The February TIC data

And the money keeps rolling in ...

In February, the US raised $84.5 billion in financing abroad through the sale of long-term securities. Net sales of debt ("debt exports") of $92.5 billion let the US finance a net equity outflow of $8 billion, as US purchases of foreign equities exceeded foreign purchases of US equities.

$85 billion annualized works out to $1020 billion -- which is about what the US will need to finance a current account deficit in the $850-900 billion range AND continued (one assumes) net FDI outflows.

In other words, after a bit of foreign interest in the US equity market in January, old patterns reasserted themselves in February. The US financed itself exclusively by the sale of long-term debt securities.

Foreign interest in Treasuries was particularly strong -- "private" purchases of $31.2 billion and official purchases of close to $11.3 billion add up to $42.5 billion. That’s a large sum -- annualized, it is close to $510 billion, more than enough to finance this year’s fiscal deficit. What’s more, only $1 billion of that inflow seems to have come from Japan. Judging from the changes in stocks, the big buyers of Treasuries seem to have to been in the UK ($10 b), the Caribbean ($11 b), and places like Switzerland, Canada and Ireland (around $8 b). The buying attributed to the UK could have been from anywhere -- the Middle East, Asia, or even elsewhere in Europe. The Caribbean and Ireland are hedge fund havens, so those flows may not provide any net financing (the purchases were financed by a loan from a US bank or broker-dealer, so there is an offsetting outflow). The Canadians almost certainly were buying for their own accounts though ... IMPORTANT UPDATE. Brian Garvey of State Street has drawn my attention to the Chinese flow data, which suggests very strong Chinese buying of both Treasuries ($9.1 billion) and Agencies ($4.7 billion). Total net Chinese buying, by my calculations, was $16.3 billion. I am not quite sure why so little of that "flow" shows up in the "stock" data -- maybe valuation adjustments because of rising interest rates on Treasuries during the month of February? It certainly seems that China starting putting some of its huge q4 reserve buildup to work in February, and did so in ways the US data reporting system measured. END OF UPDATE

I was a bit surprised that reported foreign central bank purchases only totaled $18.5 billion ($11.3 b in treasuries, $5.2 billion in agencies). The 11 biggest emerging economies added about $45 billion to their reserves in February, so there was more money out there. Plus, the foreign central banks probably still need to invest some of the funds they built up during the reserve buying spree in the fourth quarter. UPDATE: this is all the more surprising in light of the Chinese buying. Were the state banks doing the buying? END OF UPDATE

But that is not the primary reason why I was surprised -- the US data consistently fails to pick up all central bank purchases. I was surprised because the Federal Reserve’s custodial data showed a much bigger increase in central bank holdings between February 3 and March 3 (the dates on the Fed’s data, admittedly, don’t quite match up with the TIC data). The Fed data showed a $18.1 billion increase their custodial holdings of Treasuries, and a $14.2 billion increase in custodial holdings of Agencies -- for a total of $32.3 billion.

I am told that there are lots of reasons why the Fed data and the TIC data do not match up -- foreign central banks can, for example, shift funds between private accounts and their custodial holdings at the Fed without necessarily adding to their total reported holdings or bringing funds into the US. But I was still surprised by the size of the gap.

Bottom line: there is no perfect measure of even the direct impact of foreign central banks on the US market.