, bought $10.9b of agencies and bought 2.3b of corporate debt, for a new flow of negative $5.9b (data is in the Treasury Bulletin).    That from a big surplus country.

China shows up in the data in the expected way.   But on a scale that is a bit too small.  In May, China bought $0.9b of l-term Treasuries, increased its short-term Treasury holdings by $2b (largely by running down its bank deposits), bought $4.4b of Agencies and $2b of corporate debt.   The net inflow was (ballpark) around $9.3b.  That is still small relative to its $30b headline reserve growth in May – most of which didn’t come from valuation gains. 

Russia used to show up in the data, but no more – whether because of a new investment strategy or diversification away from the dollar.  Its short-term holdings are down about $8b so far this year, and the increase in its long-term holdings doesn’t offset this.  In May, Russia sold 0.4b in Treasuries and bought $1.6b in Agencies.  Chump change relative to its reserve growth.

And don’t get me started about the Middle East.

If you combine Japan, China, Russia and the Middle East, you pretty much have the world’s current account surplus.    Europeans are buying US debt, tis true – but those purchases are financed in some sense by inflows into Europe from Japan, China, Russia and the Middle East!

p.s. RGE premium subscribers interested in my more professional analysis of the reserve and capital inflow data should check out the Russia and China reserve watches. Apologies for the ad, but to find these items it helps to know where to look.