Elevated caution — markets are priced with a thin margin of safety.
Credit conditions
A separate signal, not part of the score above: whether corporate credit spreads are widening fast enough to suggest genuine financial-system stress, rather than just an expensive market. Based on the Baa–10Y spread's 6-month change.
The three inputs
2s10s Yield Curve Spread
0.38%
10Y minus 2Y Treasury yield. Both ends have risen sharply this year — the 10Y is near a 20-year high — leaving a shallow spread, still below the C-tier line of 0.50%.
Shiller PE (CAPE)
41.1
Price over 10-year real earnings. Past the F-tier line of 40 — second-highest reading on record.
Excess CAPE Yield
1.0%
Earnings yield over the real 10Y Treasury yield. Thin premium — the smallest since 2002.
Last 36 months
Recent trend, monthly
Yield curve is real every month. The most recent 12 months are fully real for all three inputs; the prior 24 are interpolated between real quarterly/annual anchors. Hover or touch the line to see any month's reading.
Find similar periods in history
Searches all 597 months for the closest matches to the current reading, based on all three tiers together, not just the composite average.
What's driving this reading
The Shiller PE has climbed to roughly 41 — the second-highest reading in the index's 145-year history, behind only the December 1999 dot-com peak — as AI-related capital spending and expectations for continued mega-cap earnings growth have kept bidding equity prices higher, even as the rate backdrop has turned less friendly. The yield curve sits at a modest, uneasy +0.3–0.5%: both ends of the curve have risen sharply this year, with the 10-year near a two-decade high, so this is a shallow spread built on broadly rising yields rather than a healthy steepening — enough to hold D tier, not reach C. Meanwhile the Excess CAPE Yield — the cushion stocks offer over bonds — has thinned to about 1%, its smallest premium since 2002, as elevated equity prices have outrun the rise in real 10-year yields.
The backdrop behind those moves shifted this month. The Fed had been cutting rates through 2024 and 2025, but on September 16, 2026 it reversed course with its first hike since 2023, taking the target range to 3.75–4%, in response to an inflation flare-up tied to an oil-price shock from the Iran conflict, even with the labor market still holding up. So the market's still-elevated valuation is now being carried mostly by AI-capex enthusiasm and resilient earnings, without the tailwind of easier money it had for most of the past two years — a central bank actively fighting inflation, not helping the cost of capital. That combination — a very expensive market whose two supports (cheap money and a thick bond cushion) are both fading at once — is what an E-tier composite is built to flag: not the market's rarest extreme, but a period asking to be held with less complacency than usual.
Full history, month by month
Full history, monthly, backtested
Every month, 1977–2026, run through your exact tier formula. The final point is always this month's actual reading, not a January snapshot. Hover or touch the line to see any month's reading.
Amber/red marks: months where corporate credit spreads widened fast enough to suggest genuine financial-system stress (see the credit-conditions note above).
Compare two periods
Weighting: the composite now uses 25% yield curve / 37.5% Shiller PE / 37.5% Excess CAPE Yield, not an even three-way split. Backtested against six other weightings across the full 1977–2026 history: equal weighting missed the 2022 bear market by about as much as leaning harder into the yield curve did (the curve re-steepens right before recessions hit, which makes curve-heavy weightings least defensive exactly when they should be most defensive); dropping Excess CAPE Yield entirely more than doubled false alarms during the calm 2013–2019 stretch. This weighting was the one that improved on 2022 without adding false alarms elsewhere — though no combination of these three inputs would have flagged 1987 or the 2008 crisis, both of which were credit and liquidity events rather than valuation or yield-curve ones. Happy to share the full comparison table if useful.
Credit-stress overlay: added to address that exact 1987/2008 gap, but deliberately kept separate rather than folded into the score. Blending the credit spread's raw level into the composite made it read as less cautious at nearly every historical peak tested, including 2008 — because valuations look "cheap" precisely as a credit crisis crashes prices, which is correct for a valuation score but useless as a crisis flag. Its 6-month rate of change is a better-scoped signal: it flagged the 2007–08 credit crisis within a month of the market's top (16 months before the eventual bottom), stayed quiet through 1987 (correctly — that was a liquidity event, not a credit one), and showed only a mild wobble in 2022 versus a much sharper move during real crises. It's shown here as its own flag, not a fourth vote in the average.
Data behind these charts: both charts plot every month, not one point per year. The yield curve spread is real for all 597 months shown, straight from the Federal Reserve (FRED, series T10Y2YM) — no exceptions. Shiller PE and Excess CAPE Yield are exact, real monthly figures from Robert Shiller's own published dataset for January 1977 through January 2016, and again for the most recent 12 months; the 116 months in between are interpolated between real annual anchor points and recalibrated against the real October 2025 reading, rather than independently verified month by month. The final point on both charts is always the current month's actual reading, never a stale snapshot.
Current reading and monthly chart as of September 2026 · annual chart real 1977–2026