Foreign demand for US debt was kind of weak in the second quarter

• Net long-term capital inflows to the US in the second quarter.  $185.6
• Q2 trade deficit.  $193.1b
• Estimated Q2 current account deficit -- $215b-225b (personally, I think it will be toward the high end, as I suspect the income balance will turn negative)
• Q2 US financing need (assuming the net FDI outflows of q1 continue) = $240b-$250b 

I know $75b in net inflows in June topped market expectations.   But it is still a pretty bad number.   Foreign demand for US debt seems to be slowly waning.   Admittedly from a very, very high level.  The rolling 12 month total for foreign purchases of US long-term securities is now $1075b.    But it hasn’t been increasing over the past few months.  And the US needs big inflows to allow it to both invest abroad (buying foreign equities as well as financing FDI investment) and run a roughly $900b current account deficit.

Foreign demand for US debt (foreign demand for US equities remains trivial) was basically constant at around $85b in both May and June.  In June, though, American purchases of foreign securities fell from around $20b to around $10b.  That increased net inflows. 

US purchases of foreign equities basically stopped in June.   The trend is pretty clear.

• January $10b
• February $12b
• March: $12b
• April $8b
• May $5b
• June: -$1b 

We all know why – there was a bit of turmoil in emerging market equities in May and June.  That turmoil has now passed (just look at the Turkish lira).  I actually had expected US purchases of foreign equities to be more negative in June.  The US sold about $1 billion in foreign equities, less that I would have guessed.  The market turmoil seems to have come from a marked slowdown in US purchases, not large net US sales …  (edited)

Of course, there are lots of ways of financing the purchase of foreign equities offshore, and the unwinding of those positions wouldn’t necessarily show up in the US data. 

Many have noted the fall off in official inflows.    Foreign central banks bought $2.3b of US debt in June, after selling 1.6b in May.  Total purchases of long-term debt for q1 were around $22.2b.    Central bank holdings of short-term Treasuries fell by $27.5b in q2 (from $215.5 to $188.8b), with $7b of that fall coming in June.  So net inflows from central banks – using this measure – were negative in q2.  This measure counts long-term purchases and short-term purchases of treasuries, but leaves out Russian and other purchases of short-term Agency bonds and the growth in bank deposits of foreign central banks.   

No matter.   Even if you add in the increase in Russian short-term holdings ($7b), net central bank flows are still pretty much flat. 

Let me state this in as a clear terms as possible.

This makes no economic sense.  By my measure (the actual data is hidden behind RGE’s firewall, but it is a solid number), central bank reserves increased by $175b in the second quarter. 

Russian and China combined to add about $100b to their reserves in the second quarter. 

And the Gulf countries official investment funds are absolutely flush with cash. 

That money has to be going somewhere.    It just isn’t showing up in the TIC data.

The low level of central bank financing for June sort of makes sense – turmoil in emerging markets did slow global reserve growth substantially.  But the low numbers for April and May make absolutely no sense.  Global reserve growth was very, very strong in those months.

What does show up in the TIC data? 

About $11b in Chinese purchases of US debt in June (over $5b in purchase of Agencies) in June, with another $1b of Chinese purchases of foreign debt held by  US investors.  I think the PBoC has been buying some of PIMCO’s portfolio of dollar-denominated emerging market debt, as well as lots of World Bank bonds …

About $1.5b in Russian purchases of US long-term debt in June, along with  a roughly $5b increase in Russian short-term holdings. 

In the second quarter, Russia bought about $4b in US long-term debt (mostly Agencies – see a trend?) and added about $7b to its short-term holdings.   That is a change from the first quarter, when it reduced its short-term holdings by about $10b.  It bought $5.5b in long-term debt, but its net holdings of US debt still fell by between $4.5b and $5b.   

That suggests to me that Russian reduced the dollar share of its reserve portfolio in q1, though some of reduction could have come in q2.   $11b in net Russian purchases in q2 is small relative to Russia’s $40b or so reserve growth in that quarter.    Russia’s headline reserve increase was actually closer to $45b, but it enjoyed some valuation gains … 

Total inflows from Russia – recorded flows that is – have been quite weak over the course of the year.   Net inflows from Russia in the first half of the year are now around $5-$6b.  Which is small relative Russia’s $60b or so in total reserve growth, even taking into account the fact that Russia reduced its dollar allocation from 60 to 50% of its reserves.

That hints at the real problem with the TIC data right now – the TIC data just isn’t capturing the US debt purchased by other oil states.  And they account for a very large share of the world’s current account surplus and the world’s reserve growth.