The Bond Market Is Screaming. Stocks Aren't Listening — Yet.

Article Ticker Tape

The fear on your feed is real, and it is not (yet) an equity story. The 30-year Treasury yield hit 5.30–5.34% on Monday and Tuesday, its highest since 2007 and within reach of the 2002 highs; the 10-year sits at 4.72–4.75%, also a post-2007 high; Japan's 10-year is at a 30-year high, and long bonds in France, Germany and the UK printed multi-decade highs on the same day (CNBC, Bloomberg). Meanwhile the S&P 500 closed Monday 0.86% off its all-time high and high-yield credit sits 0.31% off its own high (desk summaries, 2026-08-17 close). Bonds are having a crisis. Stocks and credit are not confirming it.

What's actually true, and what isn't

True and material: the 30-year at 5.30%+ (CNBC's yield table, 08-18); a bear steepener — the 2-year fell to 4.175% while the long end rose; the last 30-year auction cleared at its highest yield since 2001 and five of the last seven 20-year auctions tailed (BMO via CNBC); the US diesel crack at a record $102 a barrel, first ever above $100 (Reuters, 08-17); the Strategic Petroleum Reserve at its lowest since 1982 (Reuters/DoE) — a scheduled drawdown, not a surprise; and the Iran window: the June MOU's 60-day clock expired Monday with no extension and Hormuz crossings at three a day (Kpler via CNBC).

True number, wrong framing: the yen has retraced roughly a third to a half of the July 30 US–Japan intervention (163.73 → 157.57 → about 159.5), not "completely wiped out." Oil was flat last week on the ETF tape, up 2.91% Monday — not "up 6%."

Not true: the "Vance ends the dollar" clip is from 2023; the "BOJ apology" account is not a BOJ official; the 2007 → 57% crash analogy is a scare, not a mechanism — see below.

Tape, 2026-08-17 close Where it sits Read
SPY $772.67 −0.86% from 52-wk high, RSI 61.6, strong-up, 200-day ~$703 the equity market has priced none of this
HYG $79.61 −0.31% from 52-wk high, RSI 57.1, above 50- and 200-day credit is not flinching — the tell
TLT $81.41 −8.35% from 52-wk high, RSI 35.3, death cross, downtrend the pain lives in duration
VLO $347.19 / PSX $240.55 RSI 74.8 / 80.0, at the highs, +10–12% in a week the diesel-crack trade is already made

So when does it get "priced in"?

Not at a yield level. At the yield peak. Take every episode since 2007 where the long end spiked to a cycle high, mark the day long-bond prices bottomed, and look at what the S&P did around it (daily closes; the yields are the well-known cycle highs):

Yield peak S&P into the peak S&P after the peak Sessions to the low Total drawdown 12 months later
Jun 12, 2007 (30-yr ≈5.4%) −2.9% +4.6% to a new high (Oct 9), then −56.5% 438 −55.8% −10.5%
Nov 2, 2018 (10-yr ≈3.25%) −7.4% −13.8% 34 −20.2% +12.6%
Oct 21, 2022 (10-yr ≈4.25%) −21.6% −0.9% 9 −22.3% +12.5%
Oct 19, 2023 (10-yr ≈5.0%) −6.9% −3.7% 6 −10.3% +37.1%
Jan 14, 2025 (10-yr ≈4.8%) −4.2% −14.7% (April tariffs) 58 −18.3% +18.6%

Three things fall out. When rates are the story (2018, 2022, 2023), stocks re-rate on the way up in yields and bottom within days to weeks of the yield peak — you don't get it priced in ahead of time; the peak is the pricing-in. The damage scaled with how far and fast yields ran and whether the Fed was leaning in: −10% (2023) to −20/22% (2018, 2022). When a yield peak was followed by a crash, rates weren't the cause — 2007 was credit (the S&P made a new high four months after the yield peak), 2025 was tariffs. 2008 is the wrong analog for a rates scare: in 2007 a 5.4% long bond was normal after years at 5–6%. Today's 5.3% arrives after fifteen years near zero — the losses sit on duration owners, and the transmission to equities runs through mortgages, refinancing walls and the multiple, not a bank run.

Where that puts the numbers. A 2023-sized scare is −10% from the high: SPY ≈ $700, which is the 200-day ($703) and the 50-week — a normal correction inside the trend. A 2018/2022-sized one is −20%: ≈ $620, below the value area ($639–696) and only plausible if credit joins. The "−22% to the 200-week average" doing the rounds is off on its own arithmetic — the 200-week sits at ≈ $554, −28% from Monday's close, which is 2022 plus more; that is a credit-event outcome, not a rates one.

What would change the read

Each new 30-year high says the peak isn't in. Watch four things: the 30-year failing to make a new high on a bad print (peak); HYG losing its 200-day (~$78.35) (credit stops confirming); SPY closing under the 200-day (~$703); Brent through $100 — Rapidan's read is China's 4-mb/d import "crash diet" is what has held crude down, and refiners can't help once it ends. The AI capex bond wave is the one new mechanism worth its own note — ~$600 billion borrowed in a year by the "collective AI enterprise" (High Frequency Economics), a second hyper-borrower competing with the Treasury for the same savings; a separate piece is coming.

Those four legs are tracked mechanically as desk calls below — SPY, HYG, TLT and the Brent proxy — so each fires as a flag on a settled close, not on a headline. Until one lights, this is a bond-holder's problem and a headline-writer's opportunity.

Desk Call

Ticker Stance Entry / condition Invalidation Review by
SPY Watch no new entry; hold the trend while the 200-day (~$703) and credit hold close < ~$703 with HYG under its 200-day 2026-09-19
HYG Watch (the tell) not a trade — the confirmation signal close < ~$78.35 (200-day) 2026-09-19
TLT Watch (turn) own it only after the yield peak prints: reclaims the 20-day (~$82.60) on a bad-for-bonds day new price low after the reclaim 2026-09-19
BNO Watch (the oil leg) not a trade — BNO ~$57 ≈ Brent ~$100 at today's fund/futures ratio settled close > ~$57 2026-09-19
VLO / PSX Pass the diesel-crack trade is made (RSI 74.8 / 80.0 at the highs) n/a 2026-09-19

Sources