The marginal buyer, on a bench

Two instruments from the article. The first decides whether Chinese bank demand for Treasuries exists at all. The second decides which channel compresses valuations first — and in what year.

1 · The demand gate

The carry trade is an AND gate. Both conditions must hold: the spread has to pay, and the yuan has to be pushing the wrong way. Lose either input and the marginal buyer leaves.

What banks pay large accounts to keep dollars onshore
Left: depreciation pressure · Right: appreciation pressure
Carry spread +1.36 pts Retention motive yuan appreciating and demand

Demand on. Banks bid for deposits and buy duration. This is the state Reuters reported — and it is reversible on either input.

What flips the gate: the PBOC caps deposit rates, depositors leave, the spread closes, or the yuan turns and the retention motive disappears while the carry still pays — the test of the article's central claim.

2 · The transmission board

Set the rate regime. The board reports the state of each compression channel and, for the two mechanical ones, the year they bite. The cohort is corporate America outside the mega-caps, with flat operating earnings — the assumption is stated in the notes below.

The 2020–21 term-out being defended

Discount rate

Live pressure, no date. It compresses only when the growth story cracks — sentiment holds the door.

−28%implied multiple gap

Refinancing

The coupon stack rolls on a calendar, not on sentiment.

Year 4coverage breaks 3×

Credit

Spreads anticipate the wall. Credit fires before equity notices.

Year 3spreads reprice

When each channel bites

refinancing bites credit reprices carry demand present at this yield

First to fire: credit, year 3. Spreads widen while equity multiples are still holding — the leading indicator the article commits to.

Blended coupon as the stack rolls