Thread by @Brad_Setser

Would like to see a bit more backing for this:

"Foreigners sold more than $500bn of US government bonds during the second quarter, and perhaps one-third of that was unloaded by Chinese entities." (From Michael Howell)

TIC data stops April, so how were May/June estimated?

1/n

The flow that is unambiguously in the reported data (only through April) is a shift from Treasuries to Agencies.

Transactional data is imperfect, but if anything it is likely that it slightly understates China's Agency purchases

2/n

A lot of China's purchases (both of Treasuries and Agencies) have historically shown up in the transactional data as UK purchases

3/n

With the Fed buying lots of long-term Treasuries (more than the Treasury issued in q2), the stock in the market fell and China could well have reduced its holdings --

but on q1 trends it would be moving into Agencies not into euros

4/n

And this claim -- that Chinese flows reduced Treasury yields in 15/16 -- runs against the visible flow data.

The rise in private outflows from China then was met by a big fall in China's reserves, and large selling of Treasuries.

5/n

The mechanisms through which China held down Treasury yields (global weakness, concerns about deflation with a falling CNY) then were indirect, and not tied to a direct Chinese bid

6/n

A shift in China's portfolio toward the euro would be big news -- and form my point of view, a welcome one. The dollar has been too strong for the long-run health of the US economy since 2014.

But need a bit more evidence ...

7/7

ft.com/content/8e7379…